1 Algebra Basics

§1.1 Systems of Equations

Systems of equations appear in problems among all subjects, so it’s essential to know how to
solve them. There are generally two ways in which we solve systems of equations: elimination
and substitution.

Definition 1.1. Elimination is where we combine the results of a system of equations to get
a new equation with fewer variables. We will most often use elimination to solve systems of
linear equations, which you are probably already familiar with.

Example 1.2. Solve the following system of equations:

5x + 6y = 107
2x + 3y = 50.

Solution. Multiplying the second equation by 2, we get that

4x + 6y = 100.

Subtracting this new equation from the first equation, we eliminate the y variable and we find
x = 7. Plugging this back into our first equation, we have that 6y = 107 − 35 = 72, so y = 12.
Therefore, the solution to the system of equations is (x, y) = (7, 12). 

Definition 1.3. Substitution is where we express one variable in terms of another, and we
substitute the result into another equation, effectively eliminating a variable. Substitution can
almost always be used, even with more complicated systems with nasty terms.

Example 1.4. Solve the following system of equations:

x−y =2
2 2
x = y + 2y = 40.

Solution. Adding y to both sides of the first equation gives us x = y + 2. Substituting y + 2

in for x in the second equation gives us

(y + 2)2 + y 2 + 2y = 40.

After expanding and rearranging, we can factor the quadratic as 2(y + 6)(y − 3) = 0. Thus,
y = −6 or y = 3. Since x = y +2, our two solutions are (x, y) = (−4, −6) and (x, y) = (5, 3). 

Of course, systems of equations won’t always be so tame, nor will they necessarily have two
variables. However, we can still use the general principles of substitution and elimination; it
may just take a little more manipulation.

1
Everaise Academy (2020) Math Competitions I

Example 1.5 (AMC 12). If x, y, and z are positive numbers satisfying

x + 1/y = 4, y + 1/z = 1, and z + 1/x = 7/3

then what is the value of xyz?

Solution. The most straightforward way to solve this is to substitute in the expressions. Sub-
tracting y1 from both sides of the first equation, we get x = 4 − y1 . Substituting this into the
last equation, we have that
1 7
z+ 1 = 3.
4− y
After simplifying the fraction and rearranging, we get
7 y
z= − .
3 4y − 1
which yields
25y − 7
z= .
12y − 3
12y−3
Substituting this in to our second equation, we get that y + 25y−7 = 1. Multiplying both sides
by 25y − 7 and rearranging, we get the quadratic

25y 2 − 20y + 4 = 0.

This factors into (5y − 2)2 = 0, so y = 25 . Solving for x and z, we have that x = 3
2 and z = 35 .
Thus, the value of xyz is 1. 

Remark. There’s a much cleaner solution — try multiplying all three equations and continue
from there!

Sometimes, the equations may not look as nice. After substituting or eliminating, we may end
up with an equation just as messy as the original. There might seem to be too little information
to solve them! In such situations, we have to get more creative.

Example 1.6. If x2 + y 2 + z 2 = 14 and x + y + z = 6, find xy + xz + yz.

Solution. We are given three unknowns, but only two equations; we will probably have to
somehow introduce xy + xz + yz by manipulating the equations. In such problems, a good
strategy is some wishful thinking and persistence.

The first equation contains a sum of squares; how can we get this from the other equations?
We might think of squaring the second equation to get x2 , y 2 , and z 2 terms and hoping it
expands nicely. Indeed, squaring the second equation gives us

x2 + y 2 + z 2 + 2xy + 2xz + 2yz = 36.

Subtracting the first equation from this and dividing by 2, we get that xy + xz + yz = 11. 

2
Everaise Academy (2020) Math Competitions I

Remark. This manipulation with x2 + y 2 + z 2 and xy + xz + yz comes up very often in

algebra problems—it’s a good trick to remember.

Example 1.7. Solve the following system of equations:

3x2 − 8xy + 4y 2 = 28
x2 − 3xy + 2y 2 = 56.

Solution. Our options here are limited, and the terms look complicated. Looking for numerical
coincidences is always a good idea, though; notice that we can multiply the first equation by
two to get it equal to the second equation. This will cancel out the constant term. Why is this
important?
Well, if we eliminate the constant term, we will have a quadratic in x and y that we can factor,
giving us a relation between x and y after we use the zero product property. We’ll demonstrate
that now. Multiplying the first equation by 2 and setting the two equations equal, we have

6x2 − 16xy + 8y 2 = x2 − 3xy + 2y

Rearranging all the terms to one side gives us

5x2 − 13xy + 6y 2 = 0

we can treat the equation as a quadratic in terms of x with y being constant. Factoring yields
(x − 2y)(5x − 3y) = 0—we can get this by noting that if we were to factor the expression on
the left-hand-side into two binomials, one would have an x and the other would have a 5x, so
we could test out possibilities for y.
This gives us two possibilities: either x = 2y or x = 35 y. Now we can plug both possibilities
into the first equation.

Setting x = 2y into our first equation, we get that 3(2y)2 − 8(2y)y + 4y 2 = 28, which yields
0 = 28. Thus the case x = 2y gives no solutions. Setting x = 53 y, we get that 25
7 2
y = 28, which
yields y = ±10. This gives us the solutions (x, y) = (6, 10), (−6, −10). Both pairs satisfy our
original equations, so we are done. 

In this problem, we dealt with the given information by finding a quadratic that we could
factor. This is a common strategy in algebra problems. Try the following exercise, and look for
a way to make it a quadratic!
Exercise 1.8. Solve 4x + 5 · 6x + 6 · 9x = 0.

Example 1.9 (Brilliant). Solve the system of equations:

x2 + xy = 20
y 2 + xy = 30.

Solution. Looking at the two equations individually doesn’t seem to be too helpful. However,
if we add the two equations together, we can again get a quadratic in x and y. We get x2 +

3
Everaise Academy (2020) Math Competitions I

2xy + y 2 = 50, which factors into (x + y)2 = 50. This looks like a good start, but we still need
more information.
Since we added the two equations initially, we might consider subtracting them, too. Indeed,
subtraction cancels the xy-terms, yielding y 2 − x2 = 10, which factors into (y + x)(y − x) = 10.
Notice that the two equations we obtained share a (x+y) factor, which motivates us to divide
x+y
them. Dividing the first equation by the second gives us y−x = 5, so y = 32 x. Substituting this
into our (x + y)2 = 50 equation, we obtain
25 2
x = 50.
4
√ √ √ √ √ √
so x = 2 2, −2 2. Thus, our two solutions are (x, y) = (2 2, 3 2), (−2 2, −3 2). Plugging
these in, our original equations both hold, so we are done. 

Let’s end with a harder example.

Example 1.10 (2010 AIME I). Let (a, b, c) be the real solution of the system of equations
x3 − xyz = 2, y 3 − xyz = 6, z 3 − xyz = 20. The greatest possible value of a3 + b3 + c3
can be written in the form m
n , where m and n are relatively prime positive integers. Find
m + n.

Solution. Given that we’re trying to find a value for a3 + b3 + c3 , the most obvious first thing
to consider doing is rewriting the equations as

a3 = 2 + abc
b3 = 6 + abc
c3 = 20 + abc.

So a3 + b3 + c3 = 28 + 3abc, and we’re looking for abc, essentially. Therefore we make the
substitution abc = n, yielding

a3 = 2 + n
b3 = 6 + n
c3 = 20 + n.

Our next step should make a lot of sense: it’s hard to deal with one equation at once, but if
we multiply them all and use abc = n, we get

n3 = (n + 2)(n + 6)(n + 20).

This is actually a quadratic — the n3 terms cancel out. Solving it yields n = − 15 7 , −4

as solutions; from here, given that we’re actually looking for the maximum possible value of
a3 + b3 + c3 , we plug in n = abc = − 15 151
7 , yielding an answer of 7 . In the form the question
asks for, our answer is 151 + 7 = 158, and we are done. 

That was a tough problem, and we had to use a lot of tools to solve it. The key steps were
figuring out we only needed to find abc, then rearranging the equations to get a polynomial in
abc.

4
Everaise Academy (2020) Math Competitions I

Remark. A recurring theme throughout all our examples is that if you don’t see the solution
immediately, mess around with the givens, and eventually you will get a useful equation.
Look for expressions that could be key to solving the problem, like a difference of squares. If
you have more variables in equations, then don’t try to solve for individual values directly;
maybe you can’t get an equation only in terms of a, but maybe you can get an equation
in terms of just a + b or ab.

Exercise 1.11. Solve the system of equations.



 a+b+c=7
b+c+d=6

 c+d+a=9


d+a+b=8

Exercise 1.12 (ARML 1990). If a + b = c, b + c = d, c + d = a, and b is a positive integer,

compute the greatest possible value for a + b + c + d.
Exercise 1.13. Solve the following system of equations:

x2 − 6y = 7
y 2 − 10x + 41 = 0.

§1.2 Mean, Median, and Mode

We learn the definitions of mean, median, and mode early on in school, but here we’ll review
them in detail. Mean, median, and mode problems usually appear early in the AMC and are
generally quite simple. It would be a shame to ignore these concepts and miss very easy points!

Definition 1.14. The mean, or average, of a set of numbers is equal to the total sum of all of
the numbers, divided by however many numbers are in the set. More succinctly, for numbers
a1 , a2 , ..., an , the mean is equal to
a1 + a2 + · · · + an
.
n
One can find the median of a set of numbers by ordering the numbers in increasing order,
then finding the number in the middle. If there is an even number of numbers, then the median
is found by taking the mean of the middle two numbers. For example, the median of {1, 3, 5}
is 3, while the median of {1, 3, 5, 6} is 3+5
2 = 4.
The mode is the number that appears most frequently in a set of numbers. We can have one
or multiple modes. For example, the unique mode of {1, 2, 2, 4} is 2 while {1, 1, 2, 2, 4} has two
modes, 1 and 2.

Let’s get right into the examples.

Example 1.15 (2016 AMC 10). The mean, median, and mode of the 7 data values

60, 100, x, 40, 50, 200, 90

are all equal to x. What is the value of x?

5
Everaise Academy (2020) Math Competitions I

Solution. The mean is equal to the total sum, divided by 7. Also, we’re given that this mean
is equal to x, so we can write
60 + 100 + x + 40 + 50 + 200 + 90 540 + x
= x =⇒ = x.
7 7

Hence, 540 + x = 7x, so x = 90. Indeed, if x = 90, then the mode and median of the numbers
are both 90, we can be sure that we’re right. 

Example 1.16 (2019 AMC 12B). What is the sum of all real numbers x for which the
median of the numbers 4, 6, 8, 17, and x is equal to the mean of those five numbers?

Solution. In terms of x, the mean of the set is

4 + 6 + 8 + 17 + x x + 35
=
5 5
The median, however, can take three values, depending on the value of x.
1. If x ≥ 8, then 8 is the median.
2. If x ≤ 6, then 6 is the median.
3. If 6 < x < 8, then x is the median.
Make sure to verify our above claims for yourself! Now, we can take each of them on case at a
time.
x + 35
1. = 8, so x = 5.
5
x + 35
2. = 6, so x = −5.
5
x + 35
3. = x, so x = 8.75.
5
35 35
However, we cannot simply conclude that our answer is 5 − 5 + = . We first have to check
4 4
that that the solutions satisfied the conditions for each case.
Let’s look at case 1: Does it make sense? Notice that we said that x ≥ 8, but our solution
was x = 5, a contradiction. So this case doesn’t actually work.
The second case clearly works, as x = −5 ≤ 6. The third case is invalid because 8.75 is not
between 6 and 8. Thus, our answer is just −5. 

Example 1.17 (2017 AMC 10B). Last year Isabella took 7 math tests and received 7 different
scores, each an integer between 91 and 100, inclusive. After each test she noticed that the
average of her test scores was an integer. Her score on the seventh test was 95. What was
her score on the sixth test?

Solution. First, let’s make use of the divisibility condition: the sum of all seven scores must
be divisible by 7, as we’re given that the average of Isabelle’s scores is an integer. The largest
possible sum of scores is

100 + 99 + 98 + 97 + 96 + 95 + 94 = 679 = 7(97).

6
Everaise Academy (2020) Math Competitions I

The smallest possible sum is

91 + 92 + 93 + 94 + 95 + 96 + 97 = 658 = 7(94).

This means that the only possible sum of scores of all seven tests are 7(94), 7(95), 7(96), and
7(97), or 658, 665, 672, and 679 respectively.

In order to get the sum of the first six test scores, we can subtract 95 from each, getting 563,
570, 577, and 584. The sum should be divisible by 6, since the average of the first 6 test scores
is an integer. 570 is the only one of these integers divisible by 6. Thus, 570 is the sum of the
first six test scores.

Now, note that the sum of the first five scores must be divisible by 5. Thus, if 570 minus a
test score yields a multiple of five, then that test score must be a multiple of five, namely either
95 or 100. Since all scores must be different, 95 can’t be used, so the sixth test score is 100. 

Example 1.18 (2019 AMC 10A). Melanie computes the mean µ, the median M , and the
modes of the 365 values that are the dates in the months of 2019. Thus her data consists
of 12 1s, 12 2s, . . . , 12 28s, 11 29s, 11 30s, and 7 31s. Let d be the median of the modes.
Which of the following statements is true?

(A) µ < d < M

(B) M < d < µ
(C) d = M = µ
(D) d < M < µ
(E) d < µ < M

Solution. Let’s try to determine what is easy to solve for first. In most cases (including this
problem), the mode is easiest to determine. There are 28 of them, since the numbers 1, 2, 3, ...28
all appear 12 times each, making each the most common. Thus, d = 14.5.

Now let’s locate the median. Since there are 365 values, the middle value is the 183rd one.
Thus we can determine the 183rd smallest value by counting up by 12s starting from 1. As
12 · 15 = 180, it follows that the median M is 16, as the 183rd day would appear in a group of
16s.

Determining how big the mean is takes some ingenuity. Direct computation seems difficult, so
let’s do something more clever. Assume that every month had 31 days—then the mean would
clearly be 16. But since some months have less than 31 days, the mean must be less than 16,
so µ < M .

Lastly, let’s compare µ and d. Note that if every month had 28 days, then µ would be equal
to 14.5; however, there are more days than that in some months. Therefore d < µ.
It thus follows that d < µ < M , and choice (E) is our answer. 

7
Everaise Academy (2020) Math Competitions I

Remark. Two strategies come to mind as we solved the above problem:

1. We never actually computed the exact value of µ; we just found it relative to d and
M . This was crucial, as computing the mean would be incredibly time-consuming.
It is important to note that you don’t always have to find exact values if the question
doesn’t ask for them.
2. We also solved similar but easier problems to help us gain insights on the original
problem (like considering if all months had 31 and 28 days.) This strategy is almost
always helpful, too.

Exercise 1.19. Suppose that A and B are disjoint sets. Assume set A has three elements that
have a mean of 7, and set B has four elements that have a mean of 8. What is the mean of the
7 elements in the union of A and B?
Exercise 1.20 (2000 AMC 12). When the mean, median, and mode of the list

10, 2, 5, 2, 4, 2, x

are arranged in increasing order, they form a non-constant arithmetic progression. What is the
sum of all possible real values of x?

§1.3 Rate Problems

Rate problems are especially common in math competitions and can be quite tricky without
practice. They usually require making an observation about a system of moving objects. Typ-
ically, the three quantities to pay attention to are time (t), speed (r), and distance (d). These
have a special relationship:
d = rt.
Often, this equation can help find an unknown quantity if two quantities are known and the
other one isn’t. Note that r can be any kind of rate, from miles traveled per hour to the number
of books read in a day.
Make sure that you convert units accordingly! For example, if the rate is in mph,
make sure that distance is in miles and time in hours.

Example 1.21 (2018 AMC 10B). Sam drove 96 miles in 90 minutes. His average speed
during the first 30 minutes was 60 mph (miles per hour), and his average speed during the
second 30 minutes was 65 mph. What was his average speed, in mph, during the last 30
minutes?

Solution. During the first 0.5 hours, Sam’s average speed was 60 mph. Using d = rt, his
distance traveled was 60 × 0.5 = 30 miles during the first 30 minutes. During the second 30
minutes, using d = rt, we get that Sam traveled 65 × 0.5 = 32.5 miles.
We are given that during all of the 90 minutes, Sam traveled 96 miles. Thus, he traveled
96−30−32.5 = 33.5 miles total in the last half-hour. Thus, his speed then was d/t = 33.5/0.5 =
67 mph. 

8
Everaise Academy (2020) Math Competitions I

Example 1.22. Peter drives to work in the morning at an average speed of 30 mph. He
gets hungry at night so he drives quicker on the way back home, at an average speed of 45
mph. Assume he takes the same route going to and coming back from his work place. If
the combined time taken for this back-and-forth trip is 1 hour, how long does it take for
him to drive to work in the morning, in minutes?

Solution. Let the time in hours that Peter takes to drive to work in the morning be t1 , and let
the time in hours that Peter takes to drive back home at night be t2 . We know that t1 + t2 = 1
— note that we selected hours as our unit because the speeds we’re given are in mph. Let the
distance between his home and work be d. Then, we have d = 30t1 for the morning commute
and d = 45t2 for the night trip.
d d
Rearranging these equations, we get t1 = 30 and t2 = 45 . Substituting this into our first
equation gives us
 
d d 1 1 d
t1 + t2 = + = 1 =⇒ d + = 1 =⇒ = 1 =⇒ d = 18.
30 45 30 45 18

To finish, we go back to the question. It’s asking us to find t1 , so we plug in d into our
equation for t1 , to get t1 = 18
30 hours or t1 = 36 minutes. 

This next problem shows up very often in competitions in different forms, so make sure to
understand what is going on.

Example 1.23. It takes Alice 3 hours to complete a job. It takes Bill 7 hours to complete
the same job. If the two collaborate on this job, how long will it take for the job to be
completed?

Solution. If it takes Alice 3 hours to complete a job, this means that in 1 hour, she completes
1 1
of the job. Similarly, in 1 hour, Bob completes of his job. This means that if both of them
3 7
work together, they can complete
1 1 10
+ =
3 7 21
21
of the job in 1 hour. This means that to complete the job, it takes both of them hours. 
10
Rate problems generally just boil down to identifying what quantity is represented by each of
the variables d, r, and t, then applying d = rt as necessary.
Exercise 1.24 (2007 AIME I). A 100 foot long moving walkway moves at a constant rate of 6
feet per second. Al steps onto the start of the walkway and stands. Bob steps onto the start
of the walkway two seconds later and strolls forward along the walkway at a constant rate of 2
feet per second. Two seconds after that, Cy reaches the start of the walkway and walks briskly
forward beside the walkway at a constant rate of 4 feet per second. At a certain time, one of
these three persons is exactly halfway between the other two. At that time, find the distance
in feet between the start of the walkway and the middle person.
Exercise 1.25 (1951 AHSME). Tom, Dick and Harry started out on a 100-mile journey. Tom
and Harry went by automobile at the rate of 25 mph, while Dick walked at the rate of 5 mph.
After a certain distance, Harry got off and walked on at 5 mph, while Tom went back for Dick
and got him to the destination at the same time that Harry arrived. How many hours did the
trip take?

9
Everaise Academy (2020) Math Competitions I

§1.4 Absolute Value

Definition 1.26. The absolute value of a real number x, denoted by |x|, is defined as follows:
(
x for x ≥ 0,
|x| =
−x for x < 0.

Note: We will only deal with real numbers in this section. |x| has a different meaning when x
is a complex number, which will be defined on Day 5.

When dealing with a problem involving absolute values, the most common approach in AMC
and early AIME is to use casework with two cases: either the expression between the absolute
value signs is negative, or it is nonnegative.

Example 1.27 (2004 AMC 10A). What is the value of x if |x − 1| = |x − 2|?

Solution. We have two cases:

1. Both |x − 1| and |x − 2| are greater than or equal to 0, or both are less than or equal
to zero. Then equation simplifies to x − 1 = x − 2. This gives us −1 = −2, which is a
contradiction. Thus, we know only one expression is positive, while the other is negative.
2. Either |x − 1| or |x − 2| is positive, the other is negative. The equation simplifies to
x − 1 = 2 − x, so x = 23 .
It follows that our only working value is x = 32 , so that is our answer. 

Example 1.28 (2011 AMC 12A). Suppose that |x + y| + |x − y| = 2. What is the maximum
possible value of x2 − 6x + y 2 ?

Solution. We have four cases this time, and these cases include all of the possible values of x
and y.
1. (x + y) + (x − y) = 2 yields x = 1. Note that this case also forces y to lie between −1 and
1.
2. −(x + y) − (x − y) = 2 yields x = −1. Note that this case also forces y to lie between −1
and 1.
3. (x + y) − (x − y) = 2 yields y = 1. Note that this case also forces x to lie between −1 and
1.
4. −(x + y) + (x − y) = 2 yields y = −1. Note that this case also forces x to lie between −1
and 1.
These cases imply that y 2 = 1 is always the maximum possible value of y 2 , and that x2 = 1 is
always the maximum possible value of x2 . It turns out, then, that we can just maximize −6x
by setting x = −1, which yields x2 − 6x + y 2 = 8 for (x, y) = (−1, 1). Note that it is clear that
we cannot obtain a number any larger than this, because −6x is maximized at x = −1, and x2
and y 2 both achieve a maximum possible value of 1. 

Exercise 1.29. Prove the following properties of |x|:

√
1. |x| = x2
2. |x| ≥ 0,

10
Everaise Academy (2020) Math Competitions I

3. |x| = 0 ⇐⇒ x = 0,
4. |xy| = |x||y|,
5. |x + y| ≤ |x| + |y|.
Exercise 1.30 (2000 AMC 12 Problem 5). If |x − 2| = p, where x < 2, then what is x − p?
Exercise 1.31. There are exactly N distinct rational numbers k such that |k| < 200 and
5x2 + kx + 12 = 0
has at least one integer solution for x. What is N ?
The absolute value problems in these sections weren’t so bad; when graphs are involved,
though, they become harder. We’ll go over graphs in general as well as absolute value problems
involving graphs in the following section.

§1.5 Graphs
Definition 1.32. A graph is a visual representation of a function or relation. We will usually
obtain a graph from a function like f (x) = x2 +2x by plotting each value of x as the x-coordinate
and the corresponding value of f (x) as the y-coordinate of a point on the xy-plane.

In this section, we will be looking at the graphs of several different types of equations. In
a graph, the x and y-axes separate the plane into 4 regions. The top right is called the first
quadrant, the top left is the second quadrant, the bottom left is the third quadrant, and the
bottom right is the fourth quadrant.

§1.5.1 Lines

When we are graphing lines, there are two common forms: point-slope form and the slope-
intercept form.
Given a point on the line and that line’s slope, we can use point-slope form. If the point
we are given is (x0 , y0 ) and its slope is m, then the equation of the line is
y − y0 = m(x − x0 ).
1
For example, the line passing through (2, 5) with slope 2 has equation y − 5 = 12 (x − 2).
Slope-intercept form is the more commonly used of the two forms; if m is the slope of a
line and b is its y-intercept — that is, the line passes through (0, b) — then the equation of the
line is
y = mx + b.
There are two important facts to remember regarding the slopes of lines. First, two lines are
parallel if and only if their slopes (m) are equal. Second, two lines are perpendicular if and only
if their slopes multiply to −1 (in other words, one slope is the negative of the reciprocal of the
other slope.)

§1.5.2 Higher Degree Polynomials and Absolute Value

While graphs of higher degree polynomials and absolute values are not seen too often in com-
petitions, it is still good to know what they look like. The most important thing to remember
is the relationship between the y and x values, as that is what’s going to help you solve the
problem. Drawing a diagram can definitely help, but the equation is what truly matters.

11
Everaise Academy (2020) Math Competitions I

With that being said, here are a few graphs that you should know. If you forget what they
look like, you can always make a table and plug in values of x to get a rough idea of what graph
looks like.

Quadratic (y = x2 ):

Cubic (y = x3 ):

Absolute Value (y = |x|):

Exercise 1.33. Sketch the graph of y = x3 + 3x2 + 5x + 6.

Exercise 1.34. Sketch the graph of y = x4 + 4x3 + 2x2 + 1.

12
Everaise Academy (2020) Math Competitions I

§1.5.3 Example Problems

Example 1.35. (2005 AMC 12B) What is the area enclosed by the graph of |3x|+|4y| = 12?

Solution. We begin by using casework, which almost always works when dealing with absolute
values.
1. 3x + 4y = 12 yields y = − 43 x + 3
2. 3x − 4y = 12 yields y = 34 x − 3
3. −3x + 4y = 12 yields y = 43 x + 3
4. −3y − 4y = 12 yields y = − 34 x − 3
We can graph the equations above (because they are in slope-intercept form) to get:

The area enclosed by the graph of |3x| + |4y| = 12 is a rhombus with vertices at (4, 0), (−4, 0),
(0, 3), and (0, −3). The area of a rhombus is half the product of its diagonals, which in this
case is 21 · 8 · 6 = 24. 

Example 1.36 (2006 AMC 12A). Which of the following describes the graph of the equation
(x + y)2 = x2 + y 2 ?
(A) the empty set (B) one point (C) two lines (D) a circle (E) the entire plane

Solution. We begin by simplifying the equation; note that

x2 + y 2 = (x + y)2
= x2 + y 2 + 2xy.

This means that 2xy = 0. Then either x = 0 or y = 0, so the graph actually is just the x and
y-axes, or two lines. Our answer is (C). 

13
Everaise Academy (2020) Math Competitions I

Example 1.37 (2007 AMC 12A). A piece of cheese is located at (12, 10) in a coordinate
plane. A mouse is at (4, −2) and is running up the line y = −5x + 18. At the point (a, b)
the mouse starts getting farther from the cheese rather than closer to it. What is a + b?

Solution. The hardest part of this problem is figuring out what we actually want. The easiest
way to approach this problem is to draw the line y = −5x + 18 and label the points (12, 10)
and (4, −2).

Let’s do a bit of geometry: the mouse is closest to the cheese when it is standing at the foot of
the perpendicular from (12, 10) to the mouse’s path. Therefore we want to find the intersection
of the perpendicular from (12, 10) to y = −5x + 18 with the line y = −5x + 18.
Calling this perpendicular line `, we note that ` has slope 15 by our negative reciprocal rule,
and that it passes through (12, 10). Therefore, by point-slope form, the equation of ` is
1
y − 10 = (x − 12).
5
But we already know that the other line has equation y = −5x + 18, so this yields a system
of two equations. It’s possible to solve for their intersection and get (2, 8), so a = 2 and b = 8
implies that our answer is 2 + 8 = 10. 

§1.5.4 Polynomial Graph Intuition

The next problem we tackle requires a lot of graphical intuition, so before we get into it, we’ll
go over what polynomial graphs tell us.

The first thing we should know is that cubics (also known as third-degree polynomials) have
at most two “bends” (extreme values), or two places where the graph changes from increasing
to decreasing. The locations at which the graph goes from increasing to decreasing are called
local maxima (plural of “maximum”), and the locations where the graph goes from decreasing
to increasing are called local minima (plural of “minimum”). This can be generalized for a
n-degree polynomial as well; they have at most n − 1 extreme values.

14
Everaise Academy (2020) Math Competitions I

For example, here is a graph of a cubic:

Note that point A is a local maximum and point B is a local minimum. Here’s an arbitrary
quartic:

This graph has local minima at points A and C and a local maximum at point B.

Of course, polynomials can have fewer than n − 1 extreme values, they just can’t have more.
One such example of a n-degree polynomial with fewer than n − 1 extreme values is the cubic
function y = x3 . If you look at the image earlier in the handout, you will realize that it only
has one “extreme value”, and this value is actually neither a maxima nor a minima!

15
Everaise Academy (2020) Math Competitions I

The reason this is possible is because extreme values are not limited to maxima and minima;
they could be neither. The definition we have here for an extreme value is rather vague, for we
are not yet equipped with the knowledge to formally define these extreme values (that would
have to wait until calculus). However, for now, it’s fine if we just think about extreme values
as places where the graph of a function “bends,” or changes direction.

Another thing to remember is that as the x value of a polynomial with an odd degree ap-
proaches infinity, it goes in the opposite direction from where it came from. If the polynomial
has an even degree, however, as x approaches infinity, the function travels in the same direction
as where it came from. This is known as a polynomial’s “end behavior.”

1 3 1
Let’s take f (x) = 10 x + 2x2 + 4x + 3 as an example. The leading coefficient is 10 , which is
positive, so we know that f (x) approaches infinity as x gets bigger. f (x) has degree 3, which
is odd, so as x gets smaller and smaller, f (x) approaches negative infinity. We can check that
this is accurate by taking a look at the graph of this function:

Exercise. Consider f (x) = x4 + 3x2 + 5x + 7 and g(x) = −x5 + 3x4 + 2x3 + 10. What is the
end behavior of f (x)? What about g(x)?
Now that we are armed with this knowledge, we are ready to tackle the next problem.

Example 1.38 (2015 AIME I). Let f (x) be a third-degree polynomial with real coefficients
satisfying
|f (1)| = |f (2)| = |f (3)| = |f (5)| = |f (6)| = |f (7)| = 12.
Find |f (0)|.

Solution. We are given six values of |f (x)| that are equal. However, only three values of x can
give the same value of f (x).

16
Everaise Academy (2020) Math Competitions I

There are two ways we could have deduced that; one way is by drawing a graph and seeing
that a horizontal line could not intersect it in more than 3 locations, and the other is by noticing
that f (x) + 12 and f (x) − 12 both have degree 3 and so can have at most 3 roots.

This means that three of the values must be −12, while the other three must be 12. Now,
we can assume without loss of generality that the leading coefficient is negative; this shouldn’t
change our answer since if f (x) satisfies the conditions in the problem, then −f (x) must satisfy
them as well, since the problem statement uses absolute value signs.

Now we can graph some arbitrary cubic with positive leading coefficient with six distinct
values of x such that f (x) = ±12.

Consider an arbitrary cubic graph with negative leading coefficient (sketch one out yourself!).
There are three regions of interest: label the left of the first bend as A, the region between the
first and second bend as B, and the region to the right of the second bend as C (if you drew the
graph correctly, f (x) should be decreasing in region A, increasing in region B, and decreasing
in region C).

By drawing horizontal lines, it is clear that f (x) = 12 exactly once in each region A, B, and
C, and the same is true for f (x) = −12. Then x = 1, 2 are in region A, x = 3, 5 are in region B,
and x = 6, 7 are in region C. From the shape of the graph it is clear that in regions A and C,
we have that f (x) = 12 before f (x) = −12, while the opposite is true in region B. So we have
f (1) = f (5) = f (6) = 12 while f (2) = f (3) = f (7) = −12.
Now, let f (x) = ax3 + bx2 + cx + d. If we plug in x = 1, 2, 3, 5, 6, 7, we get the following

17
Everaise Academy (2020) Math Competitions I

system of equations:

a + b + c + d = 12
8a + 4b + 2c + d = −12
27a + 9b + 3c + d = −12
125a + 25b + 5c + d = 12
216a + 36b + 6c + d = 12
343a + 49b + 7c + d = −12

We have six equations and four variables, so in reality we only need the first four equations to
solve for the different values (we could use any four of the six, but the first four have smaller
coefficients). Solving the system gives us d = f (0) = 72. 

This was a pretty long and tough problem, but make sure you understand all of the deductions
we made using the graph. Visualizing a problem oftentimes leads to much cleaner solutions than
focusing only on the algebra.
Exercise 1.39. Finish the solution for the previous question by solving the system of equations.
Exercise 1.40 (1961 AHSME). If the graphs of 2y + x + 3 = 0 and 3y + ax + 2 = 0 are to meet
at right angles, the what is the value of a?
Exercise 1.41 (2006 AMC 10B). The lines x = 14 y + a and y = 41 x + b intersect at the point
(1, 2). What is a + b?
Exercise 1.42 (1998 AIME). The graph of y 2 + 2xy + 40|x| = 400 partitions the plane into
several regions. What is the area of the bounded region?
Exercise 1.43 (2016 AMC 12A). Which of these describes the graph of x2 (x + y + 1) = y 2 (x +
y + 1)?
(A) two parallel lines
(B) two intersecting lines
(C) three lines that all pass through a common point
(D) three lines that do not all pass through a common point
(E) a line and a parabola

§1.6 Homework Problems

Problem 1.1. [3] Kevin runs at a speed of 350 meters per minute and Vincent at a speed of
380 meters per minute. After half an hour, how far apart are Kevin and Vincent assuming they
run in opposite directions?

Problem 1.2 (2020 AMC 10B). [5] How many ordered pairs of integers (x, y) satisfy the
equation
x2020 + y 2 = 2y?

Problem 1.3 (2001 AIME I). [5]

A finite set S of distinct real numbers has the following properties: the mean of S ∪ {1} is 13
less than the mean of S, and the mean of S ∪ {2001} is 27 more than the mean of S. Find the
mean of S.

18
Everaise Academy (2020) Math Competitions I

Problem 1.4 (2020 AMC 10A). [5] What is the median of the following list of 4040 numbers?

1, 2, 3, ..., 2020, 12 , 22 , 32 , ..., 20202

Problem 1.5 (2010 HMMT). [5] Suppose that x and y are positive reals such that

x − y 2 = 3, x2 + y 4 = 13.

Find x.

Problem 1.6 (2018 AMC 12A). [5] For positive integers m and n such that m + 10 < n + 1,
both the mean and the median of the set {m, m + 4, m + 10, n + 1, n + 2, 2n} are equal to n.
What is m + n?

Problem 1.7 (2002 AMC 12A). [5] The mean, median, unique mode, and range of a collection
of eight integers are all equal to 8. What is the largest integer that can be an element of this
collection?

Problem 1.8 (2018 AMC 10A). [8] Which of the following describes the set of values of a for
which the curves x2 + y 2 = a2 and y = x2 − a in the real xy-plane intersect at exactly 3 points?
1 1 1 1 1 1
(A) a = 4 (B) 4 <a< 2 (C) a > 4 (D) a = 2 (E) a > 2

Problem 1.9 (2016 AMC 10B). [9] What is the area of the region enclosed by the graph of
the equation x2 + y 2 = |x| + |y|?

Problem 1.10 (1989 AIME). [10] Assume that x1 , x2 , . . . , x7 are real numbers such that

x1 + 4x2 + 9x3 + 16x4 + 25x5 + 36x6 + 49x7 = 1

4x1 + 9x2 + 16x3 + 25x4 + 36x5 + 49x6 + 64x7 = 12
9x1 + 16x2 + 25x3 + 36x4 + 49x5 + 64x6 + 81x7 = 123.

Find the value of 16x1 + 25x2 + 36x3 + 49x4 + 64x5 + 81x6 + 100x7 .

19
2 Exponents, Logarithms, Radicals
In this handout, we’ll go over everything related to powers in algebra problems.

§2.1 Exponents & Logarithms

For an integer n, we define an (pronounced “a to the nth power”) as a × a × · · · × a (where a is
| {z }
n
repeated n times). We can extend this definition to the rationals by defining a1/m as the mth
root of a, or the number that yields a when we raise it to the power of m.
The following are some properties of exponents:

an × am = an+m (2.1)
1
a−n = n (2.2)
a
(am )n = amn (2.3)

Let’s start with some very basic examples.

48
Example 2.1. What is 45
?

4×4×4×4×4×4×4×4
Solution. Using the definition at the beginning of the handout, 4×4×4×4×4 = 48−5 =
48
43 = 64 . We could also use properties 2 and 1 to get that 45
= 48 4−5 = 48−5 = 43 . 

Example 2.2. What is (22 )2 ?

Solution. Using the third property, we get that (22 )2 = 24 = 16. Alternatively, we could just
say that (22 )2 = 42 = 16. 

Let’s move on to the logarithm (or “log” for brevity). Here’s the motivation for defining the
logarithm: let’s say we wanted to find x such that 2x = 64. How would we do this? We would
need a function that “undoes” exponentiation (i.e. is the inverse of exponentiation).
In general, if bx = a, then we say that logb a = x (“log base b of a equals x”). Here are some
basic properties of logs that follow from the properties of exponents.
logb 1 = 0 (2.4)
logb x + logb y = logb xy (2.5)
k logb x = logb xk (2.6)
logc a
logb a = (2.7)
logc b
Property (1.7) is known as the change of base formula, and is very useful when we want to
get all our logarithms to have the same base to simplify equations. Once all our logarithms
have the same base, we can repeatedly apply (1.5) and (1.6) to combine our logarithms.

1
Everaise Academy (2020) Math Competitions I

Exercise 2.3. Prove properties (1.4) through (1.7) using the definition of a logarithm and ex-
ponent rules.
1
Exercise 2.4. Prove that logbk x = k logb x.
Exercise 2.5. Prove that − logb x = logb x1 .
1
Exercise 2.6. Prove that logb a = loga b .

Remark. We might also omit the base from a log, in which case it comes in the form log a.
This is usually because the base is assumed to be a number like 10.

Let’s tackle some tricky problems.

Example 2.7 (2013 AIME II). Positive integers a and b satisfy the condition log2 (log2a (log2b (21000 ))) =
0. Find the sum of all possible values of a + b.

Solution. Using the properties of logarithms, we can repeatedly reduce the initial equation.

log2 (log2a (log2b (21000 ))) = 0

log2a (log2b (21000 )) = 1
2a = log2b (21000 )
a
(2b )2 = 21000

Equating exponents, we get

b(2a ) = 1000.
Since a must be positive, and the only powers of two that divide 1000 are 2, 4, and 8, we get
that our three solutions are (a, b) = (1, 500), (2, 250), (3, 125). Thus, our desired sum is 881. 

Example 2.8. If log10 5 ≈ 0.7, how many digits are in 256 ?

Solution. Notice the floor of the base 10 logarithm of a number gives us one less than than the
number of digits it has, since log10 10k = k. For example 1 < log10 400 < 2 and 3 < log10 2000 <
4. Taking the base 10 logarithm of 256 gives us log10 256 = 56 log10 2. Notice that

log10 2 = log10 10 − log10 5 ≈ 0.3.

Since 56(0.3) = 16.8, it follows that 256 has 17 digits. 

Example 2.9 (2019 AMC 12A). Positive real numbers a and b have the property that
p p √ √
log a + log b + log a + log b = 100

and all four terms on the left are positive integers, where log denotes the base-10 logarithm.
What is ab?

2
Everaise Academy (2020) Math Competitions I

Solution. Since we are given that the base of each logarithm is 10, it makes sense to substitute
x = log10 a, y = log10 b. With this substitution, our equation now becomes
√ √ x y
x + y + + = 100.
2 2
Since we know that all terms must be integers, we test even perfect squares for the values of
x, y. After some plugging and chugging, we find that (x, y) = (100, 64) is a solution. Thus our
desired product ab = 10100 · 1064 = 10164 . 

Example 2.10 (2005 AIME I). The equation 2333x−2 + 2111x+2 = 2222x+1 + 1 has three real
roots. Find their sum.

Solution. Notice that the exponents all have some multiple of 111x; thus, it makes sense to
substitute y = 2111x to get a polynomial in y. Solving for x in terms of y, we have that x = log 2y
111 .
If our three roots are x1 , x2 , x3 , then our desired sum is
log2 y1 y2 y3
x1 + x2 + x3 = .
111
3
With the substitution, our equation now becomes y4 + 4y = 2y 2 + 1. Multiplying by 4 and
rearranging, we have that y 3 − 8y 2 + 16y − 4 = 0. By Vieta’s, y1 y2 y3 = 4, so our desired sum
2
is 111 . 

Example 2.11 (1984 AIME). Determine the value of ab if log8 a + log4 b2 = 5 and log8 b +
log4 a2 = 7.

2
Solution. Subtracting the first equation from the second, we have that log4 ab + log8 ab = 2.
Since the bases aren’t the same, we are motivated to change them so we can combine the logs
by multiplying their arguments. Rewriting the equation and repeatedly simplifying, we get that

2
log2 ab log2 ab
+ =2
log2 4 log2 8
a log2 ab
log2 − =2
b 3
a
log2 = 3
b
a
= 8.
b
In our simplification, we used the fact that loga bc = c loga b and loga b = − loga 1b . Substituting
a = 8b back into the first equation, we have that log8 8b + log4 b2 = 5. Rewriting, we get that

log2 8b log2 b2
+ =5
3 2
2 log2 8b + 3 log2 b2 = 30
log2 (64b2 )(b6 ) = 30
64b8 = 230
b = 8.

Thus a = 64 and our desired product ab = 512. 

3
Everaise Academy (2020) Math Competitions I

The main idea in this problem was noticing that the bases of each logarithm were powers of
two. Then, we used change of base to simplify the equations and the problem was much more
straightforward from there. In general, given an equation with logarithms, the first thing you
should try to do is make all the bases the same, if possible.

Example 2.12 (2010 AIME II). Positive

p numbers x, y, z satisfy xyz = 1081 and (log10 x)(log10 yz)+
(log10 y)(log10 z) = 468. Find (log10 x)2 + (log10 y)2 + (log10 z)2 .

Solution. Since we are dealing with a lot of base 10 logarithms, it makes sense
√ to make the
substitution a = log10 x, b = log10 y, and c = log10 z. Thus, we wish to find a2 + b2 + c2 .
Taking both sides log10 in the equation xyz = 1081 gives us

log10 xyz = 81 =⇒ a + b + c = 81.

We also have that

a(b + c) + (bc) = 468 =⇒ ab + bc + ac = 468

from the second equation. Now, we can square both sides of a + b + c = 81 to get a2 + b2 + c2 +
2ab + 2bc + 2ac = 6561. Subtracting 2(ab + bc + ac) = 2(468), we get that a2 + b2 + c2 = 5625.
(Remember when we said this manipulation comes up a lot in algebra problems?) Taking the
square root of both sides gives us our desired answer of 75. 

In this problem, the key strategy was to make some clever substitutions with the given equations,
allowing us to notice the needed algebraic manipulation much more quickly.

Example 2.13 (HMMT 2020 February). Let a = 256. Find the unique real number x > a2
such that
loga loga loga x = loga2 loga2 loga2 x

Solution. Although we are given that a = 256, let’s not substitute it in for the moment. Notice
that all the logarithms have bases with powers of a; this is a clear sign that we should use change
of base. Let’s change the outermost logarithm on the RHS to base a. Our equation becomes
loga loga2 loga2 x
loga loga loga x = .
loga a2

Since loga a2 = 2, we can simplify further:

2 loga loga loga x = loga loga2 loga2 x

loga (loga loga x)2 = loga loga2 loga2 x

Since both sides share the same base, their arguments must be equal, so we can further simplify
the equation to
(loga loga x)2 = loga2 loga2 x.
m
Now, we can make the substitution m = loga x. From change of base, we find that 2 = loga2 x.
Now we can reduce our equation to (loga m)2 = loga2 m
2.

4
Everaise Academy (2020) Math Competitions I
m
loga m
Using change of base on the RHS, we find that 2
2
= loga2 2. Thus, our equation becomes
m
2(loga m)2 = loga .
2
We have a square of loga m on the LHS; this means we probably want to turn this equation
into a quadratic in terms of either loga m or loga m
2 . Let’s try the second option.

Notice that we can write loga m = loga m 8

2 + loga 2. Since a = 256 = 2 , we have that loga 2
m 1
is just a constant; we can write it as loga m = loga 2 + 8 . We now have a quadratic in terms
of loga m2:
m 1 2 m
2(loga + ) = loga .
2 8 2
This factors into (8(loga m 2 m 1 m
2 ) − 1) = 0. Thus, loga 2 = 8 =⇒ 2 = 2 =⇒ m = 4. Substituting
4 32
m back into m = loga x = log256 x, we have that x = 256 = 2 . 

That was a tough problem, and we had to use a lot of substitutions and base changes to solve it.
Many log problems will look similarly intimidating, but they can all be simplified and eventually
solved through substitutions and manipulations.
Hopefully you have a better feel for logarithm and exponent problems now; problems will
usually boil down to having the intuition for what to do next (e.g. getting all the terms to have
the same base or substituting variables in for logarithm terms that are used repeatedly) and
knowing which logarithm properties to use to accomplish it.
Exercise 2.14 (AHSME). If 60a = 3 and 60b = 5, find 12[ (1 − a − b)/(2 − 2b)]
Exercise 2.15 (CMC). Let a and b be positive real numbers satisfying

(log2 a)(log4 b) + a2 = 135 and (log4 a)(log2 b) + b2 = 263.

√
Then the value of ab can be written as p q for positive integers p and q such that q is not
divisible by the square of any prime. What is p + q?

§2.2 Radicals
In the previous section, we discussed exponents in general. Here, we’ll discuss fractional expo-
nents, paying special attention to square roots. Roots have a tendency to show up on algebra
problems that are more number-theoretic, and often require a specific kind of manipulation to
deal with.
We know how to find squares of numbers very easily. For example, to find the square of
41, multiply: 41 ∗ 41 = 1681. However, the operation can be reversed. What number, when
multiplied by itself, yields 1681? This is a lot less trivial than the previous problem because
our methods are much more inefficient; we usually resort to guessing and checking or using a
calculator.

Remark. There actually are algorithms to find the square root of any number by hand,
precise for as many digits as we want (usually numerical methods derived from calculus).
However, they are still too slow to justify using over efficient guessing and checking when
it comes to math competitions.

5
Everaise Academy (2020) Math Competitions I

Finding a number x that multiplies by itself to produce a given number n is equivalent to

looking for the square root of n, which is x. In notation form, the radical is used, as shown.
√
144 = 12

Here, the radical symbol above the 144 means that the square root of 144 is equal to 12.
You may start to wonder why −12 isn’t a solution. Doesn’t (−12)(−12) = 144, making −12 a
possible answer? While the negative version does multiply to give the same result, radicals and
square roots are defined as functions that produce only nonnegative values. You will never see a
square root of a number yield a negative number. The nonnegative square root is known as the
principal square rootprincipal square root of a number. Unless otherwise stated, when
the square root symbol is used, we refer to the principal square root. Observe below:

Example 2.16. What solution(s) exist for the equation x2 = 400?

Solution. Plugging in values or factoring using difference of squares, we find that x = 20 and
x = −20 both work as solutions. 

Compare the above problem with this next one:

√
Example 2.17. What solution(s) exist for the equation x = 400?

Solution. This seems to be the same problem as before. x = 20 works as a solution. However,
due to the nature of principal square roots, they never output negative values. Thus, 20 is our
only answer. 

Of course, there is a reverse operation for finding the cube of a number, fourth power of a
number, and so on. For example, if you want to write the ”4th root of 81”, you can write
√4
81 = 3

Note that if you wish to take an odd root of a negative number, you can expect to get a
√
negative number back. For example, 5 −32 = −2. The idea of a principal root applies only to
even powers.
As we stated before, radicals are essentially exponents. Taking the square root of a number
is equivalent to raising it to the power of 12 . In general, taking the nth root of any number is
equivalent to raising the number to the n1 -th power. This can help simplify expressions:
√
Example 2.18. The quantity 7
16 can be expressed as 2n for some n. What is n?

Solution. We should first notice that 16 is equal to 24 . Also, as discussed before, taking the
7th root is equivalent to raising the number to the 71 -th power. Using exponent rules, this is
equivalent to
1 4
(24 ) 7 = 2 7
Thus, n = 74 . 

6
Everaise Academy (2020) Math Competitions I

Simplifying square roots are also essential in math contests, given that your answer can only
be correct if expressed in simplest form. Note that
√ √ √ √
180 = 2 45 = 3 20 = 6 5.
√
Thus, if you get 180, you have to simplify it. This process is outlined below:
1. Determine a perfect square (greater than 1) that divides into the radicand (the number
under the radical). Usually start with 4 and work your way up.
2. Divide the radicand by your chosen perfect square and replace your old radicand with this
new quotient. Also, put the square root of your perfect square outside of the radicand.
3. Repeat steps 1 and 2 until you cannot locate any more perfect squares. At this point, you
finished.
Let us demonstrate this process:
√
Example 2.19. Simplify 240.

Solution. Using the process outlined above, we’ll try to determine perfect squares that divide
into 240. Let’s start with 4. Since 240
√ is indeed divisible by 4, we now follow step 2 and replace
√
240
240 with 4 = 60 while putting 4 = 2 on the outside, getting the more simplified 2 60.
However, this isn’t the answer because 4 yet again divides into the radicand, 60. Thus, we
follow step 2√again and replace 60 with 604 = 15 and the number √ outside the radical, 2, with
itself times 4 = 2, giving us the even more simplified radical 4 15. √ Noting that no more
perfect squares divide into the new radicand 15, our answer stays as 4 15. 
√
Exercise. Simplify 735.
√
4
√
4
√
We can extend this process to nth roots as well. For example, 324 = 4 · 34 = 3 4 4.

Example 2.20 (Adapted from 2020 MATHCOUNTS State). A sequence is defined by Sn =

n(n + 1)
for all positive integers n. Another sequence is defined as Pn = 2Sn ∗ 2Sn+1 ∗
2 √
2Sn+2 ∗...∗2S2n−1√for all positive integers n. The value of P2019 can be written in simplest
radical form as a b , where a and b are positive integers. What is the value of b?

Solution. Note that

2Sk = k(k + 1)
This means that we can simplify the expression of Pn :

Pn = n(n + 1) ∗ (n + 1)(n + 2) ∗ (n + 2)(n + 3) ∗ ... ∗ (2n − 1)(2n)

Pn = n(n + 1)2 (n + 2)2 ...(2n − 1)2 ∗ 2n

√
Now we should derive a formula for Pn rather than Pn . Note that all the squared terms can
be taken out of the radical as shown:
p p √
Pn = n(n + 1)2 (n + 2)2 ...(2n − 1)2 ∗ 2n = (n + 1)(n + 2)...(2n − 1) 2n2
√
= n(n + 1)(n + 2)...(2n − 1) 2
The problem only asks for the value of b, the value inside the radicand. Clearly, no matter what
n is, the value of b will always stay the same: 2. 

7
Everaise Academy (2020) Math Competitions I

§2.2.1 Conjugates

If we have an expression involving square roots, the conjugate is an expression we can multiply
the original expression by to eliminate all square roots. For a binomial expresson x + y, the
conjugate is x − y as the difference of squares identity will usually resolve any square roots. For
√ √
example,
√ if a, b, and
√ c are integers, the conjugate of a + b c is a − b c, and the conjugate of
√ √
a + b is a − b. Conjugates are very useful for simplifying expressions involving radicals.
For example, if we have a fraction with a radical in the denominator, like
3
√ ,
1+2 5
then we can make the denominator rational by multiplying the top and bottom by the conjugate
of the denominator: √ √
3 · (1 − 2 5) −3 + 6 5
√ √ =
(1 + 2 5)(1 − 2 5) 19
Conjugates also appear very often when we have to deal with nasty radical expressions. Let’s
try a few examples:
p √ p √
Example 2.21 (2011 AMC 10A). Which of the following is equal to 9 − 6 2+ 9 + 6 2?
√ √ √ √
(A) 3 2 (B) 2 6 (C) 7 2 2 (D) 3 3 (E) 6

Solution. Note that the expressions in thep square

√ roots are conjugates
p √ of each other. How can
we make use of this fact? If we let x = 9 − 6 2 and y = 9 + 6 2, we know that x2 + y 2
will be an integer since the square root terms will cancel. We also know x2 y 2 is an integer.
In fact, x2 y 2 = 81 − 72 = 9 is a perfect square, so xy is also an integer. Then notice if we
square the entire expression to get x2 +2xy +y 2 , we will end up with an integer value, which will
give us a simplified form of the expression when we take the square root again! So, computing
(x + y)2 , we get
√ q √ √ √
(x + y)2 = x2 + 2xy + y 2 = 9 − 6 2 + 2 (9 − 6 2)(9 + 6 2) + 9 + 6 2 = 18 + 2(3) = 25.
√ √
Thus, x + y = 24 = 2 6 is our answer. 

Example 2.22 (2018 AMC 10A). Suppose that real number x satisfies
p p
49 − x2 − 25 − x2 = 3
√ √
What is the value of 49 − x2 + 25 − x2 ?

√ √
Solution. We have a pretty messy square root expression here. Since it is in the form a − b,
we should try multiplying
√ by its conjugate to see if we get anything nice. The conjugate is of
√
the form a + b, so we can multiply as follows:
p p p p
( 49 − x2 + 25 − x2 )( 49 − x2 − 25 − x2 ) = 49 − x2 − 25 + x2 = 24

That worked nicely. We are given that

p p
49 − x2 − 25 − x2 = 3

8
Everaise Academy (2020) Math Competitions I

so p p 24
49 − x2 + 25 − x2 = = 8.
3


Though conjugation is useful for square roots, we often have to rationalize denominators with
expressions containing cube roots as well. In these cases, the difference/sum of cubes identities
can be very helpful:
a3 + b3 = (a + b)(a2 − ab + b2 )
and
a3 − b3 = (a − b)(a2 + ab + b2 )

Example 2.23 (2013 AIME

√
I). The real root of the equation 8x3 − 3x2 − 3x − 1 = 0 can be
√
3 3
a+ b+1
written in the form c , where a, b, and c are positive integers. Find a + b + c.

Solution. Looking at the last three terms of the LHS of the equation reminds us of the expan-
sion (x + 1)3 . So it seems like a good idea to rewrite

8x3 − 3x2 − 3x − 1 = 9x3 − (x + 1)3 = 0

Now we can solve this equation quite easily:

√
9x3 − (x + 1)3 = 0 =⇒ 9x3 = (x + 1)3 =⇒
3
9x = x + 1
1
since real cube roots are unique. This gives us x = √
3 . But to get this in the desired form,
9−1
we have to rationalize the denominator. This is where the difference of cubes identity comes in:
we know that √ √ √
3 3 3
( 9 − 1)( 81 + 9 + 1) = 9 − 1 = 8.
Then
√
3 8
9−1= √
3
√ ,
81 + 3 9 + 1
so √ √
3
81 + 3 9 + 1
x= .
8
Thus, our answer is 81 + 9 + 8 = 98. 

Conjugates serve as an important tool for eliminating roots. More generally, its important
to remember algebraic identities like difference of cubes and difference of squares, as these are
essential to eliminating roots and rationalizing expressions, greatly simplifying problems.
5
Exercise 2.24. Rationalize the denominator of √
3 .
7+1
6√
Exercise 2.25. Rationalize the denominator of √
4 .
3+ 4 2

9
Everaise Academy (2020) Math Competitions I

§2.3 Homework Problems

Problem 2.1 (2010 AMC 12A). [4] The solution of the equation 7x+7 = 8x can be expressed
in the form x = logb 77 . We can write b as a fraction pq in simplest form, where p and q are
integers. What is p + q?
Problem 2.2 (1984 AIME). [4] Determine the value of ab if log8 a + log4 b2 = 5 and log8 b +
log4 a2 = 7.
Problem 2.3 (2013 AMC 12A). [4] The sequence log12 162, log12 x, log12 y, log12 z, log12 1250
is an arithmetic progression. What is x?
Problem 2.4 (2008 AMC 12B). [4] A circle has a radius of log10 (a2 ) and a circumference of
log10 (b4 ). What is (loga b)/π?
Problem 2.5 (2000 AIME I). [6] The system of equations
log10 (2000xy) − (log10 x)(log10 y) = 4
log10 (2yz) − (log10 y)(log10 z) = 1
log10 (zx) − (log10 z)(log10 x) = 0

has two solutions (x1 , y1 , z1 ) and (x2 , y2 , z2 ). Find y1 + y2 .

Problem 2.6 (2008 AMC 12B). [7] The sum of the base-10 logarithms of the divisors of 10n
is 792. What is n?
p √ √ √
Problem 2.7 √ (2006
√ AIME
√ I). [8] The number 104 6 + 468 10 + 144 15 + 2006 can be
written as a 2 + b 3 + c 5, where a, b, and c are positive integers. Find abc.
Problem 2.8 (2019 AMC 12A). [10] Define binary operations ♦ and ♥ by
1
a ♦ b = alog7 (b) and a ♥ b = a log7 (b)
for all real numbers a and b for which these expressions are defined. The sequence (an ) is defined
recursively by a3 = 3 ♥ 2 and
an = (n ♥ (n − 1)) ♦ an−1
for all integers n ≥ 4. To the nearest integer, what is log7 (a2019 )?
Problem 2.9 (2019 AIME I). [10] There are positive integers x and y that satisfy the system
of equations
log10 x + 2 log10 (gcd(x, y)) = 60
log10 y + 2 log10 (lcm(x, y)) = 570.
Let m be the number of (not necessarily distinct) prime factors in the prime factorization of x,
and let n be the number of (not necessarily distinct) prime factors in the prime factorization of
y. Find 3m + 2n.
Problem 2.10 (2018 AIME I). [7] For each ordered pair of real numbers (x, y) satisfying
log2 (2x + y) = log4 (x2 + xy + 7y 2 )
there is a real number K such that
log3 (3x + y) = log9 (3x2 + 4xy + Ky 2 ).
Find the product of all possible values of K. Provide a detailed solution, or explain what
progress you made, even if you didn’t get an answer.

10
3 Sequences and Series
Sequences and series are everywhere on the AMC 10/12 and AIME — this is no exag-
geration. This handout will cover the various types of sequences and series you are likely to
encounter.

§3.1 Arithmetic and Geometric Sequences and Series

Before we get into series, we’ll define arithmetic and geometric sequences.

Definition 3.1. A arithmetic sequence is a sequence an where consecutive terms differ by a

constant value, known as the common difference. For example, an = 2n + 3 is an arithmetic
sequence with a0 = 3 and common difference 2.

Definition 3.2. A geometric sequence is a sequence gn where consecutive terms have a con-
stant ratio, known as the common ratio. For example, gn = 5 · 3n is a geometric sequence with
g0 = 5 and common ratio 3.

Now let’s get into arithmetic series; recall that an arithmetic series is a sum of any number
of consecutive terms in an arithmetic sequence.
Here’s a simple example that illustrates a general process that can be used to sum arithmetic
series:

Example. Sum the series 1 + 2 + 3 + · · · + 100.

Solution. The solution most commonly attributed to a young Gauss proceeds as follows: pair
up 1 and 100; 2 and 99; 3 and 98; and similarly pair up the rest of the terms until 50 and 51.
Then each pair of terms sums to 101, and there are 50 pairs in total. (Why?) Thus it follows
that the total sum is 50 · 101 = 5050. 

Remark. Pretty much all arithmetic series can be summed using the same manner — while
there is a formula, it’s quite easy to just derive on the spot, so we wouldn’t recommend
memorizing it. Pair up terms, find how many pairs there are, and multiply!

Example. Find the sum of the series 1 + 3 + 5 + · · · + (2n − 1) in terms of n.

Solution. Let’s try our pairing-up method again: 1 goes with 2n − 1; 3 goes with 2n − 3; and
so on — all looks the same as before, right?
There’s one small hurdle we might run into: all our pairs have an average of n and sum to
2n — they’re kind of “centered around n”, in a way. But what if n itself is odd? Then n − 2
and n + 2 would make a pair, but n itself wouldn’t have another n to pair with.
So we take cases. First, suppose that n is indeed odd. There are n terms in total in the series
(why?) so there must be n−12 pairs, with one middle element n alone. Each pair sums to 2n, so
n−1
our total is 2 · 2n + n = n2 . Thus the sum comes out to be n2 if n is odd.

1
Everaise Academy (2020) Math Competitions I

What if n is even? Then the two terms nearest n are n − 1 and n + 1, which pair up very
nicely and sum to 2n, just like 1 and 2n − 1, 3 and 2n − 3, and all the other terms. In this
case, there are n terms and therefore n2 pairs of terms, each pair summing to 2n, so we get
n 2 2
2 · 2n = n as the sum of our series. It turns out the the sum is n whether n is even or odd —
how nice! 

Before we move on, we’re going to quickly introduce some helpful notation:
P Pn
Definition 3.3. The symbol denotes a sum. Specifically, i=1 f (i), where P
f (i) is any
function of i, is just telling us to add f (i) for i from 1 all the way up to n — that is, ni=1 f (i) =
f (1) + f (2) + · · · + f (n).
For the first example we provided in the section, we might have
100
X
i = 1 + 2 + · · · + 100.
i=1

As for the second, it would be

n
X
(2i − 1) = [2(1) − 1] + [2(2) − 1] + [2(3) − 1] + · · · + [2(n) − 1]
i=1
= 1 + 3 + 5 + · · · + (2n − 1).

For a tougher example, if we wanted 1 − 4 + 9 − 16 + 25 − 36 + · · · − 10000, we could write

this as
100
X
(−1)i+1 i2 .
i=1

We’ve introduced the −1 power to handle the alternating nature of the sum, and the i2 is used
to generate squares. (Try to figure out why this works on your own — plug in i = 1, 2, and 3,
and convince yourself that these are equivalent representations of the sum!)
We can also write sums to infinity: we know that 1 + 21 + 14 + · · · = 2, where successive terms
are generated by dividing previous terms by 2. But this can equivalently be written as
∞  i  0  1  2
X 1 1 1 1
= + + + · · · = 2.
2 2 2 2
i=0

The variable i is called a dummy variable in that it doesn’t actually matter in the end —
P100 P100
i=1 i = k=1 k = 1 + 2 + 3 + · · · + 100, for example. The most common dummy variables
that we use are i, j, and k for sums.
It turns out that we can write products in much the same way, although we won’t consider
that in this handout: 10
Q
i=1 i = 10!, for example.
Pn
Exercise 3.4. Derive a general formula for the sum of the series j=1 j = 1 + 2 + 3 + · · · + n.
(Yes, you’ve probably seen this before, but give it another try!)
Now we’ll move onto a somewhat harder task: summing geometric series.
P∞ i,
Example. Derive a formula for the sum of the infinite geometric series i=0 ar with
|r| < 1. (Here, we refer to r as the common ratio for the series.)

2
Everaise Academy (2020) Math Competitions I
P∞ i
Solution. Unpacking the notation, we see that i=0 ar = ar0 + ar1 + ar2 + · · · . Let’s let
S = a + ar + ar2 + ar3 + · · · . We can see that

rS = ar + ar2 + ar3 + ar4 + · · · .

Great! It turns out that S and rS actually look a lot like each other. Indeed, we actually
get S = a + rS — there’s an entire copy of rS embedded in S. Now, with this new equation
S = a + rS, moving all the S terms to one side and dividing yields
a
S=
1−r
and so we’ve derived a formula for S, the sum of the series. 

A few things to note: there’s nothing special, actually, about our a and r here. Indeed, for any
a
starting value a and common ratio r with |r| < 1, it’s true tha the sum a + ar + ar2 + · · · = 1−r .
We haven’t actually mentioned why |r| < 1 is necessary. Suppose that r = 2, and a = 1, for
example. Then our series becomes

S = 1 + 2 + 4 + 8 + ···
1
which we might sum as S = 1−2 = −1. That doesn’t look right — all the terms are positive!

Remark. An infinite geometric series is said to diverge if the common ratio r satisfies
|r| ≥ 1. That means that when you try to sum the series, it goes out to infinity or negative
infinity — it doesn’t approach some finite number. In this case, you cannot use the
formula we’ve learned to sum the series.

Let’s put everything together one last time.

Theorem 3.5 (Sum of Geometric Series). The sum ∞ i 2

P
i=0 ar = a + ar + ar + · · · is equal
to
a
1−r
where a is the first term of the series, and r is the common ratio of the series.

Remark. This is just a formula for the sum of an infinite geometric series — if you have to
deal with finite geometric series, the formula is a bit uglier. To that end, finite geometric
series can be summed using the same method as infinite geometric series, though: let S be
the sum of the finite series, multiply the entire thing by the common ratio (let’s call it r),
and use the fact that S and rS are super similar. For example,

S = 3 + 9 + 27 + · · · + 729

gives us
3S = 9 + 27 + 81 + · · · + 2187
so therefore, S = 3 + (3S) − 2187. From there you can find S = 1092.
In general, the sum of the first n terms of a geometric series is

a(1 − rn )
1−r

3
Everaise Academy (2020) Math Competitions I

where a is the first term and r is the common ratio. Try proving this yourself, using the
method we just outlined.

Exercise 3.6. Argue that 0.9 = 0.999 . . . = 1 with Theorem 3.5.

Great! Let’s move onto some harder examples, from competitions.

Example (2016 AIME I). For −1 < r < 1, let S(r) denote the sum of the geometric series

12 + 12r + 12r2 + 12r3 + . . . .

Let a between −1 and 1 satisfy S(a)S(−a) = 2016. Find S(a) + S(−a).

Solution. We’re trying to deal with the condition S(a)S(−a) = 2016, so our first step should
probably be to address S(a) and S(−a). We have, by the infinite geometric series formula,
12
S(a) = 12 + 12a + 12a2 + 12a3 · · · =
1−a
2 3 12
S(−a) = 12 − 12a + 12a − 12a + · · · =
1+a
as 12 is the first term of both series, and the first series has common ratio a and the second has
common ratio −a.
Great. So now we have
12 12 144
· = = 2016.
1−a 1+a (1 − a)(1 + a)

One thing we’ll continuously emphasize is that before rushing forward and solving equations
like the above one, it’s best to first check what we’re looking for — that is, S(a) + S(−a). Using
the previously-derived formulas, we have
12 12
S(a) + S(−a) = +
1−a 1+a
12(1 + a) + 12(1 − a)
=
(1 + a)(1 − a)
24
= .
(1 + a)(1 − a)

144
Nice! It turns out that we already have (1−a)(1+a) = 2016, so all we have to do is divide by
6, giving us
24
= 336
(1 − a)(1 + a)
so our answer is 336. 

A few takeaways: remember firstly that infinite series can be summed really nicely, whether
their common ratios are positive or negative! And as discussed before in Section 1.1, remember
to check what you’re looking for before blindly running ahead and finding the values of variables
— sometimes it’s easier without explicitly finding them, such as in this problem where we didn’t
need to actually solve for a.

4
Everaise Academy (2020) Math Competitions I

Example (2012 AIME I). The terms of an arithmetic sequence add to 715. The first term
of the sequence is increased by 1, the second term is increased by 3, the third term is
increased by 5, and in general, the kth term is increased by the kth odd positive integer.
The terms of the new sequence add to 836. Find the sum of the first, last, and middle
terms of the original sequence.

Solution. We’ve got a bunch of words here — the first thing we should do is translate them
into math. Suppose that our arithmetic sequence is a1 , a2 , a3 , · · · , an ; keep in mind that we still
don’t know how many terms there are! Then

a1 + a2 + a3 + · · · + an = 715
(a1 + 1) + (a2 + 3) + (a3 + 5) + · · · + (an + (2n − 1)) = 836

where we’ve noticed that each term of the new sequence is of the form ak + (2k − 1). (Why?)
Note that in our second equation, there’s actually an a1 + a2 + a3 + · · · + an — it’s just hidden!
So we get

836 = (a1 + a2 + a3 + · · · + an ) + (1 + 3 + 5 + · · · + (2n − 1))

= 715 + (1 + 3 + 5 + · · · + (2n − 1).

Recall from an earlier example that the sum 1 + 3 + 5 + · · · + (2n − 1) is just n2 . Therefore
836 = 715 + n2 , and n = 11.
Now we’re looking for the sum of the first, last, and middle terms of the original sequence.
This means that we want to find a1 + a6 + a11 . Note that as a1 , a2 , a3 , · · · , a11 is an arithmetic
sequence, it’s “centered” around a6 — that is, a5 and a7 ’s average is a6 , a4 and a8 ’s average is
a6 , and so on. (Why? Think about the fact that if the common difference is d, the terms before
and after a in a sequence are a − d and a + d.)
So thus we know that a1 + a11 = 2a6 ; a2 + a10 = 2a6 ; and reducing all the terms this way, we
find the equation
11a6 = 715.
3
Meanwhile, we’re just looking for a1 + a6 + a11 = 3a11 . Thus our final answer is 11 · 715 = 195,
and we’re done. 

Example 3.7 (2020 AMC 10). There exists a unique strictly increasing sequence of non-
negative integers a1 < a2 < . . . < ak such that

2289 + 1
= 2a1 + 2a2 + . . . + 2ak .
217 + 1
What is k?

Solution. At first glance, this seems like a pretty tough problem; 2289 + 1 and 217 + 1 are
massive numbers and essentially impossible to prime factorize, so we can’t do the brute force
approach of finding (2289 + 1)/(217 + 1) then computing its binary representation. However,
the idea of representing the expression as a sum of powers of 2 brings to mind geometric series.
Can we find a connection? Noting that 289 = 172 , the expression
2
217 + 1
217 + 1

5
Everaise Academy (2020) Math Competitions I

looks suspiciously like the sum of a geometric series with first term 1, if we let r = −217 .
Converting the expression to the form of the geometric series formula gives

1 − (−217 )17 1 − r17

= .
1 − (−217 ) 1−r

Then we know this is the sum of the first 17 − 1 = 16 terms of a geometric sequence with first
term 1, so this is equal to

1 + r + r2 + . . . r16 = 20 − 217 + 234 − · · · + 2272 .

This isn’t quite what we were looking for, since we have some negative powers of 2. How can
we fix this? The simplest option seems to be adding every negative power of 2 to the positive
power of 2 which follows it:

20 + (−217 + 234 ) + (−251 + 268 ) − · · · + (−2255 + 2272 ).

The reason this is promising is because every negative-positive pair differs from the next pair
by a factor of 234 , so the binary representations will have 1s in different positions. This means
we’ll get distinct powers of 2 from each pair, so we can put them in strictly increasing order
just like the problem asks. Note that each pair factors as 217k (−1 + 217 ) for some integer k.
Here we can use the fact that

−1 + 2n = 1 + 2 + 22 + · · · + 2n−1 .

This is a pretty common trick that helps to memorize and is easy to derive using geometric
series. Using this fact, each pair can be turned into a sum of powers of two:

217k (−1 + 21 7) = 217k (20 + 2 + 22 + · · · + 216 ) = 217k + 217k+1 + 217k+2 + · · · + 217k+16 .

It’s clear that all the powers of 2 will be distinct since pairs differ by a factor of 234 . Then since
there are 8 pairs with 17 powers of 2 each along with the term 20 at the beginning, there are a
total of 17 × 8 + 1 = 137 powers of two. 

It turns out that most arithmetic and geometric sequences and series problems on math
competitions end up being a bit formulaic. In the next section, we’ll go over some different
types sequences and series — ones that aren’t simply arithmetic and geometric.
Exercise 3.8 (2012 AIME II). Two geometric sequences a1 , a2 , a3 , . . . and b1 , b2 , b3 . . . have the
same common ratio, with a1 = 27, b1 = 99, and a15 = b11 . Find a9 .

§3.2 Other Sequences and Series

While arithmetic and geometric series are the most common types of sums, there are many
other series problems that don’t fit well in either of these two categories. One of these is a mix
of the two, and it is creatively named the “arithmetico-geometric series.”

∞
X i 1 2 3 4
Example. Compute the infinite sum i
= + + + + ···.
3 3 9 27 81
i=1

6
Everaise Academy (2020) Math Competitions I

Solution. It turns out that using the S and rS trick used to derive Theorem 3.5, with r = 3,
and then applying the same theorem directly on the resulting sum solves the problem (try it
out for yourself!). However, we provide an alternate approach for the sake of variety.
Let’s try to view this problem from a different perspective. Right now, the infinite sum is
expressed in such a way so that we don’t have a formula to arrive at an immediate answer.
What if we break it down into different chunks that are easier to work with? We know, for
example, how to quickly find 31 + 19 + 27
1
+ · · · , which looks suspiciously similar to the problem
we have at hand. If we continue down this train of thought, we get the following:

1 1 1 1
S= + + + + ···
3 9 27 81
1 1 1
+ + + + ···
9 27 81
1 1
+ + + ···
27 81

If we sum the expression above vertically, we obtain the expression that we originally started
with. However, if we sum it horizontally using Theorem 3.5, we get S = 12 + 61 + 18 1
+ ···.
3
Applying the theorem once again gives us a final value of 4 . 

We solved this problem by quite literally changing perspectives. This is a good strategy to
use whenever you’re dealing with something that looks suspiciously familiar, but not quite right
— turn the problem into something that you do know how to work with.
∞
X 2j − 1 1 3 5 7
Exercise 3.9. Sum the series = + + + + ···.
2j 2 4 8 16
j=1

Another common type of infinite series is the telescoping series, which is when a lot of the
terms in the sequence cancel out rather nicely, leaving one clean expression at the end. We
use the word “telescope” to describe such a series, since each next term collapses the series,
similarly to how a long, retractable telescope can collapse to a size no bigger than a pencil. This
description may be rather confusing, so it’s best to illustrate the concept with an example:

∞
X 1 1 1 1 1
Example. Find = + + + + ···
i · (i + 1) 1·2 2·3 3·4 4·5
i=1

Solution. Note that

1 n+1 n 1 1
= − = −
n(n + 1) n(n + 1) n(n + 1) n n+1

This is great, because our original problem can now be written as the following:
∞          
X 1 1 1 1 1 1 1 1 1 1
− = − + − + − + − ···
i i+1 1 2 2 3 3 4 4 5
i=1

As you can see, each next term collapses on itself, leaving a rather nice looking sum! Taking
this expression as it approaches infinity, it’s easy to see that the sum itself approaches 1. 

Let’s take a look at the motivation behind this problem — we mentioned before that in general,
telescoping sums collapse after each term. To see what this means, let’s let Sn = ni=1 i·(i+1)
1
P
.

7
Everaise Academy (2020) Math Competitions I

Trying out a few values by hand, we see that S1 = 12 , S2 = 23 , and S3 = 43 . From here, it’s
1
reasonable to conclude that Sk = 1 − k+1 , but can we prove why this must be the case?
We’ve calculated already that S1 = 1 − 12 , S2 = 1 − 13 , and S3 = 1 − 41 . These Sn ’s have
another nice property in general, though: they share a lot of common terms. That is, if we have
1
Sn−1 and Sn , all their terms up to the last n·(n+1) will be the same. When we have a situation
like this it’s often nice to subtract Sn−1 from Sn , then — recall our proof of the geometric series
formula, in which we ended up with a bunch of common terms and subtracted.
1
So Sn − Sn−1 = n·(n+1) , from definition. But now let’s look back at our smaller cases:
S2 − S1 = 2 − 3 ; S3 − S2 = 13 − 14 . Miraculously enough, our equation at the beginning
1 1
1 1
of this paragraph yields S2 − S1 = 2·3 , and S3 − S2 = 3·4 . At this point, the pattern that
1 1 1
n(n+1) = n − n+1 should be quite visible — and we can proceed as in the solution.

P
Remark. Sometimes, it’s not easy to determine how exactly a series ai telescopes. If you
ever get stuck, try to relate each ai to a different sequence, bi , such that ai = bi+1 − bi . In
1
this particular example, ai = i·(i+1) and bi = − 1i . With this in mind, the sum of the first
n terms of the sequence of ai ’s will simply evaluate to bn+1 − b1 .

Let’s look at a harder example; here, the telescoping substitution will be trickier to spot.

Example (2018 Math Prize for Girls). Consider the sum

n
X 1
Sn = √ .
k=1
2k − 1

Determine bS4901 c. Recall that bxc is the greatest integer less than or equal to x.

Solution. Accurately summing square roots is next to impossible, so let’s see if we can bound
the expression at hand in between two values. This usually requires a telescoping approach, so
we’ll go with that route here. To account for edge cases that will arise later, let’s define
n
X 1
Tn = √
k=2
2k − 1

so that 1 + Tn = Sn . Now, notice that

√ p √ p √
2k − 1 + 2(k − 1) − 1 < 2 2k − 1 < 2(k + 1) − 1 + 2k − 1.

Taking the reciprocals and rationalizing the denominators will leave us with
√ p p √
2k − 1 − 2(k − 1) − 1 1 2(k + 1) − 1 − 2k − 1
> √ > .
2 2 2k − 1 2

So in reality, we have that

n p n 
X √  X √ p 
2(k + 1) − 1 − 2k − 1 < Tn < 2k − 1 − 2(k − 1) − 1 .
k=2 k=2

This sum telescopes — if we set n = 4901, we see that a lot of the summands in both the
lower and upper bounds of the inequalities cancel, leaving us with just
√ √ √ √
9803 − 3 < T4901 < 9801 − 1

8
Everaise Academy (2020) Math Competitions I
√ √
The upper bound is exactly 98, and, using crude estimates of 3 ≈ 1.71 and 9803 ≈ 99, the
lower bound is a little bit more than 97. Thus, bT4901 c = 97, and so bS4901 c = 98. 

The main idea behind solving this problem is bounding this sum between two easier-to-find
integers, as we’re only asked to find the floor of this sum. To do so, we’ll probably want the lower
and upper bounds to be in the form of a telescoping series involving square roots. Referring
back to the remark made after the previous problem, our telescoping series should most likely
√ √
be one whose summands are in the form n + 2 − n, as we’re only summing the reciprocals
of the square roots of every other number in our original expression. It turns out that these key
insights are all that we needed, and the rest of the solution comes quite naturally!

§3.3 Bashing Sequences

There are some functions that just feel bashable. You’ll know which these are through intuition
and experience. They’re either repeating, or their pattern becomes clear after you compute
three or four terms. All you have to do is plow through.
One type of sequence problems should be bashed directly with the numbers given; the other
type should be bashed in terms of variables. The key (and often only) observation is figuring
out which is which.

Example 3.10 (2019 AMC 10A). A sequence of numbers is defined recursively by a1 = 1,

a2 = 73 , and
an−2 · an−1
an =
2an−2 − an−1
for all n ≥ 3. Then a2019 can be written as pq , where p and q are relatively prime positive
integers. What is p + q?

3 3
Solution. We calculate a few terms to get a feel for the problem. Note that a3 = 11 , a4 = 15 ,
3 3
and a5 = 19 . This leads us to believe an = 4n−1 , and checking that this holds for a1 is enough
3 3
to convince us that a2019 = 4·2019−1 = 8075 , so the answer is 8078.
3
In fact we can verify with induction that an = 4n−1 . The base case a1 is already done, and
3 3
an−1 · an · 4n−5 9 3
an+1 = = 4n−1 3 3 = =
2an−1 − an 2 · 4n−5 − 4n−1 12n + 9 4n + 3
finishes the inductive step. 

This is a problem where you bash directly with the given numbers, because you feel the given
numbers might be special in some way.

The next example contains a periodic sequence. We say a sequence is periodic if its terms
eventually repeat. The period of a periodic sequence is the minimum number of terms be-
fore each repeat. For example, 1, 2, 3, 1, 2, 3, . . . is periodic, and has period 3. The sequence
1, 5, 2, 4, 8, 1, 2, 3, 1, 2, 3, . . . is periodic as well, because it eventually repeats forever, even though
the first few terms did not repeat.
Exercise. Assume we have a sequence an such that each term depends only on the previous two
terms. Show that if a100 = a102 and a101 = a103 , then the sequence is periodic. In general, if we
have a sequence where each successive term depends on the previous k terms, how many terms
need to repeat before we can confirm the sequence is periodic?

9
Everaise Academy (2020) Math Competitions I

Example 3.11 (2020 AIME II). Define a sequence recursively by t1 = 20, t2 = 21, and
5tn−1 + 1
tn =
25tn−2

for all n ≥ 3. Then t2020 can be written as pq , where p and q are relatively prime positive
integers. Find p + q.

Solution. Playing with this sequence for a little bit doesn’t reveal much about it, so we might
assume there’s a pattern we have to bash to find. We’ll write everything in terms of t1 and t2 .
Note
5t2 + 1
t3 =
25t1
5t2 +1
5· 25t1 +1 5t1 + 5t2 + 1
t4 = =
25t2 125t1 t2
5t1 +5t2 +1
5· 125t1 t2 + 1 5t1 + 5t2 + 1 + 25t1 t2 (5t1 + 1)(5t2 + 1) 5t1 + 1
t5 = = = =
25 · 5t25t
2 +1
1
2
125t2 + 25t2 25t2 (5t2 + 1) 25t2
5t1 +1
5· 25t2 + 1 5t21 + 5t1 t2 + t1
t6 = 5t1 +5t2 +1 = = t1 .
25 · 125t1 t2 5t1 + 5t2 + 1
5·20+1 101
So we see that this sequence repeats with a period of 5. Thus t2020 = t5 = 5·21 = 125 , so the
answer is 226. 

This is a problem where you bash with the given variables because you feel that the recursion
should have some fixed period no matter the starting values.

10
Everaise Academy (2020) Math Competitions I

§3.4 Problems
Problem 3.1 (2015 AMC 8). [4] An arithmetic sequence is a sequence in which each term after
the first is obtained by adding a constant to the previous term. For example, 2, 5, 8, 11, 14 is an
arithmetic sequence with five terms, in which the first term is 2 and the constant added is 3.
Each row and each column in this 5 × 5 array is an arithmetic sequence with five terms. What
is the value of X?

Problem 3.2 (1969 AHSME). Let Sn and Tn be the respective sums of the first n terms of
two arithmetic series. If Sn : Tn = (7n + 1) : (4n + 27) for all n, the ratio of the eleventh term
of the first series to the eleventh term of the second series can be expressed as a : b where a and
b are relatively prime positive integers. What is a + b?

Problem 3.3 (2016 AMC 10B). [6] The sum of an infinite geometric series is a positive number
S, and the second term in the series is 1. What is the smallest possible value of S?

Problem 3.4 (2011 AIME II). [6] The sum of the first 2011 terms of a geometric sequence is
200. The sum of the first 4022 terms is 380. Find the sum of the first 6033 terms.

Problem 3.5 (2002 AMC 12A). [7] Consider the sequence of numbers: 4, 7, 1, 8, 9, 7, 6, . . . For
n > 2, the n-th term of the sequence is the units digit of the sum of the two previous terms.
Let Sn denote the sum of the first n terms of this sequence. What is the smallest value of n for
which Sn > 10, 000?

Problem 3.6 (2002 AIME II). [7] Two distinct, real, infinite geometric series each have a sum
of 1 and have the same second term. The third √ term of one of the series is 1/8, and the second
term of both series can be written in the form m−np , where m, n, and p are positive integers
and m is not divisible by the square of any prime. Find 100m + 10n + p.

Problem 3.7 (1989 AIME). [7] If the integer k is added to each of the numbers 36, 300, and
596, one obtains the squares of three consecutive terms of an arithmetic series. Find k.

Problem 3.8 (2001 AMC 12). [8] Consider sequences of positive real numbers of the form
x, 2000, y, . . . in which every term after the first is 1 less than the product of its two immediate
neighbors. For how many different values of x does the term 2001 appear somewhere in the
sequence?

Problem 3.9 (USAMTS). [10] Let

r r r
1 1 1 1 1 1
A = 1 + 2 + 2 + 1 + 2 + 2 + ··· + 1 + 2
+ .
1 2 2 3 2019 20202
Then we can write A/2019 as a fraction pq , where p and q are integers, in lowest terms. Find
p + q.

11
Everaise Academy (2020) Math Competitions I

Problem 3.10 (2005 AIME II). [8] An infinite geometric series has sum 2005. A new series,
obtained by squaring each term of the original series, has 10 times the sum of the original series.
What is the common ratio of the original series?
Provide a detailed solution, or explain what progress you made, even if you didn’t
get an answer.

12
4 Polynomials

§4.1 General form

Polynomials appear very often in math competitions, and are a major topic in algebra. Let’s
start with the definition of a polynomial:

Definition 4.1. For constants a0 , a1 , . . . an , an 6= 0, and nonnegative integer n, the expression

f (x) = a0 + a1 x + a2 x2 + . . . an xn

is a polynomial.

Example 4.2. Which of the following are polynomials?

1. 5
1
2. x + x7 + 3
3. x−2 + x2
4. x2 + 124
√
5. x3 + 3x + 2
123456789
6. x12
+ 13
2
7. x + 6x3

Solution. What makes something not a polynomial? The condition that strikes the eye is the
nonnegative integer part. For an expression to be a polynomial, all of the exponents of the x
terms must be 0 or greater. So all the expressions except 3, 6, and 7 are polynomials.


Now we’ll define some important terms we use to describe a polynomial. As we said before,
we can express a polynomial as

f (x) = a0 + a1 x + a2 x2 + . . . an xn

for nonnegative integer n and constants a0 , a1 , . . . an with an nonzero.

The degree of a polynomial is n, the highest power of x that appears in the polynomial. A
coefficient is any of the a0 , a1 , . . . an , and we call an the leading coefficient of f (x). A root
of f (x) is some constant r such that f (r) = 0.

An integer polynomial is a polynomial where all the coefficients are integers, and this can
be similarly defined for polynomials with coefficients that are rational, real, etc.

Constants denote polynomials of degree 0, linear terms denote polynomials of degree 1

(e.g. 3x + 4), quadratic terms denote polynomials of degree 2 (e.g. 3x2 + 2x + 5), cubic terms
denote polynomials of degree 3, and quartic terms denote polynomials of degree 4.

1
Everaise Academy (2020) Math Competitions I

We can think of polynomials as functions of x. The variable x does not necessarily need to
come from the same set as the polynomial’s coefficients; for example, we can plug in x = i into
the polynomial f (x) = x2 + 2 to get f (i) = 1.

However, polynomials are interesting in their own right, not just as functions. Polynomials
have their own arithmetic, which we’ll discuss now. For the sake of clarity, we will define
P (X) = p0 + p1 x + . . . pn xn and Q(x) = q0 + q1 x + . . . qm xm .

Remark. The reason for the capital X is that we’re referring to polynomials as objects
rather than functions of x when we define arithmetic on them, though the distinction isn’t
that important for our purposes.

Definition 4.3. Polynomial Addition Without loss of generality, assume n ≥ m. Then,

P (X) + Q(X) = (p0 + q0 ) + (p1 + q1 )x + . . . (pm + qm )xm + . . . pn xn

with subtraction being defined similarly. Next, we define multiplication:

Definition 4.4. Polynomial Multiplication Without loss of generality, assume n ≥ m.

Then,

P (X)Q(X) = p0 q0 + (p1 q0 + p0 q1 )x + (p2 q0 + p1 q1 + p0 q2 )x2 + . . . pn qm xm+n

In regular integer arithmetic, division has quotients and remainders. For example, if we divide
7 by 3, we get a quotient of 2 and remainder 1, since 7 = 3 ∗ 2 + 1. Putting in this a polynomial
perspective, suppose by we divide F (X) by P (X).

Definition 4.5. Polynomial division If F (X) is a polynomial being divided by P (X), we

express this as
F (X) = P (X)Q(X) + R(X)
where deg(R(X)) ≤ deg(P (X)) − 1.

In this case, Q(X) represents the quotient while R(X) represents the remainder. The re-
quirement that deg(R(X)) ≤ deg(P (X)) − 1 ensures that our remainder is “smaller” than our
divisor. For example, if you divide a cubic by a quadratic, the remainder must either be linear
or a constant.

§4.1.1 Polynomial Long Division

Polynomial long division is similar to regular long division, except instead of tens places, we
have powers of x. It is an important tool for factoring polynomials; if you know that x − r is a
factor of a polynomial f (x), then you can use polynomial division to find q(x) = f (x)/(x − r)
and factor f (x) into (x − r)q(x).
If you’re not familiar with polynomial long division, we recommend that you learn this using
other online resources, as it should be fairly easy. However, we’ll show a few examples here as
a refresher.

2
Everaise Academy (2020) Math Competitions I

Example. Find the quotient and remainder when x4 + 4x2 + 5x + 3 is divided by x2 + 1.

Using polynomial long division, we get

x2 +3
x2 x4 4x2


+1 + + 5x + 3
− x4 −x 2

3x2 + 5x + 3
− 3x2 −3
5x

So our quotient is x2 + 3 and our remainder is 5x.

Example. Find the remainder when x4 + 5x3 + x2 + 8x + 11 is divided by x2 + 2x + 1.

Solution. Notice that x2 + 2x + 1 = (x + 1)2 . Then, we can first divide by x + 1, then divide
the result by x + 1.

x3 + 4x2 − 3x + 11
x4 + 5x3 + x2 + 8x + 11


x+1
− x4 − x3
4x3 + x2
− 4x3 − 4x2
− 3x2 + 8x
3x2 + 3x
11x + 11
− 11x − 11
0

Note that we first multiplied by x3 . Then, we can divide again by x + 1 to get our answer:

x2 + 4x − 7
x3 + 5x2 − 3x + 11


x+1
− x3 − x2
4x2 − 3x
− 4x2 − 4x
− 7x + 11
7x + 7
18
Hence, our remainder is 18. 

The remainder theorem is an important theorem pertaining to polynomial division.

Theorem 4.6 (Remainder Theorem). Assume we have a polynomial F (X) being divided by

3
Everaise Academy (2020) Math Competitions I

a linear polynomial P (X) = ax + b. If R(X) is the remainder polynomial as a result of

b


this division, then we have R(X) = F − a .

Proof. We first start with our expression for polynomial division, and let P (X) = ax + b. Note
that this makes R(X) a constant as it must have degree less than 1. Then, we have that

F (X) = (ax + b)Q(X) + R(X).

Clearly, since a(− ab ) + b = −b + b = 0, if we plug in X = − ab then we have that

   
b b
F − =0+R −
a a
   
b b
F − =R − .
a a

But since R(X) is constant, R(X) + R(− ab ), so

 
b
F − = R(X)
a

and we are done. 

Though the remainder theorem in itself is mostly a trick to save time, we can apply the funda-
mental idea of the remainder theorem— that plugging in roots simplifies division expressions
greatly— to solve problems.

Example 4.7 (BMC). Find the remainder when x81 + x49 + x25 + x9 + x is divided by x3 − x.

Solution. We first make our polynomial division expression:

x81 + x49 + x25 + x9 + x = (x3 − x)Q(x) + R(x)

Since we are dividing by x3 − x, this means that deg(R(X)) is at most 2. Hence, we can let
R(X) = ax2 + bx + c for some constants a, b, c.

Now, although we haven’t covered roots yet, the next part is fairly intuitive. We need three
values of R(X) as we have 3 variables to solve for. However, if we were to say choose some
arbitrary x = 2, then we have the huge equation

281 + 249 + 225 + 29 + 2 = (23 − 2)Q(2) + R(2).

There are two major problems: the numbers are huge and we actually have no idea what Q(x)
is, let alone Q(2).

However, we know Q(x) is being multiplied by x3 − x. Thus, we can at least eliminate one
major problem by setting x3 − x = 0 so we don’t need to deal with Q(x). Hence, we solve for
the different values of x :

x3 − x = 0
x(x2 − 1) = 0
x(x − 1)(x + 1) = 0

4
Everaise Academy (2020) Math Competitions I

The values of x that satisfy this equation are 0, 1, −1 which is also very convenient because this
eliminates our other major problem: huge numbers. Plugging in these three values into the
expression, we get the following system of equations:

R(1) = a + b + c = 5
R(0) = c = 0
R(−1) = a − b + c = −5

Solving this systems of equations gives a = 0, b = 5, c = 0. Hence, our remainder is 5x. 

§4.2 Roots
Perhaps one of the most important concepts of polynomials are their roots. Why should we
care about when a polynomial is equal to 0? Well, the most obvious point comes directly from
the zero product property.

Theorem 4.8 (Factor Theorem). A number r is a root of the polynomial P (x) if and only
if x − r is a factor of P (x).

Proof. First, we’ll prove the forwards direction: if r is a root of P (x), then P (r) = 0. Assume
the degree of P is n. Write P (x) as

P (x) = a(x − r1 )(x − r2 ) · · · (x − rn )

so that r1 , r2 , · · · , rn are the roots of P . Then, as P (r) = 0, we know that a(r−r1 ) · · · (r−rn ) = 0,
so at least one of r − r1 , · · · , r − rn must be equal to 0 by the zero product property; therefore
r = ri for some i between 1 and n. It follows indeed that x − r is a factor of P (x), as desired,
as all the x − ri ’s are factors of P (x).
Now, if x − r is a factor of P (x), then P (x) = (x − r)Q(x) for some other polynomial Q(x).
It follows that P (r) = (r − r)Q(r) = 0, so therefore f is a root of P (x), as desired. 

There is another fact that makes roots important.

Theorem 4.9 (Fundamental Theorem of Algebra). A polynomial P (x) of degree n has n

roots in total, real and complex.

This allows us to express any polynomial P as

P (x) = an (x − r1 )(x − r2 ) · · · (x − rn )

where r1 , · · · , rn are the roots of P , and an is its leading coefficient.

Exercise. Prove the above statement using the fundamental theorem of algebra and the factor
theorem. (If you want to do this formally, you might have to use induction!)
Note that roots can appear more than once in the factorization of a polynomial. For example,
x2 −2x+1 = (x−1)2 , which has two factors of x−1. To deal with this, we define multiplicity. A
root r has multiplicity k if (x−r) appears exactly k times in the factorization of the polynomial.
When we say a polynomial has roots r1 , r2 , . . . rn , they are not assumed to be distinct; if a root
has multiplicity k, then it appears k times in r1 , r2 , . . . rn .

5
Everaise Academy (2020) Math Competitions I

§4.2.1 Vieta’s

We know how to factor polynomials with rational roots, like quadratics. Whenever we can
directly find roots, we should do so. Unfortunately, we will often be given problems where
directly finding roots is impossible. Thankfully, it turns out that we can still figure out a lot
about the roots of a polynomial without knowing what they are.

Specifically, the fact that we can break down a polynomial into its linear factors (x − ri ) gives
rise to a very important series of formulas known as Vieta’s formulas. Before we get into Vieta’s
formulas, though, we first have to define symmetric sums.

Definition 4.10. Define the kth elementary symmetric sum of n variables a1 , a2 , · · · , an

as the sum of all combinations of products of k of these variables at a time. We’ll denote this
as Sk . (The original number of variables will be clear from context.)
Of course, it’s hard to understand this without some examples. The first elementary sym-
metric sum of any n variables a1 , · · · , an , denoted S1 , is just

S1 = a1 + a2 + · · · + an .

The second, S2 , is
S2 = a1 a2 + a1 a3 + · · · + an−1 an ,
where S2 is the sum of all products of all pairs of elements.
Here a few specific cases: if there are five variables a, b, c, d, and e, then

S4 = abcd + abce + abde + acde + bcde.

If there are four variables a, b, c, and d, then

S2 = ab + ac + ad + bc + bd + cd

and
S4 = abcd.

Theorem 4.11 (Vieta’s Formulas). Given a polynomial P (x) = an xn + an−1 xn−1 + · · · +

a1 x + a0 with roots r1 , r2 , · · · , rn , Vieta’s Formulas state that
an−k
Sk = (−1)k · ,
an
where Sk denotes the kth elementary symmetric sum of the roots r1 , · · · , rn .

Proof. We know that we can express P (x) as

P (x) = a(x − r1 )(x − r2 ) · · · (x − rn )

for some constant a. In this case, it is clear that a = an .

Expanding this yields

P (x) = an xn + an (−r1 − r2 − · · · − rn )xn−1

+ an (r1 r2 + r1 r3 + · · · + rn−1 rn )xn−2
+ . . . + (−r1 )(−r2 ) · · · (−rn ).

6
Everaise Academy (2020) Math Competitions I

Look at these coefficients. Miraculously, they’re the elementary symmetric sums of the n
roots. Additionally, depending on which coefficient we’re dealing with, we have a power of −1
to deal with — and as such, by comparing coefficients,
an (Sk )(−1)k = an−k .

This rearranges to the desired equation. 

Now let’s try some examples.

Example 4.12. For the polynomial x4 + 5x2 + 8x + 9, find the sum and product of all of
its roots.

Solution. We first solve for the sum of all of the roots. We know that a4 = 1, a3 = 0, a2 = 5,
a1 = 8, and a0 = 9. From Vieta’s, we have that
a3 0
r1 + r2 + r3 + r4 = − = − = 0,
a4 1
so the sum of all of the roots is 0. Similarly, the product of all of the roots is
9
r1 r2 r3 r4 = (−1)4 · = 9.
1
Thus, the product of all of our roots is 9. 

Example 4.13. If p, q, r, and s are the roots of P (x) = 2x4 − 6x3 − 30x2 + 12x + 10, then
find
1. p + q + r + s,
2. pq + pr + ps + qr + qs + rs,
3. pqr + pqs + prs + qrs, and
4. pqrs.

Solution. We know that p + q + r + s = S1 , the first elementary symmetric sum, so our answer
is
−6
S1 = (−1) · = 3.
2
Next, pq + pr + ps + qr + qs + rs = S2 , so our answer is
−30
S2 = (−1)2 · = −15.
2
Now, pqr + pqs + prs + qrs = S3 , so our answer is
12
S3 = (−1)3 · = −6.
2
Finally, pqrs = S4 , so our answer is
10
S4 = (−1)4 · = 5.
2


It is important to remember that Vieta’s Formulas only apply to polynomials; however, it is

sometimes possible to manipulate equations to turn them into polynomials, as we will see in
the following example:

7
Everaise Academy (2020) Math Competitions I

Example 4.14. (2003 AMC 10A) What is the sum of the reciprocals of the roots of the
equation 2003 1
2004 x + 1 + x = 0?

Solution. We want to find a symmetric expression containing all the roots, so Vieta’s Formula
should come to mind. However, we cannot apply Vieta’s immediately, as 2003 1
2004 x + 1 + x is not
1
a polynomial. We can multiply both sides of the equation by x to get rid of the x , and instead
find the roots of
2003 2
x + x + 1 = 0.
2004
We must ensure x = 0 is not a root (otherwise we would be dividing by zero in the original
equation).
It is easy to see that it is not, so we can apply Vieta’s without worrying about extraneous
roots.
Let the roots of this equation be x1 and x2 . We are asked to find the sum of the reciprocals of
the roots, or
1 1 x1 + x2
+ = .
x1 x2 x1 x2
−1 1
x1 + x2 = 2004 and x1 · x2 = 2004 , so our answer is
2003 2003
 
−1
2004
−1
 2003  = = −1.
1 1
2004
2003

Remark. If we have a polynomial f (x) with degree n (so f (x) has n roots r1 , . . . , rn ) then
a common trick for finding the sum of the reciprocals of the roots is using the fact that
1 1 1 Sn−1
+ + ··· + =− .
r1 r2 rn Sn

Example 4.15. If r, s, t are the three roots of polynomial f (x) = x3 + 5x2 − 8x + 9, find
(r + s + 1)(s + t + 1)(t + r + 1).

Solution. It seems rather tempting to just expand out the terms. After all, we already know
that

r + s + t = −5
rs + st + tr = 8
rst = −9

so we can just group together the terms and plug and chug. And while this certainly is a viable
solution, there exists a much cleaner and efficient way. We know that we can express any cubic
polynomial as
f (x) = an (x − r1 )(x − r2 )(x − r3 ).

8
Everaise Academy (2020) Math Competitions I

Since the leading coefficient of f (x) is 1, we have that an = 1 and we know that our roots are
r, s, t so we can express f (x) as
f (x) = (x − r)(x − s)(x − t).
We know we can substitute any x. Note the symmetry of (r + s + 1)(s + t + 1)(t + r + 1). If we
let x = r + s + t + 1, we then find that
f (r + s + t + 1) = (s + t + 1)(t + r + 1)(r + s + 1)
which is exactly what we need to find! We even know r + s + t from Vieta’s, which is −5, so
we just need to calculate f (−5 + 1) = f (−4). This is the easiest step: we just evaluate our
polynomial with x = −4. Doing so gives us
−64 + 80 + 32 + 9 = 57
as our answer. 

Whenever applying Vieta’s, don’t forget to include the alternating signs, and make sure you
remember to divide by the leading coefficient as not every polynomial given will be monic (i.e.
have leading coefficient 1).

We’ll move on to a problem that’s much harder than the previous examples, but should still
be solveable with our tools.

Example 4.16. (2017 AMC 10A) For certain real numbers a, b, c the polynomial

g(x) = x3 + ax2 + x + 10

has three distinct roots, and each root of g(x) is also a root of the polynomial

f (x) = x4 + x3 + bx2 + 100x + c.

What is f (1)?

Solution. Note that since f (x) and g(x) are both monic and they share 3 roots, we can set up
the following equation
f (x) = (x − r)g(x)
for some real r. Now, we can substitute in g(x) and then compare coefficients with f (x):
f (x) = (x − r)(x3 + ax2 + x + 10)
f (x) = x4 + (a − r)x3 + (1 − ar)x2 + (10 − r)x − 10r

This gives us
x4 + x3 + bx2 + 100x + c = x4 + (a − r)x3 + (1 − ar)x2 + (10 − r)x − 10r.
By matching coefficients, we can set up a system of four equations:
a−r =1
1 − ar = b
10 − r = 100
−10r = c

9
Everaise Academy (2020) Math Competitions I

Before we start solving for any values, let’s first examine the problem statement so we know
exactly what we need to do. The problem says to find f (1), and plugging in 1 gives us 1 + 1 +
b + 100 + c, so we need to find b and c. Let’s get right into it.
From the third equation, we find that r = −90. Thus, we can directly plug this into the
fourth equation to get c = 900.
Now, it remains to find the value of b. From the first equation, we obtain that a−(−90) = 1, so
a = −89. Then, plugging this into the second equation gives 1−(−90)(−89) = b. So, b = −8009.
Finally, we can plug in all the values we got. Our final answer is 1 + 1 − 8009 + 100 + 900
which conveniently simplifies down to −7007.


Now, we’ll introduce another application of Vieta’s formulas, Newton Sums. Newton Sums
are a method of expressing pn = an + bn + cn as a combination of symmetric sums and lower
power sums like pn−1 and pn−2 .

Theorem 4.17 (Newton Sums). Suppose that we are given a polynomial P (x) = ni=0 an xn ,
P

where an 6= 0. Denote the nth symmetric sum by Sn . Given that x1 , x2 , ..., xn are the roots
of P (x), let Pn denote ni=1 xni . Then, for all 1 ≤ m ≤ n,
P

m
X
m
(−1) mSm + (−1)k+m Pk Sm−k = 0.
k=1

The first few indentites are as follows:

1. P1 = S1
2. P2 = S1 P1 − 2S2
3. P3 = S1 P2 − S2 P1 + 3S3
4. P4 = S1 P3 − S2 P2 + S3 P1 − 4S4

You can ignore the scary sum and just focus on the pattern in the identities.
The proof is quite long and tedious, but if you want to prove it on your own, try plugging in
all the roots of P (x), set these expressions equal to 0, then use Vieta’s to express the coefficients
in terms of the roots.
Note that if we only have 3 variables, then S4 is just 0, and we can further ignore any
expressions with S5 and so forth.

The following is a very nice problem that utilizes a small amount of trig. For now, if you
have not learned trig, do not look at the solution and add sin2 x + cos2 x = 1 to your workspace
to further attempt the problem.

11
Example 4.18. (2019 AIME) Let x be a real number such that sin10 x + cos10 x = 36 .
Then, sin12 x + cos12 x = m m
n . where n is in lowest terms. Find m + n.

Solution. To simplify your work, it is very easy to simply let sin2 x = a and cos2 x = b. Thus,
11
we also have that a + b = 1 by the Pythagorean identity and we are given a5 + b5 = 36 . We are
6 6
looking for a + b .
We can now apply Newton sums– you may wonder why, because there’s no polynomial in
this problem, and Newton sums are usually given in that context.

10
Everaise Academy (2020) Math Competitions I

However, if you glance at the identities again, you will see we’re only concerned with sym-
metric sums and power sums. We have our power sums directly from the problem, and since
we only have two variables, our only symmetric sums are s1 = a + b = 1 and s2 = ab. Thus,
we can use the information we get using Newton sums to solve for s2 and then get the value of
p6 = a6 + b6 , which is what we’re looking for.
We know that s1 = p1 = 1. Hence, we have that
p6 = s1 p5 − s2 p4
p5 = s 1 p4 − s 2 p3
p4 = s 1 p3 − s 2 p2
p3 = s 1 p2 − s 2 p1
p2 = s1 p1 − 2s2
Note that we’re just taking the identities for the first six power sums and ignoring the terms
with symmetric sums higher than the second symmetric sum.
So, we want to further simplify p5 ’s expression as we already know its value. To also simplify
the visuals, we also let s2 = c. Note that p2 = 1 − 2c.
11
= p4 − c(s1 p2 − cp1 )
36
11
= (p3 − cp2 ) − c(p2 − c)
36
11
= (p2 − c − cp2 ) − cp2 + c2
36
11
= 5c2 − 5c + 1
36
So, we now have a handy in one variable. Now, we solve for c. If you are unfamiliar with
quadratics, you can skim over the next section, but the solution doesn’t require anything fancy.

25
5c2 − 5c + =0
36
36c2 − 36c + 5 = 0
(6c − 5)(6c − 1) = 0
Thus, we either get c = 56 or c = 16 . If you don’t have a trig background, take our word that
c = 56 is impossible. The reason why is because
sin2 (2x)
c = sin2 x · cos2 x =
4
However, sin2 (2x) < 1, so c < 14 which is why c = 16 . Now armed with this knowledge, we can
7
simply plug and chug. After solving for p4 , which is 18 , we have that p6 = 13
54 .


§4.3 Quadratics
Let’s now examine one of the most commonly seen polynomials: quadratics. While the name
might be a bit misleading as ”quad” generally means 4, a quadratic equation is any polynomial
with degree 2. Formally, it is any f (x) = ax2 + bx + c where a, b, and c are real numbers and
a 6= 0. Quadratics are very nice and a great introduction to solving roots because there exists
a handy formula called the...

11
Everaise Academy (2020) Math Competitions I

Theorem 4.19 (Quadratic Formula). The solutions of the equation ax2 + bx + c = 0 are
√
−b ± b2 − 4ac
2a
.

Proof of Theorem 4.19. We utilize a technique called “completing the square.” That is, we wish
to create a squared term z(x − y)2 for constants z, y as this allows us to take the square root of
an equation and thereby reduce the degree of our x terms by 1.

ax2 + bx + c = 0
 
2 b c
⇐⇒ a x + x + =0
a a
 2
b2
  
b c 2 b
⇐⇒ a − =a x + x+ 2
4a2 a a 4a
2 2
 
b − 4ac 2 b b b
⇐⇒ =x + x+ 2 = x+
4a2 a 4a 2a
√
−b ± b2 − 4ac
⇐⇒ x =
2a

The quadratic formula allows us to directly compute roots, which makes factoring larger poly-
nomials even more potent as we only really need to factor into quadratics to compute all roots.
√
We call the part containing the radical, b2 − 4ac, the discriminant.
√
Proposition 4.20 (Discriminant). Let D = b2 − 4ac. The quadratic ax2 + bx + c has two
real solutions when D > 0, one real solution when D = 0, and no real solutions when
D < 0.

Exercise. Prove Proposition 4.20.

We’ve established some nice algebraic facts about quadratics. However, it is also useful to
think about quadratics from a graphical perspective. We start with vertex form:

y = a(x − h)2 + k

This is called vertex form because this directly gives us the vertex of the parabola: (h, k). To
think about why this is the vertex, suppose x = h. If a > 0, then if x < h or x > h will yield a
positive value of (x − h)2 , hence any other value of x will make the y greater than when x = h.
Similar reasoning applies when a < 0.
Vertex form can be obtained from the standard form of a quadratic, y = ax2 + bx + c, by
completing the square (the same method we used in the proof of the quadratic formula).
As an example, here is the graph of y = 3(x − 2)2 + 4 :

12
Everaise Academy (2020) Math Competitions I

Now we’ll talk about the graphical implications of the discriminant.

We know that if D > 0, a quadratic has two real solutions, so it intersects the x-axis twice.
If D = 0, then it is tangent to the x-axis (i.e. it touches the x-axis but does not pass through
it). For example, here is the graph of y = x2 :

Note how it touches the x-axis, but bounces off. This behavior occurs in any quadratic with
discriminant 0.
Finally, if the discriminant is less than 0, the quadratic never touches or passes through the
x-axis.

13
Everaise Academy (2020) Math Competitions I

Now let’s apply our knowledge of quadratics to some problems.

Example 4.21. (2009 AMC 12A) Functions f and g are quadratic, g(x) = −f (100 − x),
and the graph of g contains the vertex of the graph of f . The four x-intercepts on the two
graphs have x-coordinates x1 , x2 , x3 , and x4 , in increasing order, and x3 − x2 = 150. The
√
value of x4 − x1 is m + n p, where m, n, and p are positive integers, and p is not divisible
by the square of any prime. What is m + n + p?

Solution. Something that can immediately observed is a kind of symmetry about x = 50.
Substituting in 50 − x for x gives us that g(50 − x) = −f (50 + x).

AoPS
Thus, if a is a point a distance of x to the left of 50, and b is a point a distance of x to the right
of 50, we know that g(a) = −f (b).
Thus, we can simplify our calculations by simply shifting our graph to the left by 50. Instead
of g(x) and f (x) being symmetric about x = 50, they are now centered about the y-axis! The
question asks for the distance between two roots. Since we translated every point equally,
distance should be preserved, so we don’t need to worry about doing this.
Then we know that x3 = −x2 since both roots must be equidistant from the y-axis.
Thus, x3 = (75, 0) and x2 = (−75, 0). We know that we can express a quadratic as

f (x) = a(x − r1 )(x − r2 )

for roots r1 and r2 as well as constant a. For the sake of simplicity, suppose x4 = −b. Applying
this expression to both f (x) and g(x) gives:

f (x) = a(x − 75)(x − b)

g(x) = −a(x + 75)(x + b)

Note that it doesn’t matter if f (x) gets the negative or positive root - symmetry will give the
same results either way. From here, we want to use the info that g(x) contains the vertex of
f (x). So, we get the vertex form of f (x).

f (x) = ax2 − a(75 + b)x + 75ab

14
Everaise Academy (2020) Math Competitions I

75+b
Thus, 2 is the x value of our vertex. Hence, we have the following relationship:
   
75 + b 75 + b
p =q
2 2

Plugging this into p(x) and q(x) respectively gives:

    
75 + b −75 + b 75 − b
p =a
2 2 2
    
75 + b 225 + b 75 + 3b
q = −a
2 2 2

Setting these two equal to each other and then simplifying gives:
     
−75 + b 75 − b 225 + b 75 + 3b
=−
2 2 2 2
(−75 + b)(75 − b) = −(225 + b)(75 + 3b)
b2 − 150b + 752 = 3 · 752 + 3b2 + 750b
2b2 + 900b + 2 · 752 = 0

Thus, we get that b2 + 450b + 752 = 0. Applying the quadratic formula gives:
√
−450 ± 4502 − 1502
b=
2
Further simplifying this expression gives:
√
−450 ± 150 32 − 12
b=
2 √
b = −225 ± 150 2

The magnitude of b must√be larger than 75 since b and −b will be our roots x1 and x4 . Then√we
have that b = −225 − 150 2. Thus, we are looking for x4 − x1 = −b − (b) = −2b = 450 + 300 2.
Then the answer is 450 + 300 + 2 = 752. 

Example 4.22. (2018 AMC 10A) Find the interval for the set of values of a such that the
curves x2 + y 2 = a2 and y = x2 − a in the real xy-plane intersect at exactly 3 points.

Solution. We begin by drawing the diagram. Note that this problem will require the reader to
know the general equation for a circle. Conveniently, they gave our quadratic already in vertex
form, so we can derive that the vertex is (0, −a). Moreover, simply by plugging in (0, −a) into
x2 + y 2 gives a2 , so we know that there already is one intersection between the parabola and
circle.

15
Everaise Academy (2020) Math Competitions I

So, we just need to make sure that our parabola only intersects the circle at 2 more points,
requiring the curve to be ”inside” of the circle. In your own time, try playing around with
values of a to see how the graph adjusts. The easiest way to do so is to first see what happens
when we assume that the circle and parabola intersect.
We first set the two equations to each other, but in order to simplify calculations we instead
substitute x2 = y + a into the first equation to get
y 2 + y + a = a2 .
We can then manipulate the equation to get
y 2 + y + 1 = a2 − a + 1.
From here, we complete the square on both sides:
1 2 1 2
   
y+ = a−
2 2
From here, we can break this down into two cases:
1 1
y+ =a−
2 2
1 1
y + = −(a − )
2 2
We’ll first work with the second case. In this case, we have that y = −a. However, this doesn’t
give us any information as we already know that (0, −a) is an intersection.
Now, we return to the first case. We obtain that y = a − 1, and then we can substitute this
into the quadratic:
(a − 1) = x2 − a
2a − 1 = x2

16
Everaise Academy (2020) Math Competitions I

x will have 2 solutions as long as 2a − 1 > 0 which is what we want as we need 3 total
intersections. Hence, this gives the desired interval a > 12 .
Note that you don’t need to draw the diagram and could directly go into calculating the
intersections between the two curves; outside of stimulating intuition or ideas, it didn’t have
much purpose. 

§4.4 Homework Problems

Problem 4.1 (2000 AMC12). [4]Let f be a function for which f (x/3) = x2 + x + 1. The the
sum of all values of z for which f (3z) = 7 is a fraction − m
n for relatively prime positive integers
m and n. Find m + n.

Problem 4.2 (2010 AMC 10A). [5] The polynomial x3 − ax2 + bx − 2010 has three positive
integer roots. What is the smallest possible value of a?

Problem 4.3 (2007 AMC 10A). [5] Suppose that the number a satisfies the equation 4 =
a + a−1 . What is the value of a4 + a−4 ?

Problem 4.4 (2014 AIME II). [6] Real numbers r and s are roots of p(x) = x3 + ax + b, and
r + 4 and s − 3 are roots of q(x) = x3 + ax + b + 240. Find the sum of all possible values of |b|.

Problem 4.5 (2019 AMC 12A). [6] Let sk denote the sum of the kth powers of the roots of
the polynomial x3 − 5x2 + 8x − 13. In particular, s0 = 3, s1 = 5, and s2 = 9. Let a, b, and c be
real numbers such that sk+1 = a sk + b sk−1 + c sk−2 for k = 2, 3, .... What is a + b + c?

Problem 4.6 (2001 AMC12). [6] The polynomial p(x) = x3 + ax2 + bx + c has the property
that the average of its zeros, the product of its zeros, and the sum of its coefficients are all
equal. The y-intercept of the graph of y = p(x) is 2. What is |b|?

Problem 4.7 (1991 AIME Problem). [8] For how many real numbers a does the quadratic
equation x2 + ax + 6a = 0 have only integer roots for x?

Problem 4.8 (2003 AIME II). [10] Consider the polynomials P (x) = x6 − x5 − x3 − x2 − x
and Q(x) = x4 − x3 − x2 − 1. Given that z1 , z2 , z3 , and z4 are the roots of Q(x) = 0, find
P (z1 ) + P (z2 ) + P (z3 ) + P (z4 ).
Show your work for this problem, and include any progress even if you didn’t
manage to solve it!

17
5 Trigonometry and Complex Numbers
In this handout, we’ll go over basic trig and trig identities. We’ll also cover complex numbers
and Euler’s formula, where our knowledge of trig will come in handy.

Radians

Before we get into examples, we’ll first define radians, the SI(standard) unit of angle measure-
ment. Instead of having 360 degrees, a circle has 2π radians. This is a more useful unit than
degrees because the length of the arc of a unit circle is equal to the angle it subtends when
measured in radians.
It’s common convention to use radians when dealing with trigonometric functions, mainly
because the more natural definition allows us to formulate results a lot more easily and elegantly
in fields that use trig like complex analysis

§5.1 Trigonometric Functions

We’ll define trigonometric functions using a right triangle. Suppose that 4ABC has ∠B = π2 ;
then we label side BC as the side opposite of angle ∠A, and we label side AB as the side
adjacent to angle ∠A.

o h

B a A

The basic trigonometric functions are then defined as follows:

Function Name Symbol Value
BC
sine sin(∠A) AC
AB
cosine cos(∠A) AC
AB
tangent tan(∠A) AC
AB
cosecant csc(∠A) AC
AB
secant sec(∠A) AC
AB
cotangent cot(∠A) AC

A nice mnemonic to remember these formulas is SOH CAH TOA – Sine is Opposite over
Hypotenuse, Cosine is Adajcent over Hypotenuse, and Tangent is Opposite over Adjacent.
We can also observe right off the bat that
sin(∠A)
tan(∠A) =
cos(∠A)
sin(∠A) = cos(∠B) = sin(90◦ − ∠B)

1
Everaise Academy (2020) Math Competitions I

Similar relations hold for the other trig functions. We also see that

1
csc(∠A) =
sin(∠A)
1
sec(∠A) =
cos(∠A)
1
cot(∠A) =
tan(∠A)

Thus, we’re mainly going to be focusing on sine and cosine identities, since the other trig
functions can all be expressed in terms of these two.
√ √
3 2
Exercise 5.1. Show that sin( π3 ) = 2 and that cos( π4 ) = 2 .

Hint: View some common triangles with relevant angles.

Though f 2 (x) usually means f (f (x)) in the context of functions, this notation is instead used
for exponentiation when we’re dealing with trig functions; thus, sin2 x = sin x · sin x. However,
the convention for inverse is still the same (so that sin−1 (sin(x)) = x and sin(sin−1 (x)) = x for
x in the correct domains).
This can get confusing, so mathematicians often prefix an arc in front of trig functions to
denote inverses instead, such as arcsin(sin(x)) = x. We will be using this notation for the rest
of the handout. We also will omit the parentheses when it’s obvious what the input is. Now
let’s present our first identity:

Proposition 5.2 (Pythagorean Identity). sin2 x + cos2 x = 1 for 0 < x < π2 .

Exercise 5.3. Prove the Pythagorean Identity. Hint: There’s a reason it’s called the Pythagorean
Identity.

Exercise 5.4. Also prove that tan2 x + 1 = sec2 x and cot2 x + 1 = csc2 x.
In fact, using this identity along with some algebraic manipulations, we can actually already
solve some easy and middle difficulty AIME problems:

Example 5.5 (2003 AIME I #4). Given that log10 (sin x) + log10 (cos x) = −1 and that
log10 (sin x + cos x) = 21 (log10 n − 1), find n.

Solution. Immediately, it appears that the equations can be simplified and somehow combined
because the log parts share a common base. First, combine log10 (sin x) + log10 (cos x) in the
first equation to get log10 (sin x cos x) = −1. Exponentiating both sides with base 10, we get
1
sin x cos x = (5.1)
10
Simplifying the second equation,
1 n
log10 (sin x + cos x) = (log10 )
2 r10
n
log10 (sin x + cos x) = (log10 )
10

2
Everaise Academy (2020) Math Competitions I

Exponentiating both sides with base 10, we get

r
n
sin x + cos x = (5.2)
10
Square both sides, apply the Pythagorean Identity, and substitute equation 5.1 in.
n
sin2 x + cos2 x + 2 cos x sin x =
  10
1 n
1+2 =
10 10

Thus, n = 12. 

Remark. Squaring some form of sin x + cos x then replacing the sin2 x + cos2 x part with 1
is a common trick that appears on problems involving trig.

22 m
Example 5.6 (1991 AIME I #9). Suppose that sec x+tan x = 7 and that csc x+cot x = n,
where m
n is in lowest terms. Find m + n.

Solution. Looking at sec x + tan x and csc x + cot x, we notice that they bear a tiny bit of
resemblance to the identities tan2 x + 1 = sec2 x and cot2 x + 1 = csc2 x that we asked you to
prove in a previous exercise.
We wish to have tan2 x and sec2 x on opposite sides of the equation, so subtract sec x from
both sides of sec x + tan x = 22
7 to get

22
sec x = − tan x
7
Then, square both sides and apply the tan2 x + 1 = sec2 x identity:
   2
2 2 22 22
sec x = tan x − 2 tan x +
7 7
   2
2 2 22 22
sec x + 1 = tan x + 1 − 2 tan x +
7 7
   2
22 22
sec2 x + 1 = sec2 x − 2 tan x +
7 7
   2
22 22
1 = −2 tan x +
7 7

Thus, tan x = 435308 . We then attempt to further simplify the expression csc x + cot x in a
similar manner so that we get something in terms of cot x. This would be very helpful because
cot x = tan1 x and we already know the value of tan x. For ease of writing, let y = m
n . We use
the exact same procedure as before:

csc x = y − cot x
csc2 x = cot2 x − 2y tan x + y 2
csc2 x + 1 = cot2 x + 1 − 2y cot x + y 2
1 = −2y tan x + y 2

3
Everaise Academy (2020) Math Competitions I

1 308
Substituting in cot x = tan x = 435 and rearranging, we get
 
2 308
0=y −2 y−1 (5.3)
435
0 = 435y 2 − 2(308)y − 435 (5.4)
This quadratic can be solved by factoring it into (15y − 29)(29y + 15) or by the quadratic
formula. We now have two solutions, one of which is extraneous. Since the signs of the two
solutions differ, it would be nice to find out whether our answer is positive or negative.
Note that we showed that tan x = 435308 , which is positive. Since tan x =
sin x
cos x , sin x and cos x
are either both positive or both negative.
22
Another useful equation is the given one, sec x + tan x = 7 . We can rewrite this as
22
sec x + tan x =
7
1 sin x 22
+ =
cos x cos x 7
1 + sin x 22
= .
cos x 7
Note that 1 + sin x is always nonnegative, meaning that cos x is positive. Therefore, both sin x
and cos x are positive.
29
Since sin x, cos x are positive, we must have that csc x + cot x is positive, so y = 15 and not
15
− 29 . Our answer is m + n = 29 + 15 = 44. 

11
Example 5.7 (2019 AIME I #8). Let x be a real number such that sin10 x + cos10 x = 36 .
Then sin12 x + cos12 x = m
n where m and n are relatively prime positive integers. Find
m + n.

Solution. Substitute a = sin2 x and b = cos2 x. (This is a natural choice because all the
exponents are even). Note that we have a + b = 1 by the Pythagorean Identity as well as
11
a5 + b5 = 36 from the problem statement.
Then we can substitute in b = 1 − a to get
11
a5 + (1 − a)5 =
.
36
Expanding out the left side can be made easier by substituting c = 1/2 − a as this will cancel
out the odd-powered terms by symmetry. Performing this substitution and expanding gives
 5  5
11 1 1
= −c + +c
36 2 2
1
= (80c4 + 40c2 + 1).
16
Now we can use the quadratic formula or factor (note that the equation is a quadratic in c2 )
1 1
to get c2 = 12 . This gives c = ± 2√ 3
. We can choose either the positive or negative value of c.
We want the value of  6  6
6 6 1 1
a + (1 − a) = −c + +c .
2 2
13
Some computation (made easier with difference of cubes) gives us a value of 54 , so our answer
is 13 + 54 = 67.


4
Everaise Academy (2020) Math Competitions I

Notice that in our solution, we pretty much used no geometry, even though the trigonometric
functions are defined geometrically! This is actually pretty much true of almost all problems
that explicitly state trig functions: algebraic manipulation and gratuitous usage of identities
are usually your best bets, and geometric insights are rarely necessary or even useful.

Remark. In fact, it’s even common practice in computational geometry to blindly use the
Law of Cosines and Law of Sines to reduce problems to pure algebra, which is really the true
power of trigonometry. We’ll cover these two theorems in Week 3 when we do geometry.

§5.2 The Unit Circle

It turns out that there’s a generalized definition of the trigonometric identities. The idea is to
consider the circle x2 + y 2 = 1 plotted on the Cartesian plane; if we take an arbitrary point
(x1 , y1 ) on this circle, then if θ is the counterclockwise angle made by the line between (0, 0)
and (x1 , y1 ) and the x-axis, we can define cos θ = x1 and sin θ = y1 . Similar definitions hold for
the other trig functions; for instance, tan θ = xy11 .

(cos θ, sin θ)

This might seem a little confusing. What do we do with negative angles, or angles greater
than or equal to 2π?.

The key is to note that 2π is a full rotation around the unit circle, so cos(2π) = cos(0) = 1
and sin(2π) = sin(0) = 0, and the same is true for any multiple of 2π.
This means that the sine and cosine of an angle θ are equal to the sine and cosine of θ modulo
2π, respectively, so sine and cosine are periodic functions: they take all of their possible
values in a fixed interval, [0, 2π), and then repeat themselves from [2π, 4π), [4π, 8π], and so on
to infinity. Thus, we say that sin has period 2π, or sin is periodic modulo 2π.

We can interpret negative angles as clockwise angles instead; therefore, an argument of − 3π

2
inputted to the sine and cosine functions would return the same outputs as an argument of π2 .
That way, the intervals also repeat to the left: the values of [0, 2π] are repeated in the intervals
[−2π, 0], [−4π, −2π), etc. This also implies that sine is an odd function and that cosine is an
even function; that is, sin(−x) = − sin(x) for all real x and cos(−x) = cos(x) for all real x.
Here’s a graph of sin x and cos x:

5
Everaise Academy (2020) Math Competitions I

1
f (x) = cos x
0.5

−0.5 f (x) = sin x

−1
0 π π 3 2π
2 2π

Exercise 5.8. Show that tan is periodic modulo π, and try graphing the function.
Exercise 5.9. Find cos( π2 ) and cos( 5π
3 ).

Note that the sine graph is actually the cosine graph shifted by π2 because of our earlier
observation that sin x = cos( π2 − x). It’s also useful to note that sin x = sin(π − x) along with
cos x = − cos(π − x). Applying these can often come in handy with making an expression in
terms of sines or cosines only, as we will see in the next example.

Example 5.10. Find the value of

π π 3π π 9π
tan tan tan tan . . . tan .
20 10 20 5 20

Solution. Let’s try putting everything in terms of sines and cosines; this is often a useful
strategy if nothing clever comes to mind. The expression is equal to
π
sin 20 sin 2π 3π 4π 9π
20 sin 20 sin 20 . . . sin 20
π 9π ,
cos 20 cos 2π 3π 4π
20 cos 20 cos 20 . . . cos 20

but cos(x) = sin( π2 − x), so the expression becomes

π
sin 20 sin 2π 3π 4π 9π
20 sin 20 sin 20 . . . sin 20
π = 1.
sin 9π 8π 7π 6π
20 sin 20 sin 20 sin 20 . . . sin 20

Now we can present some more identities:

Proposition 5.11 (Angle Addition Identities). For real x, y, we have

sin(x + y) = sin x cos y + sin y cos x

cos(x + y) = cos x cos y − sin y sin x

Proof. We first show the sine Angle Addition Identity; we will only really prove the identity for
the case 0 < x + y < π. The cases where π < x + y < 2π or x + y = 0, π are left as an exercise.

6
Everaise Academy (2020) Math Competitions I

Consider a triangle 4ABC with largest angle at ∠A. Let D be the foot of the altitude from
A to BC, and let x = ∠BAD and y = ∠CAD. Then we must have

[ABC] = [ABD] + [ACD]

AB · AC sin(x + y) AD · BD AD · CD
= +
2 2 2
1
1
cos x· cos y sin(x+y)AB · cos x · AB · sin x AC · cos y · AC · sin y
= +
2 2 2
sin(x + y) sin x sin y
= +
cos x cos y cos x cos y
sin(x + y) = sin x cos y + sin y cos x.

x y

B D C

Now we show the cosine Angle Addition Identity, which this time requires no geometry and
as a result will be proven for all real x, y. We mentioned earlier that cos( π2 − x) = sin x, so
π 
cos(x + y) = sin −x−y
π2  π 
= sin − x cos(−y) + sin(−y) cos −x
2 2
= cos x cos y − sin x sin y

as desired. 

Exercise 5.12. Prove the sine Angle Addition Identity for the cases not shown above without
using geometry.
These are probably the most important trig identities you need to know. From them, we can
quickly derive a host of new identities:

Proposition 5.13 (Angle Subtraction Identities). We have

sin(x − y) = sin x cos y − sin y cos x

cos(x − y) = cos x cos y + sin x sin y

for real x, y.

Proposition 5.14 (Double Angle Identities). We have

sin(2x) = 2 sin x cos y

7
Everaise Academy (2020) Math Competitions I

cos(2x) = cos2 x − sin2 x

= 1 − 2 sin2 x
= 2 cos2 x − 1.

for real x.

Proposition 5.15 (Half Angle Identities). We have

r
x 1 − cos x
sin =±
2 2
x r
1 + cos x
cos =±
2 2
for real x. Of course, cos( x2 ) and sin( x2 ) do not actually take on two values, and the sign
is dependent upon which quadrant the angle x2 is in.

Proposition 5.16 (Product-to-sum Identities). We have

1
sin x sin y = (sin(x + y) + cos(x − y))
2
1
cos x cos y = (cos(x + y) + cos(x − y))
2
for real x, y.

Proposition 5.17 (Sum-to-product Identities). We have

   
x+y x−y
sin x + sin y = 2 sin cos
2 2
   
x+y x−y
cos x + cos y = 2 cos cos
2 2
for real x, y.

Exercise 5.18. Derive these identities yourself. Hint: The Pythagorean Identity is useful for the cosine
Double Angle Identity, and the Double Angle Identities are useful for the Half Angle Identities. Use the Angle
Addition Identities to prove the Product-to-sum Identities, and then make a substitution into the Product-to-sum
Identities to get the Sum-to-product Identities.
tan x+tan y
Exercise 5.19. Prove that tan(x + y) = 1−tan x tan y for real x, y.
√ √
π 6− 2
Exercise 5.20. Show that sin( 12 )= 4 .

Example 5.21 (2012 AIME II #9). Let x and y be real numbers such that sin x
sin y = 3 and
cos x 1 sin 2x cos 2x p
cos y = 2 . The value of sin 2y + cos 2y can be expressed in the form q , where p and q are
relatively prime positive integers. Find p + q.

8
Everaise Academy (2020) Math Competitions I

sin 2x
Solution. To find sin 2y , we can just observe by the sine Double Angle Formula that this is
equal to
2 sin x cos x 3
= .
2 sin y cos y 2
cos 2x
That was pretty quick, but unfortunately cos 2y isn’t as easy. Let’s try to get rid of one variable.
1 1
We’ll start with x first because is nicer to deal with than
2 3. Since sin x = 3 sin y and
cos x = 12 cos y, we can use the Pythagorean Identity to get

1 1
1 = sin2 x + cos2 x = 9 sin2 y + cos2 y = 9 sin2 y + (1 − sin2 y).
4 4
3
Solving for sin2 y, we get 35 . Now we can plug and chug with the cosine Double Angle Formula:
we get
54 19
cos 2x = 1 − 2 sin2 x = 1 − 18 sin2 y = 1 − =−
35 35
and
6 29
cos 2y = 1 − 2 sin2 y = 1 − = ,
35 35
so we get
sin 2x cos 2x 3 19 49
+ = − =
sin 2y cos 2y 2 29 58
so our final answer is p + q = 107. 

We’ll end this section with a list of the values of trigonometric functions of some common
angles, which is definitely worth memorizing.

Angle sin cos tan

0 0 √
1 0
√
π 1 3 3
6 √2 √2 3
π 2 2
4 √2 2 1
π 3 1
√
3 2 2 3
π
2 1 0 undefined
2π
√
3
√
3 √2
−√12 − 3
2
3π
4 2 − √22 −1
√
3
5π
6
1
2 − 2 − 33
π 0 −1√
0
√
7π
6 −√12 − √23 3
3

5π
4 − √22 − 2
2
1
4π 3
√
3 − 2 − 12 3
3π
2 −1 0 undefined
√ √
5π
3 − √23 1
√2
− 3
7π
4 − 22 √2
2
−1
√
3
11π
6 − 12 2 − 33
2π 0 1 0

9
Everaise Academy (2020) Math Competitions I

√ π
2π 1 3 2 (0, 1) √
(− 2 2 )
, π 3
√
3
3 ( 12 ,
2√)
3π 3 1
4 (− 2 , 2 )
π 3 1
√ 4 ( 2 , 2)
5π 3 1 √
6 (− 2 , 2 )
π 3 1
6 ( 2 , 2)

0 (1, 0)
π (−1, 0)

√ √
7π 3 1
(− 2 , −2)
3
11π 1
6 2 , −2)
6 (
√ √
3
5π
4 (− 1
2 , −2)
7π
( 23 , − 12 )
4 √
√ 5π 1 3
3 (2, − 2 )
4π 1 3
3 (− 2 , − 2 ) 3π
2 (0, −1)

A nice trick to remember these values is to learn the first quadrant and then use the mnemonic
All Students Take Calculus to figure out the signs: All the trig functions are positive in the
first quadrant, only Sine is positive in the second quadrant, only Tangent is positive in the
third quadrant, and only Cosine is positive in the fourth quadrant.

Example 5.22 (2000 AIME II #15). Find the least positive integer n such that
1 1 1 1
+ + ... + = .
sin 45◦ sin 46◦ sin 47◦ sin 48◦ sin 133◦ sin 134◦ sin n◦

Solution. Whenever we see a long fraction summation equal to a nice result, telescoping should
be the first thing that comes to mind. Let’s keep this in mind.
We can observe that the 1 in the numerator on the right hand side is sort of suspicious. This
means that multiplying both sides by something nice is probably going to be useful. Now here
comes the intuitive leap: multiplying by sin 1◦ gives us what we want. Why is this true?
Well, sin 1◦ = sin((x + 1)◦ − x◦ ) = sin(x + 1)◦ cos x◦ − sin x◦ cos(x + 1)◦ , and this leads to nice
cancellation both in the fractions and of terms using the fact that cot(x) = − cot(π − x):

sin 1◦ sin 1◦ sin 1◦ sin 1◦

= + + . . . +
sin n◦ sin 45◦ sin 46◦ sin 47◦ sin 48◦ sin 133◦ sin 134◦
◦ ◦ ◦
sin 46 cos 45 − sin 45 cos 46 ◦ sin 134◦ cos 133◦ − sin 133◦ cos 134◦
= + . . . +
sin 45◦ sin 46◦ sin 133◦ sin 134◦
= (cot 45 − cot 46 ) + (cot 47 − cot 48 ) + . . . + (cot 133◦ − cot 134◦ )
◦ ◦ ◦ ◦

= cot 45◦ − (cot 46◦ + cot 134◦ ) + (cot 47◦ + cot 133◦ ) + . . . + (cot 89◦ − cot 91◦ )
= cot 45◦
= 1,

so n = 1 clearly works. 

§5.3 Complex Numbers

Say we have the following problem:

10
Everaise Academy (2020) Math Competitions I

Proposition 5.23. Solve the equation x2 + 1 = 0.

√
This seems ludicrous– we’d need the square root of −1. What is −1? It can’t be any of the
numbers we’re familiar with, because whenever we take the square of a number it’s non-negative.
This motivates the idea of complex numbers.

An imaginary number is defined as any number whose square is negative. We define

√
i = −1 as the imaginary unit.. In order to distinguish between pure imaginary numbers
and the numbers we’re familiar with, we call the numbers that we normally work with real
numbers, which are numbers whose square is a non-negative value. Note that this means any
imaginary number a can be written in the form b · i, where b is a real number and i is the
imaginary unit.

Example 5.24. Solve x2 + 4.

√ √ √
Solution. Note that x2 = −4 =⇒ x = −4 = −1 · 4 = ±2i. 

Example 5.25. What is the value of 9i · 4i?

Solution.
9i · 4i = 9 · 4 · i2 = 9 · 4 · −1 = −36


Exercise 5.26. Find the values of i3 , − 1i , and i7 .

Where have we discussed taking the square root of a negative number before? When talking
about quadratics, we mentioned that the discriminant must be non-negative in order for there
to be a solution. Now we have a way to solve it when it’s negative!

Example 5.27. Solve the equation x2 + x + 1 = 0.

Solution. We use the quadratic formula:

√ √ √
−1 ± 12 − 4 · 1 · 1 −1 ± −3 −1 ± 3i
x= = =
2 2 2


Some of you may notice that we have actually added together a pure imaginary number and
real number: note that if we square it, it’s neither a non-negative nor negative real number.
In fact, it’s another real number plus an imaginary number! We call this number a complex
number, or the sum between a real number and imaginary number. Remembering that we
can look at imaginary numbers as a real number multiplied by i, we can characterize complex
numbers as follows:

Definition 5.28. A complex number is a number in the form a + bi, where a and b are real
and i is the imaginary unit.

11
Everaise Academy (2020) Math Competitions I

We refer to a and b as the real part and imaginary part of a complex number a + bi,
respectively. For a complex number z, we usually denote the real part as <(z) and the imaginary
part as =(z). Note that both <(z) and =(z) both return real numbers. One seemingly trivial
but nonetheless important idea to keep in mind is the following:

Proposition 5.29. Two complex numbers x and y are equal if and only if their real parts
are equal and their imaginary parts are equal.

This should be fairly intuitive. Let’s see how we can use this fact:

Example 5.30. If a2 + 4i = 25 − bi, where a and b are real numbers, then find all possible
pairs (a, b).

Solution. Remember two complex numbers are equal if and only if their real and imaginary
parts match up with each other. This gives us the equations

a2 = 25, 4 = −b

which imply the only solutions are (a, b) = (±5, −4). 

Exercise 5.31. Determine which of the following are complex numbers: i, 1 − 2i, 3/π.
Exercise 5.32. Express the quotient (a + bi)/(c + di) in the form p + qi.

§5.4 The Complex Plane

We’re accustomed to graphing real numbers on a number line, but what if we tried to extend
this to complex numbers? We’d probably want to keep the real and imaginary parts separate;
this means that we should have two axes, similar to the xy-plane. This leads to the idea of the
complex plane. If we wanted to graph the complex number a + bi in the complex plane, then
the “x-coordinate” would be a and the “y-coordinate” would be b; that is, the x-axis is the real
axis and the y-axis is the imaginary axis. The origin is still (0, 0), which corresponds to the
complex number 0.

12
Everaise Academy (2020) Math Competitions I

Definition√ 5.33. The magnitude of a complex number z = a + bi is the distance from the
origin, or a2 + b2 . The magnitude of z is denoted as |z|, since if z is real then this is equivalent
to the absolute value function.

Exercise 5.34. Show that for any complex numbers z and w, |zw| = |z||w|.

Definition 5.35. the conjugate of a complex number z = a + bi is defined as a − bi. The

conjugate of z is denoted as z.

Exercise 5.36. Show that for any complex numbers z and w, z = z, z + w = z + w, and
zw = z · w.
Exercise 5.37. Show that a complex number z is real if and only if z = z, and that it is imaginary
if and only if z = −z.
Exercise 5.38. Show that zz = |z|2 .
Exercise 5.39. Graphically, what does the conjugate of z represent?
Arithmetic is also doable on the complex plane. For addition, it’s quite simple to apply our
knowledge. If we have points a + bi and c+ di, then the sum is just (a + c) + (b+ d)i. Multiplying
a + bi by a real c is also very simple as we are given c(a + bi). However, multiplication of two
complex numbers is slightly more unwieldy in this form, but thankfully this is remedied in the
following section.

§5.5 Euler’s Formula

Before we introduce Euler’s Formula, we introduce some new material.
Consider a complex number z = a + bi. First, let’s divide a + bi by a real number c such that
a+bi
c has a distance of 1 from the origin; this puts z on the unit circle.
Then, it follows from our generalized definitions of sine and cosine that if θ is the measure
of the counterclockwise angle made by the line between a + bi and 0 and the real axis (called
a + bi’s argument), ac can be described as cos θ and cb can be described as sin θ.
a+bi
Therefore, c is equal to cos θ + i sin θ, and multiplying by c, we have that

a + bi = c cos θ + ic sin θ.

This is a very important interpretation of the complex numbers known as polar form, because
we’re expressing z in terms of an argument and a magnitude only, which is exactly what regular
polar coordinates are in the Cartesian plane. The representation a+bi is known as rectangular
form because a + bi, a, bi, and 0 are the vertices of a rectangle.

Notation. Let cos θ + i sin θ = cis θ.

We first introduce an important theorem that will help us with Euler’s Formula.

Theorem 5.40 (De Moivre’s Theorem). cis nθ = cisn θ for all integral n and real θ.

Proof. To prove this for all integral n, we use induction. Fix θ. First, we know that cis(0 · θ) =
cis 0 = cos 0 + i sin 0 = 1 = cis0 θ, so our base case is complete.

13
Everaise Academy (2020) Math Competitions I

We assume that this the assertion is true for some nonnegative integer k, and we prove it for
k + 1. It follows that

(cos θ + i sin θ)k+1 = (cos θ + i sin θ)(cos θ + i sin θ)k

=⇒ (cos θ + i sin θ)k+1 = (cos θ + i sin θ)(cos kθ + i sin kθ) from our assumption for k
=⇒ (cos θ + i sin θ)k+1 = cos θ cos kθ + i cos θ sin kθ + i sin θ cos kθ − sin θ sin kθ
=⇒ (cos θ + i sin θ)k+1 = cos((k + 1)θ) + i sin((k + 1)θ)
from our angle addition identities.

Hence, we are done as we’ve proven the base case and the inductive step. To extend the state-
ment to negative integers as well, observe that cis θ is the conjugate of cis(−θ), so cis θ cis(−θ) =
1 =⇒ cis(−θ) = cis−1 θ. Thus, for arbitrary n, cis(−nθ) = (cis(−θ))n = (cis−1 (θ))n = cis−n θ
as desired. 

Exercise 5.41. Prove that sin 3x = (4 cos2 x − 1) sin x.

With De Moivre, we can now understand Euler’s formula. We state Euler’s formula without
proof since the proof requires calculus, which is out of the scope of this class.

Theorem 5.42 (Euler’s Formula). eix = cis x for all real x.

This third form is called exponential form for obvious reasons. Notice that this gives us
a very natural explanation of de Moivre, and also allows us to understand the consequences of
multiplying complex numbers much more easily:

Proposition 5.43. We have

cis x cis y = cis(x + y),
cis x
= cis(x − y)
cis y
for real x, y.

This is much faster than multiplying complex numbers in rectangular form!

Try to digest the graphical significance of these facts. If we multiply a complex number by
cis θ, then we just rotate it in the complex plane by θ. So multiplication by any complex number
just corresponds to scaling by the magnitude, then rotating by the argument.
Exercise 5.44. Prove Proposition 5.43.
From Euler’s Formula, we get the beautiful relation known as Euler’s Identity:

Theorem 5.45. (Euler’s Identity) eiπ + 1 = 0.

Proof. eiπ = cis π = cos π + i sin π = −1 ⇐⇒ eiπ + 1 = 0 

The following example demonstrates the power of Euler’s Formula:

14
Everaise Academy (2020) Math Competitions I

Example 5.46. (Mu Alpha Theta 1991) Evaluate

∞
X cos(nθ)
2n
n=0

where cos θ = 15 .

Solution. First, let’s try to see why Euler’s Formula is a good candidate for this problem.
The key here is that the problem asks us to deal with the cosine of multiples of a fixed angle.
We know that we can easily deal with multiples of angles from De Moivre’s Theorem, so we
want to tie this in somehow.
Moreover, we have a sum that resembles a geometric sum, except we have cos(nθ) in the
numerator— then we want to use something that ties in De Moivre’s Theorem and geometric
series together, which is exactly what Euler’s Formula does!
So how will we apply Euler’s Formula? Well, the key idea is that cos(nθ) is the real part of
ei(nθ) .
Why? Well, this is immediate by Euler’s Formula:

ei(nθ) = cos(nθ) + i sin(nθ)

ei(nθ)
Thus, our problem is equivalent to summing the real parts of 2n , or the real part of the sum
∞ i(nθ)
X e
2n
n=0

e0 eiθ
Now we have ourselves a geometric series, with first term 20
= 1 and ratio 2 . We can evaluate
the sum with techniques we already know:
∞ i(nθ)
X e 1 2
= =
2n 1− eiθ 2 − eiθ
n=0 2

2 2
√ the Pythagorean Identity, we have cos θ + sin θ = 1. Plugging in cos θ, we get sin θ =
From
± 2 5 6 . How do we know which one to use? The answer is that we don’t— the problem doesn’t
specify which quadrant θ lies in. However, remember that we only care about the real part of
the sum, or just the cosines. Since taking plus or minus doesn’t affect the cosines, we can just
choose any value to use! Let’s take the positive value; from Euler’s Formula,
√
iθ 1 2 6
e = cos θ + i sin θ = + i
5 5
Plugging this into our sum,
2 2 2
iθ
=  √  = √
2−e 2 − 15 + 2 6 9
+ 2 6
5 i 5 5 i

Exercise 5.47. Evaluate the sum: although it’s tedious, it’s important to practice rationalizing
the denominator so that you can do it faster later on.

As we alluded to in the exercise, we can simply rationalize the denominator:

 √  √
9 2 6
2 − 5 i
18
− 4 6
2 5 5 5 i
√ = √  √  = 81+24
9 2 6 9 2 6 9 2 6
5 + 5 i 5 + 5 i 5 − 5 i 25

15
Everaise Academy (2020) Math Competitions I

Since we just want the real part, our answer is

18
5 90 6
105 = =
25
105 7

Let’s recap the steps we took again. First, we noticed that the problem involved the cosine of
multiples of a fixed angle, θ, so that told us we should use De Moivre’s and complex numbers
somehow. Combining this with the fact that we had a 2n , we decided that geometric series
was the way to go, which led to using Euler’s Formula, since it gives us an exponential form of
complex numbers. From here, we knew that we could evaluate the series using the geometric
series sum formula, and then take the real part to just get the cosines.
Exercise 5.48. Evaluate
∞
X sin (nθ)
2n
n=0

where cos θ = 15 .
Let’s look at another application of De Moivre’s Theorem and Euler’s Formula:

Theorem 5.49. The solutions to the equation xn = 1 for some natural number n are of
2πk
the form ei n , where 0 ≤ k ≤ n.

Proof. Suppose the complex number z = reiθ is a root of xn − 1. Since r is the magnitude of
z, the magnitude of z n must be rn . Since r is a positive real number and rn = 1, we must have
r = 1. Thus, z = eiθ . We need z n = einθ to equal 1. Note that ei2π = cos(2π) + i sin 2π = 1, so

einθ = ei2πk = 1

for some positive k.

By exponent rules, we must have nθ = 2πk =⇒ θ = 2πk n . Now, if k ≥ n, we can pull out
e2π ,which is equal to one, so it is a duplicate of one of the roots with 0 ≤ k ≤ n. The same
goes for the other direction: if k < 0, then we can multiply by e2π = 1 to get one of the roots
already described.
i2πk
Thus, e n for 0 ≤ k ≤ n characterizes the n distinct roots of the polynomial xn − 1 (note
that this also satisfies the Fundamental Theorem of Algebra). 

These n roots are known as the nth roots of unity. Note that the angle between each consecutive
root of unity is constant, and the magnitude is always 1, so geometrically, they form a regular
n-gon in the complex plane.
Here’s an example of this in action:

Example 5.50 (2014 AIME II, Modified). Let z be a complex number and P be the polygon
1
in the complex plane whose vertices are z and every w such that z+w = z1 + w1 . What
geometric shape is P ?

Solution. Let’s first interpret the given condition on z and w. Combining the fractions and
rearranging, we get
1 1 1 z+w
= + = =⇒ (z + w)2 = zw
z+w z w zw

16
Everaise Academy (2020) Math Competitions I

which means z 2 + zw + w2 = 0. But wait! This expression is a factor of difference of cubes. We

can rewrite our equation, then, in the following form:

z 3 − w3
z 2 + zw + w2 = =0
z−w
This means that the solutions for z are the solutions of z 3 − w3 = 0 and z 6= w. Since z is
nonzero, w is nonzero, so we can divide the equation by z 3 :
 w 3
z 3 − w3 = 0 =⇒ 1 − =0
z
w
As we talked about earlier, this means that z are the 3rd roots of unity, except for 1, since it
got discarded.

We want the area formed by all possible points w and z in the complex plane, so let’s interpret
our points geometrically. We have two values of w since z is fixed and wz is a third root of unity
other than 1. If ω is a third root of unity other than 1 (e.g. cis 2π 2
3 ), then ω is the other third
root of unity not equal to 1 (you can check this yourself).
Then, letting ω = wz , we get that our three vertices are z, zw, and zw2 . These are the three
third roots of unity, all multiplied by z. What does this mean? The third roots of unity form
an equilateral triangle in the complex plane, but do we still have an equilateral triangle in this
case?
It turns out we do. If we let z = |z| cis θ, then multiplication by z simply scales up each
vertex by |z| then rotates it by θ. This still preserves the distance between the three points, so
our answer is an equilateral triangle. 

Exercise 5.51. The original problem gives |z| = 2014 and asks for the area of P . Try solving it.

§5.6 Problems
Problem 5.1 (1988 AHSME). [5] The √ complex number z satisfies z + |z| = 2 + 8i. What is
|z|2 ? Note: if z = a + bi, then |z| = a2 + b2 .

Problem 5.2 (2019 AMC 12A). [6] For a certain complex number c, the polynomial

P (x) = (x2 − 2x + 2)(x2 − cx + 4)(x2 − 4x + 8)

has exactly 4 distinct roots. What is |c|2 ?

Problem 5.3 (2019 AMC 12B). [6] How many nonzero complex numbers z have the property
that 0, z, and z 3 , when represented by points in the complex plane, are the three distinct vertices
of an equilateral triangle?

Problem 5.4 (2019 AIME I). [9] Given f (z) = z 2 − 19z, there are complex numbers z with
the property that z, f (z), and f (f (z)) are the vertices of a right triangle in the complex plane
with a right angle at f (z). There are positive integers m and n such that one such value of z is
√
m + n + 11i. Find m + n.

Problem 5.5 (2018 AIME I). [10] Let N be the number of complex numbers z with the
properties that |z| = 1 and z 6! − z 5! is a real number. Find the remainder when N is divided
by 1000.

17
Everaise Academy (2020) Math Competitions I

Problem 5.6 (2017 AIME I). [8] Let z1 = 18 + 83i, z2 = 18 + 39i, and z3 = 78 + 99i, where
√
i = −1. Let z be the unique complex number with the properties that zz32 −z z−z2
−z1 · z−z3 is a real
1

number and the imaginary part of z is the greatest possible. Find the real part of z.

Problem 5.7 (2020 AMC 12A). [10] In the complex plane, let A be the set of solutions to
z 3 − 8 = 0 and let B be the set of solutions to z 3 − 8z 2 − √
8z + 64 = 0. We can express the
greatest distance between a point of A and a point of B as a b where a and b are integers and
b is not divisible by the square of any integer. What is a + b?
√
Problem 5.8 (2019 AMC 12B). [12] Let ω = − 12 + 12 i 3. Let S denote all points in the complex
plane of the form
√ a + bω + cω 2 , where 0 ≤ a ≤ 1, 0 ≤ b ≤ 1, and 0 ≤ c ≤ 1. We can express the
p q
area of S as r , where p, q, and r are positive integers. What is p + q + r?

Problem 5.9 (2011 AIME II). [10] Let z1 , z2 , z3 , . . . , z12 be the 12 zeroes of the polynomial
z 12 − 236 . For each j, let wj be one of zj or izj . Then the maximum possible value of the real
12
X √
part of wj can be written as m + n where m and n are positive integers. Find m + n.
j=1
Provide a detailed solution, or explain what progress you made, even if you didn’t
get an answer.

18
6 Basic Counting

From a young age, we learn to count. Whether it’s the number of fingers on our hands or
the toes on our feet, counting is basically an ingrained skill. In this chapter, we will explore
the mathematics of counting. This handout will review the basics of counting so we can tackle
more advanced concepts in the following days.

§6.1 As Easy As 1, 2, 3
We first define one of the most common mathematical objects, used almost everywhere.

Definition 6.1. A set is an unordered collection of distinct elements (i.e. no repeats). The
cardinality of a set with a finite number of elements is the number of elements it contains. We
denote the cardinality of a set A as |A|. We usually use capital letters to represent sets.

Definition 6.2. The complement of a set A is the set of all possible elements not in A. The
complement is denoted AC . Context is usually given as to what “all possible elements” means.

Remark. It turns out that there is a way to generalize the definition of cardinality to infinite
sets as well. In the process, we end up categorizing different “sizes” of infinities using the
so-called aleph numbers, and perform arithmetic with them. This is outside of the scope
of this handout, but if you’re interested, search up “countable sets”.

For example, the set S of solutions to the quadratic equation x2 − 1 = 0 is

S = {−1, 1},

and since sets are unordered, this is considered the same set as {1, −1}.
Sets also aren’t limited to just numbers; they can contain anything from words to even other
sets. A shorthand we use when describing a set containing all elements satisfying a certain
property is with a colon or the symbol |. Thus, another way we could describe S is

S = {x : x2 − 1 = 0},

read as “all x such that x squared minus one equals zero”.

Definition 6.3. A subset A of a set B is a set such that all of its elements are also elements
of B. We denote this by A ⊆ B, and if A is not equal to the set B (i.e. there exists an element
of B not in A), then we say that A is a proper subset of B and denote this by A ⊂ B. The
empty set, denoted ∅, is a set containing no elements. Thus, ∅ is a subset of every set.

Counting problems generally revolve around finding valid subsets, where validity is determined
by a given condition in the problem statement.

Example 6.4. How many numbers are in the set {1, 2, 3, . . . , 10}?

1
Everaise Academy (2020) Math Competitions I

Solution. You can count the number of terms aloud or write them down and count how many
numbers you wrote, but the answer is 10. 

Obviously, the previous problem wasn’t very much of a problem. The numbers were small,
and there wasn’t much thinking involved at all.

Example 6.5. How many numbers are in {101, 102, 103, . . . , 110}? Or {5, 6, . . . , 14}?

Solution. The answer is 10 in both these cases. We can figure this out by pairing up numbers
with the first 10 integers. In the first case, we pair up each number n with n + 100: for example,
7 goes to 107. In the second case, we pair up each number n with n + 4. Thus, both these lists
have 10 numbers.


Example 6.6. How many numbers are in {5, 7, 9 · · · 39}?

Solution. We can match n to 2n + 1. For instance, 1 goes to 5, 2 goes to 7, and so on. Notice
that 19 goes to 39. Thus there are 19 numbers in this set. 

What we just established was a one-to-one correspondence between the elements of

{1, 2, . . . 10} and the two sets in this problem. Whenever we have a one-to-one correspon-
dence between two sets A and B, we know that |A| = |B|. We’ll go over this in more detail
tomorrow.
Exercise.
(a) Show that if our set is {a, a + 1, . . . b} for positive integers a ≤ b, then the amount of terms
in the set is b − a + 1.
(b) How many numbers are in the list 2.1, 2.2, . . . , 4.0?

§6.2 Constructive Counting

Constructive counting is a fundamental method in combinatorics. Essentially, we count the
number of objects that satisfy some property by “constructing” them, i.e. directly creating
the possible objects that satisfy the given conditions, and keeping track of how many there
are.

Example 6.7. How many ways are there to choose a boy and a girl from a class of 12 boys
and 13 girls?

Solution. Note the choice of boys is separate from the choice of girls. Thus, the total number
of pairs is 12 × 13 = 156. 

Example 6.8. The Committee of Everaise needs to split 3 unique jobs among 8 staff
members. If each staff member can do as many jobs as they please, how many ways are
there to assign the jobs?

2
Everaise Academy (2020) Math Competitions I

Solution. Each job can be done by 8 people, and you make a choice for all 3 jobs. Thus the
answer is 8 × 8 × 8 = 512. 

Example 6.9. How many 5 digit integers are there with at least two consecutive 1’s?

Solution. Note we have four different places to place the 11: in spots 11 , 11 , 11 , or
11. However, there’s one discrepancy: a five digit integer can’t have 0 as its first digit. In
the first case, the last three digits can be anything, giving us 10 × 10 × 10 = 1000, but for the
last three cases, we only have 9 choices for the first digit, giving us 9 × 10 × 10 = 900. Thus, in
total we have 1000 + 3 × 900 = 3700. 

Exercise (2015 AMC8). How many integers between 1000 and 9999 have four distinct digits?
Exercise. Let a word be a string of letters that alternate between consonants and vowels. In
terms of n, how many words of length n are there that start with a consonant? Look at some
small specific values of n.
Exercise. A palindrome is a word that reads the same forwards and backwards. How many
3-digit palindromes are there? 4-digit? Can you generalize?

§6.3 Casework
Casework is essentially organized constructive counting: we split a problem into several smaller
problems (the “cases”) with restrictions that make them easier to count, and then we add them
up at the end to get the final answer.
Casework can often be very bashy, but on the flip side is useful for solving a huge number
of counting problems that appear on the AMC/AIME. One of the most important skills in
competition math is to be able to quickly determine whether or not a problem can feasibly be
solved with casework within the time limit. This intuition can only be gained through practice.
An important warning is that cases must be independent: otherwise, we could potentially
count a particular possibility either more than once or not at all (i.e. overcount). Let’s take a
look at an actual problem:

Example 6.10. (2004 AIME II) A jar has 10 red candies and 10 blue candies. Terry picks
two candies at random, then Mary picks two of the remaining candies at random. Given
that the probability that they get the same color combination, irrespective of order, is
m/n, where m and n are relatively prime positive integers, find m + n.

Solution. We first consider the case where Terry picks out 2 red candies. He can pick out 2
red candies with probability
10 9 9
· = ,
20 19 38
and the probability that Mary does too after Terry has picked his two red candies is
8 7
· .
18 17
14
Thus, the probability that Mary and Terry both pick two red candies is 323 . Next, we count
the case where Terry picks out 1 red and 1 blue candy - note that it doesn’t matter which order

3
Everaise Academy (2020) Math Competitions I

he picks them due to symmetry, so we just need the second ball to be different. He will pick
out this particular combination with
20 10
·
20 19
probability and Mary will pick out 1 red and 1 blue with
18 9
·
18 17
90
probability, so the probability that both pick out these combinations is 323 . The probability
that both pick 2 blue candies is the same as the probability of both picking out 2 red candies,
so we just multiply the number we got above by 2 for this last case.
Summing up the probabilities gives
90 28 118
+ = .
323 323 323

Let’s move onto another example:

Example 6.11. The numbers 1447, 1005, and 1231 have something in common: each is a
four-digit number beginning with 1 that has exactly two identical digits. How many such
numbers are there?

Solution. Looking at the given examples, we notice two ’types’ of integers: one where the 1 is
the repeated digit, and the other where a different digit is repeated twice. Hence, this gives us
the idea of casework.
Let’s look at the case where 1 is the repeated digit. Then, we have 3 places to put the other
1 as we must start with a 1. In the other two spots, we have 9 · 8 = 72 ways to fill up the other
two digits, as they must be distinct from 1, so we have 216 different integers under this case.
For the next case, we again must choose 2 other digits, similarly with 72 different ways. Now,
if our digits are x and y, we have 3 ways to arrange them:

1xxy
1xyx
1yxx

so we end up with another 216 different integers. Summing them up gives our answer of
216 + 216 = 432. 

Exercise 6.12 (2006 AMC 12A). A bug starts at one vertex of a cube and moves along the edges
of the cube according to the following rule. At each vertex the bug will choose to travel along
one of the three edges emanating from that vertex. Each edge has equal probability of being
chosen, and all choices are independent. What is the probability that after seven moves the bug
will have visited every vertex exactly once?
Exercise 6.13 (2005 AMC 10A). How many three-digit numbers satisfy the property that the
middle digit is the average of the first and the last digits?

4
Everaise Academy (2020) Math Competitions I

§6.4 Complementary Counting

Casework and constructive counting aren’t always the best approaches. There can be times
when there are just too many possibilities to consider. But, if we know the total amount of
cases, then we can just count the cases that we don’t want, and then subtract this from the total
number of cases to count what we do want. You can think of this as taking the complement of
a subset.

Example 6.14. How many 4 digit binary strings have at least one 0, assuming a binary
string can start with a 0?

Solution. It would be messy to count this directly, as we’d have to keep track of how many
0’s there are and their placement in the overall string. However, note there are 24 = 16 total 4
digit binary strings as each string either has 0 or a 1, and there only exists 1 string without a 0
(namely, 1111). Thus, there are 16 − 1 = 15 binary strings with at least one 0.
Putting this in a set perspective, our set was every 4 digit binary string. The condition was
having at least one 0, so our original set was split into two subsets: ones with at least one 0
and ones without a 0. 

Example 6.15. (2006 AMC 10A) How many four-digit positive integers have at least one
digit that is a 2 or a 3?

Solution. Again, there’s no need to go through all the cases where we have perhaps one 2
no 3s, two 2 and one 3, etc. when we can just count the numbers without a 2 or 3. We have
9 · 10 · 10 · 10 = 9000 four-digit numbers, and the amount of four-digit integers without a 2 or 3
is 7 · 8 · 8 · 8 = 3584, so our answer is simply 5416. 

We will finish off complementary counting with a harder example.

Example 6.16. (2017 AIME II) Find the number of subsets of {1, 2, 3, 4, 5, 6, 7, 8} that are
subsets of neither {1, 2, 3, 4, 5} nor {4, 5, 6, 7, 8}.

Solution. We first need to have a way to count the number of subsets. If we want to make a
subset, either an element from the original set is in the subset or it isn’t in it, so this means we
have 2n subsets for a set with n integers.
So, with complementary counting, we want to count the number of subsets that are subsets
of {1, 2, 3, 4, 5} or {4, 5, 6, 7, 8} and subtract them from 28 = 256, the total number of subsets
of the original set.
Each set has 5 elements, so each set has 25 subsets and combined gives us 64 subsets. However,
since {4, 5} overlap between the two sets, we inevitably overcount the total number of subsets
we need to subtract. Since {4, 5} has 4 subsets, our final answer is

256 − 64 + 4 = 196.

Remember that at the end of the day, you still need extensive constructive counting skills
to really nail counting problems, but complementary counting is just another strategy to make

5
Everaise Academy (2020) Math Competitions I

your constructive counting more efficient. In general, if you feel that counting the complement
involves less cases, you should go for this approach.
Exercise. (2008 AMC 12B) A parking lot has 16 spaces in a row. Twelve cars arrive, each of
which requires one parking space, and their drivers chose spaces at random from among the
available spaces. Auntie Em then arrives in her SUV, which requires 2 adjacent spaces. What
is the probability that she is able to park?

§6.5 Committees
Committees are an extension of constructive counting with a crucial distinction: the order in
which the elements are chosen doesn’t matter. Through our exploration of committees, we will
also investigate binomial coefficients, which are essential for intermediate counting problems
and the rest of this week.
Consider the following question:

Example 6.17. How many ways are there to place 5 different books on a bookshelf? Can
you find a general formula for the number of ways to place n books?

Solution. We can use constructive counting to solve this problem. There are 5 choices for the
first book, 4 for the second, 3 for the third, 2 for the fourth, and 1 remaining book for the fifth.
Thus, there are

5 × 4 × 3 × 2 × 1 = 120

ways to place the 5 books. For n books, we use the same logic and generalize it. There are n
ways to pick the first book, n − 1 for the second, and so on, until we get to the last book, for
which we have 1 choice. Thus, the number of ways to order n books is

n × (n − 1) × (n − 2) · · · 2 × 1

Notice that this is the product of all positive integers less than or equal to n. We can use the
factorial sign, an “!”, to denote this. For a positive integer n,

n! = n × (n − 1) × (n − 2) · · · 2 × 1

Note that for all positive integers n, we have (n + 1)! = (n + 1) · n!. Thus, we define 0! to be 1.
Factorials show up very often in counting because n! counts the number of ways to arrange
n objects in a certain order. Each ordering can also be referred to as a permutation. For
example, 1,3,2,4,5 is a permutation of the digits 1,2,3,4,5.
Exercise. Show that the number of ways to order any m objects from a set of n objects, where
m ≤ n, is m!
n! .

Now consider a similar, but different question:

Example 6.18. The Math Club of Newton High School has 8 board members. They must
elect a committee of 3 people to serve as directors. How many ways are there to do this?
Note that there is no difference between the director positions.

6
Everaise Academy (2020) Math Competitions I

Solution. At first glance, this may seem like a simple constructive counting problem that we
have already done. There are 8 choices for the first director, 7 for the second, and 6 for the
third– so our answer should be 8 × 7 × 6.
However, we have overcounted; the 3 director positions are not distinct. For example, choosing
Aspara, then Brussel, then Kale as our directors is the same choosing Kale, then Brussel, then
Aspara, but we counted both in our previous calculation.
So, how many times have we overcounted? Suppose one of our possible committees is Aspara,
Brussel, and Kale. There are 3! = 6 ways to order the following 3 people, each of which were
counted in our original calculation. In other words, we counted each possible committee 6 times.
Thus, the number of possible committees is equal to
8×7×6
= 56
3×2×1


Picking a few elements from a set of elements where the order in which they are chosen doesn’t
matter is a common theme in combinatorics problems. Note that if a problem uses words such
as “committee” or “combination,” it likely implies that order does not matter.

Example 6.19. Find a general formula for the number of ways to pick a committee of m
people from a pool of n people.

Solution. First, we must pick m people in a constructive counting manner, for which there are

n × (n − 1) · · · × (n − m + 1)

ways to do so. We can rewrite this as

n!
(n − m)!
Then, we must divide to compensate for overcounting. We counted each possible committee m!
times because we counted every possible permutation of each committee. Therefore, we must
divide by m!. The number of ways to pick a committee of m people from a pool of n people is
n!
(n − m)!m!


This notation is slightly messy, so it is convention to use the following notation. For nonneg-
ative integers n, m where n ≥ m,

 
n n!
=
m (n − m)!m!
We call this combination notation
because the each unique committee that can be put together
n
is a combination. However, m is also verbally referred to as “n choose m” because it is the
number of ways you can choose m elements from a set of n elements with no regards to the
order in which they are chosen.

7
Everaise Academy (2020) Math Competitions I

Remark. When simplifyinh with fractions involving factorials, make sure to cancel out
as many factors as possible between the numerator and denominator before multiplying
anything.

Now let’s look at an AMC problem involving combinations.

Example 6.20. (2017 AMC 10A) At a gathering of 30 people, there are 20 people who all
know each other and 10 people who know no one. People who know each other hug, and
people who do not know each other shake hands. How many handshakes occur within the
group?

Solution. We can split the handshakes into two categories: those that occur between the 10
people that do not know anyone and

those that occur between the group of 20 people and the
group of 10 people. There are 102 = 45 handshakes in the first category because each possible
combination of 2 people produces 1 handshake. There are 20 × 10 = 200 handshakes in the
second category because there are 20 choices for one participant of the handshake and 10 for
the other. Thus, there are 200 + 45 = 245 handshakes in total. 

Exercise. Solve this using complementary counting.

Example 6.21. We want to choose two disjoint committees of 4 people from a class of 12.
How many ways can we do this, if the committees are distinct?

Solution. Note we can choose the first committee in 12 8

4 ways and the second
8
ways (since
in 4
12
the first 4 students are already in the first committee). Thus the answer is 4 · 4 = 34650. 

Now, we will move on to a slightly more complicated application of combinations.

Example 6.22 (2013 AIME I). Melinda has three empty boxes and 12 textbooks, three of
which are mathematics textbooks. One box will hold any three of her textbooks, one will
hold any four of her textbooks, and one will hold any five of her textbooks. If Melinda packs
her textbooks into these boxes in random order, the probability that all three mathematics
textbooks end up in the same box can be written as m n , where m and n are relatively prime
positive integers. Find m + n.

Solution. First, we will use constructive counting with combinations to find the total number
12


of ways to distribute the 12 books. There are a total of 3 ways to place a combination of
9


three books in the first box and 4 ways to place a combination of 4 books from the remaining
9 books into the second box. The 5 remaining books must be placed into the last box, so there
is exactly 1 way to do so. Thus, there are a total of
   
12 9 12 · 11 · 10 · 9 · 8 · 7 · 6
· ·1=
3 4 3·2·1·4·3·2·1
= 11 · 10 · 9 · 4 · 7

ways to place distribute the 12 books.

Now, we count the number of ways to distribute the books when all 3 math books are in the
same box. We will do this with casework:

8
Everaise Academy (2020) Math Competitions I

• Case 1: The 3 math books end up in the box with three books. There are 9


4 = 126 ways
to distribute the remaining books.
• Case 2: The 3 math books end up in the box with four books. Now, we must pick one
other book to be in the same
box

and then distribute the remaining books among the
other two boxes. There are 91 83 = 504 ways to do so.
• Case 3: The 3 math books end up in the box with five books. Now, we must pick two
other books to be in the same
box

and then distribute the remaining books among the
9 7
other two boxes. There are 2 3 = 1260 ways to do so.
Adding them all up, there are a total of 1890 ways to distribute the books if all three math
books are together. Thus, the probability that all three math books end up together is equivalent
to
1890 3
= .
11 · 10 · 9 · 4 · 7 44
Therefore, m + n = 3 + 44 = 47. 

Exercise.
7


(a) Compute 3 .
7


(b) Compute 4 .
9


(c) Compute 2 .
9


(d) Compute 7 .
n n



Exercise. Show that m = n−m is true, both in terms of counting objects and algebraically.
Now we’ll move onto a rather hard example involving combinations that makes use of the
idea of a one-to-one correspondence. We’ll return to this idea in later handouts, as it will come
in handy for proving a lot of things in combinatorics.

Example 6.23. For nonnegative integers a and b with a + b ≤ 6, let T (a, b) = a6 6b a+b


6

.
Let S denote the sum of all T (a, b), where a and b are nonnegative integers with a + b ≤ 6.

Solution. Directly summing the product of these binomial coefficients looks like it would be an
algebraic mess. In problems involving obvious counting-related expressions like combinations,
its often a good idea to think of a counting argument, or double counting. A counting
argument is essentially a way of showing that two expressions count the same thing, which then
means the two expressions are equal.
Let’s look at each individual term T (a, b) first. What could this possibly represent?

We can think of having 3 sets with 6 objects each. We pick a from the first set, b from the
second set, then a + b from the third set.
Here’s the key observation: since
   
6 6
= ,
a+b 6−a−b
we can say we are choosing 6 − a − b objects from the third set instead of a + b objects.
Then we are choosing a total of (6 − a − b) + a + b = 6 objects out of a set of 18 objects. In
fact, we are choosing all such possible combinations of 6 objects out of 18 since a + b can vary
from 0 to 6.
This is just 18


6 . After some computation, this gives an answer of 18564. 

9
Everaise Academy (2020) Math Competitions I

Note how we showed both 18

6 and the expression given in the problem counted the same
abstract idea. Such arguments are the essence of combinatorics. Counting arguments won’t
appear very often on math competitions, but they are still very worthwhile to learn.
Exercise 6.24. Show that
n  
X n
= 2k
k
k=0
using a counting argument. Hint: Think about sets and subsets.

§6.6 Homework Problems

Problem 6.1. [3] [2009 AMC 10B] How many 7-digit palindromes (numbers that read the
same backward as forward) can be formed using the digits 2, 2, 3, 3, 5, 5, 5?

Problem 6.2. [4] [2004 AIME II] How many positive integers less than 10,000 have at most
two different digits?

Problem 6.3. [5] [2012 AIME II] At a certain university, the division of mathematical sciences
consists of the departments of mathematics, statistics, and computer science. There are two
male and two female professors in each department. A committee of six professors is to con-
tain three men and three women and must also contain two professors from each of the three
departments. Find the number of possible committees that can be formed subject to these
requirements.

Problem 6.4. [5] [2017 AMC 10A] Alice refuses to sit next to either Bob or Carla. Derek
refuses to sit next to Eric. How many ways are there for the five of them to sit in a row of 5
chairs under these conditions?

Problem 6.5. [7] [2006 AIME II]

Let (a1 , a2 , a3 , . . . , a12 ) be a permutation of (1, 2, 3, . . . , 12) for which a1 > a2 > a3 > a4 >
a5 > a6 and a6 < a7 < a8 < a9 < a10 < a11 < a12 . An example of such a permutation is
(6, 5, 4, 3, 2, 1, 7, 8, 9, 10, 11, 12). Find the number of such permutations.

Problem 6.6. [8] [2013 AMC 10A] Central High School is competing against Northern High
School in a backgammon match. Each school has three players, and the contest rules require
that each player play two games against each of the other school’s players. The match takes
place in six rounds, with three games played simultaneously in each round. In how many
different ways can the match be scheduled?

Problem 6.7. [10] [2018 PUMaC] In an election between A and B, during the counting of the
votes, neither candidate was more than 2 votes ahead, and the vote ended in a tie, 6 votes to 6
votes. Two votes for the same candidate are indistinguishable. In how many orders could the
votes have been counted? One possibility is AABBABBABABA.

Problem 6.8. [7] [2016 AMC 10B]A set of teams held a round-robin tournament in which
every team played every other team exactly once. Every team won 10 games and lost 10 games;
there were no ties. How many sets of three teams {A, B, C} were there in which A beat B, B
beat C, and C beat A?
Provide a detailed solution, or explain what progress you made, even if you didn’t
get an answer.

10
7 More Counting Methods
In this handout, we’ll go over a few specialized counting methods that appear often on the
AMC and AIME.

§7.1 Grid Paths

Let’s say we have a person at point A on a 2-dimensional grid trying to reach another point B
on the grid. However, the person can only move across adjacent points on the grid. How many
different paths are there from A to B?

This idea of counting grid paths is very common. Now we’ll dive into some examples to see
how to solve these kinds of problems.

Example 7.1. Bob is stuck at home, which is located at the origin of an arbitrary Cartesian
coordinate system. He wants to buy some food at the supermarket, which is 3 units up
and 4 units left of his home (so at (4, 3)). Because he can only walk across sidewalks, Bob
can only go in directions parallel to the axes. Also, because Bob doesn’t want to waste
time, he wants to travel exactly 7 units. How many ways can he walk to the supermarket?

First things first: draw a diagram.

Now, note that Bob can only go one unit at a time, and he can only go in the up and right
direction since he can’t travel more than 7 units.
There are several possibilities. He can go up 3 units and left 4 units, go left 4 units then up
3 units, or some path in between. Let’s model a move up as U and a move right as R. Since
Bob can move up exactly 3 times and he can move right exactly 4 times, his sequence of moves
can be any permutation of 3U ’s and 4R’s.

We’ve done this before! How many ways can you arrange 3 U s and 4 Rs? We can just choose
which of the seven positions will contain a U , so our answer is
 
7
= 35
3

Exercise. Show that the original problem and our new problem of arranging U ’s and R’s are in
a one-to-one correspondence.
There can also be restrictions that can make the problem trickier:

1
Everaise Academy (2020) Math Competitions I

Example 7.2. Assume the same problem as last time, but there is a police station at (2, 1)
and Bob does not want to be at that point. How many paths can Bob take en route to the
supermarket?

Solution. Since it’s more straightforward to count paths crossing through the police station,
we’ll use complementary counting.
To do this, we need to find how many paths there are from B to P and multiply this by
the number of paths from P to S; since every path we’re looking for passes through the police
station, it must be a composition of a path going to the police station and a path leading out.
There are 3 × 3 = 9 paths that cut through the police station. (Verify the results yourself.)
Thus, there are
35 − 9 = 26
ways to avoid the police department. 

We can also have irregular grid patterns.

Example 7.3. Bob moves to another city, where the grid pattern is shown below. If he
only goes up and right, how many paths does Bob have to get to the supermarket?

Solution. It appears that our current method cannot handle this complex system. When all
else fails, we can use a recursive approach.
Label a number at every point that signifies how many paths lead up to that point. This
number can be determined by adding the numerical values of the points immediately below and
to the left of the point, since a point can be reached by taking any path to the point below then
moving up, or by taking any path to the point to the left then moving right, and in no other
way since the only possible moves are up or right.
We will illustrate how this works. First, we can fill in the edges with 1 since there’s only
one possible way to travel to any of the edge points: We can fill in the rest by adding the two

2
Everaise Academy (2020) Math Competitions I

numbers that lead to it (the number to the left and the number below). Verify that the numbers
we filled in are correct. The answer is 17.


Now, let us apply this to an actual competition problem.

Example 7.4 (1992 AIME). In a game of Chomp, two players alternately take bites from a 5-
by-7 grid of unit squares. To take a bite, a player chooses one of the remaining squares, then
removes (“eats”) all squares in the quadrant defined by the left edge (extended upward)
and the lower edge (extended rightward) of the chosen square. For example, the bite
determined by the shaded square in the diagram would remove the shaded square and the
four squares marked by ×. (The squares with two or more dotted edges have been removed
form the original board in previous moves.)
The object of the game is to make one’s opponent take the last bite. The diagram shows
one of the many subsets of the set of 35 unit squares that can occur during the game of
Chomp. How many different subsets are there in all? Include the full board and empty
board in your count.

Solution. We strongly advise you to experiment with different subsets of the board first, and
see if you can relate this problem to grid paths without reading further.

At first, it seems that there are no “paths” in this game. We’ll have to find them ourself.
Instead of paths, we can think about lines of division. Note that the shaded square in each
circumstance actually divides the grid into two parts: One with unchomped squares and one
with chomped and to-be-chomped squares (the latter of which forms a rectangle).
We can configure a “path” by tracing the border of the unchomped squares. Specifically, we
can do the following:
Assume that unit squares not shown (outside of the rectangle) are chomped already. Start
at the upper left corner and only move along segments that are shared by both a chomped and
an unchomped square (crossed and shaded squares are unchomped).

3
Everaise Academy (2020) Math Competitions I

We will illustrate with the example above. If the lower left corner of the rectangle is the
origin, and each square is a unit square, our path looks like this:

(0, 5) → (3, 5) → (3, 4) → (4, 4) → (4, 2) → (7, 2) → (7, 0)

Convince yourself that finding the number of paths that only go right and down is in a one-
to-one correspondence to the above problem. When you do, it should be clear that the answer
is  
7+5
= 792
5


As with most hard counting problems, the most important step was realizing we could use a
simple method that at first didn’t seem to apply to the problem. You’ll need to think of clever
ways to apply the material in the handout to the homework problems similarly.

§7.2 Stars and Bars

Stars and Bars is a method of counting essential to problem solving.
We begin with a motivating example:

Example 7.5. There are three dogs and you have ten identical toys. How many ways are
there to distribute all of the toys to the dogs?

Solution. This seems like a hassle to count directly; we’d have to split by cases depending on
how many toys go to each dog.
Instead, we can create a depiction of the ten toys as circles: How can we divvy up these toys?

We can start by using two dividers between the circles to distinguish between the three dogs.
Each divider will be represented by a single line.

In the above example, we gave the first dog 2 toys, the second dog 0 toys and the third dog
with 8 toys.
Amazingly, just by moving the dividers, we can count the number of ways to distribute the
toys.

4
Everaise Academy (2020) Math Competitions I

Exercise. Show that the original problem and the new problem of ordering dividers and circles
are in a one to one correspondence.
We can find the answer as follows:
   
10 + 2 12
= = 66
2 2

We can apply this to algebraic equations too.

Example 7.6. How many solutions are there to the equation

a + b + c + d = 14

where a, b, c, and d are non-negative integers?

Solution. We can think about this as taking 14 identical objects and distributing them to four
boxes (i.e. our variables). This is the same situation as before. If we use 3 dividers to distinguish
the 4 variables, we see that there are
   
14 + 3 17
= = 680
3 3

Let us solve a slightly different problem now:

Example 7.7. How many solutions are there to the equation

a + b + c + d = 14

where a, b, c, and d are positive integers?

Solution. This time, all integers must be positive. This restriction is easy to take care of. Of
the 14 objects, allocate one object to each of the four boxes automatically.
Now we’ve just reduced this problem to distributing 14 − 4 = 10 objects to 4 boxes; it’s just
the case with nonnegative integers again, but with 10 instead of 14.
Thus, the answer is simply    
10 + 3 13
= = 286
3 3


Now let’s apply this to a hard problem:

Example 7.8 (2011 AIME II). Ed has five identical green marbles, and a large supply of
identical red marbles. He arranges the green marbles and some of the red ones in a row
and finds that the number of marbles whose right hand neighbor is the same color as
themselves is equal to the number of marbles whose right hand neighbor is the other color.
An example of such an arrangement is GGRRRGGRG. Let m be the maximum number of

5
Everaise Academy (2020) Math Competitions I

red marbles for which such an arrangement is possible, and let N be the number of ways
he can arrange the m + 5 marbles to satisfy the requirement. Find N .

Solution. A simpler way to put the problem condition is that the number of color continuations
(GG and RR) found in the row must be the same as the number of color switches (GR and
RG). Because of this, if we wanted to add as many red marbles as we could, we would be
limited by the number of color switches. Therefore, we want this limit of color switches to be
rather high.
The maximum number of color switches is 10, found in the sequence

RGRGRGRGRGR.

We have 0 color continuations here. Now, we should add more red marbles to the end to increase
the amount of these continuations. We end up with 16 red marbles.
Now we have to order these to keep the equality of continuations and switches. We can place
these red marbles anywhere but since they are indistinguishable, we have to use stars and bars
to count them. Note that we have 10 stars, which are the unused red marbles we wish to place,
and 5 bars, which are the green marbles. This gives us an answer of
 
10 + 5
= 3003.
5


Exercise 7.9. How many solutions are there to the equation

a + b + c + d ≤ 14

where a, b, c, and d are non-negative integers? (Note: Be creative! When you figure out the
solution, the computation is almost exactly the same as in Example 7.6.)

§7.3 Symmetry
From previous counting problems, you’ve already seen the possibility of overcounting, or count-
ing the same object multiple times. Sometimes two objects can be the “same”, or symmetric, if
they have been transformed in some way that preserves their shape. The most common types
of symmetry are reflectional symmetry and rotational symmetry.

Definition 7.10. A reflection is a mirrored image across a line (in two dimensions). In
three dimensions, reflections are mirrored images across planes. Two objects have reflectional
symmetry if they are reflections of each other.

For example, an isosceles triangle has reflectional symmetry, where the line of symmetry is
the altitude to the side of different length.

Definition 7.11. An object has rotational symmetry if it can be rotated through by an

angle 0 ≤ θ < 2π and remain the same. We say an object has n-fold symmetry iff there are n
distinct angles 0 ≤ θ < 2π that it can be rotated and stay the same.

Consider a square; it clearly has 4-fold symmetry, as we can rotate by 0, π/2, π or 3π/2 and
end up with the same image. Similarly, any regular n-gon will have n-fold symmetry. It is
visually intuitive that a regular n-gon has d-fold symmetry iff d is a positive divisor of n. For
example, an octagon also has 4-fold and 2-fold symmetry.

6
Everaise Academy (2020) Math Competitions I

Exercise. Prove that any object with n-fold symmetry also has d-fold symmetry whenever d is
a positive divisor of n.
From now on, whenever we refer to n-fold symmetry, n is maximal: so, for example, we will
only refer to a square as having 4-fold symmetry.

In counting problems with reflections and rotations, the strategy is often to use casework so
we avoid overcounting certain configurations.

Example 7.12 (2017 AMC 10B). In the figure below, 3 of the 6 disks are to be painted blue,
2 are to be painted red, and 1 is to be painted green. Two paintings that can be obtained
from one another by a rotation or a reflection of the entire figure are considered the same.
How many different paintings are possible?

Solution. Let’s label the disks as follows:

There aren’t that many possible cases because reflections and rotations are equivalent, so we
can choose a location for the three blue disks and then figure out the number of paintings for
each location. After playing around for a while, it’s easy to see there are 6 distinct locations
for the blue disks with respect to reflection and rotation. We’ll do cases based on each set of
locations.

Case 1 (1, 2, 3)
All that remains is the bottom row; we can either place the green disk in the middle or on
either edge of the bottom row (placing it on either edge is equivalent by reflection). So we have
2 paintings.

Case 2 (1, 2, 4)
We can put the green disk in the 6 spot and the red disks in 3 and 5. We can also put the
green disk in 3 or 5; both are the same by reflection, so we get 2 more paintings.

7
Everaise Academy (2020) Math Competitions I

Case 3 (1, 2, 5)
Here every position of the green disk is distinct so we have 3 paintings in this case.

Case 4 (1, 2, 6)
Again, each position of the green disk is unique, so we have 3 paintings again.

Case 5 (1, 4, 6)
This leaves the center 3 disks. Putting the green disk in any of the three remaining spots and
the red disks in the remaining spots is equivalent by rotation, so we have 1 distinct painting.

Case 6 (2, 3, 5)
Any placement of the green disk and red disks is the same by rotation so there is 1 painting.

Adding up the 6 cases gives us a total of 2 + 2 + 3 + 3 + 1 + 1 = 12 paintings. 

Example 7.13 (2006 AIME). There is an unlimited supply of congruent equilateral triangles
made of colored paper. Each triangle is a solid color with the same color on both sides of
the paper. A large equilateral triangle is constructed from four of these paper triangles.
Two large triangles are considered distinguishable if it is not possible to place one on the
other, using translations, rotations, and/or reflections, so that their corresponding small
triangles are of the same color.
Given that there are six different colors of triangles from which to choose, how many
distinguishable large equilateral triangles may be formed?

Solution. Label the top triangle “1”, the bottom-right triangle “2” and the bottom-left triangle
“3”. Let’s see what we can do using reflections and rotations.
We see that by using rotation, we can make the top triangle either “1”, “2”, or “3”. Then
the two bottom numbers can be switched by reflecting across the line of symmetry from the top
vertex to the base. Reflections and rotations cannot change the position of the center triangle.
Thus, assuming the center triangle’s color is fixed, only the combination of colors we use for
the outer triangles matters, not their order. That is, any two triangles with the same center
triangle color, and the same multiset(i.e. set including repeats) of colors for the outside triangles
are the same. So we can split this into cases based on how many distinct colors there are for
the outer triangles.

8
Everaise Academy (2020) Math Competitions I

Case 1 (3 distinct colors) There are 63 choices for the outer colors since we have 6 colors to

choose
from and order does not matter. The center triangle can be any of the 6 colors, so we
get 6 63 = 120.

Case 2 (2 distinct colors) If there are 2 distinct colors then one color appears twice and the
other appears once. There are 6 possibilities for the color that appears twice and 6 − 1 = 5 for
the color that appears once since the two colors must be distinct. Finally, the center triangle
can be any of the 6 colors, giving us 6 · 6 · 5 = 180 triangles for this case.

Case 3 (1 color) We choose one color for all the outside triangles and one for the center
triangle, yielding 6 · 6 = 36.
Adding up the cases, we get an answer of

120 + 180 + 36 = 336.

Example 7.14. How many ways can we color the vertices of a regular octagon either red,
blue, or green, where colorings are considered the same if they are rotations of each other?

Solution. Let’s start by considering the total amount of colorings without the rotation condi-
tion, which is 38 . What are we overcounting?
For a coloring with no symmetry, we know we overcount it by a factor of 8 as each of its 8
rotations is counted. However, colorings with symmetry are different; if a coloring has n-fold
symmetry, then we only count it 8/n times (since its coloring pattern repeats 8/n times if we
rotate it 8 times by an angle of π/4).
We know that the coloring can have 8-fold, 4-fold, or 2-fold symmetry, or no symmetry. Let’s
consider each of these cases independently, initially considering rotations distinct.

Case 1 (8-fold symmetry)

If we have 8-fold symmetry, then every vertex must be colored the same, so there are 3
colorings in this case.

Case 2 (4-fold symmetry)

An octagon with 4-fold symmetry is determined by the coloring of any two adjacent vertices
(the 2-vertex coloring simply repeats 3 more times), so we have a count of 32 = 9. But by doing
this we also counted the colorings with 8-fold symmetry, so we have to subtract them out to
get a total of 9 − 3 = 6 in this case.

Case 3 (2-fold symmetry)

Here we need a coloring of 8/2 = 4 adjacent vertices that will be used twice. We can color 4
vertices in 34 = 81 ways, but we need to subtract the colorings with 4-fold and 8-fold symmetry,
so we have a total of 81 − 6 − 3 = 72.

9
Everaise Academy (2020) Math Competitions I

Case 4 (no symmetry)

The total possible number of colorings is 38 , and subtracting out colorings with symmetry,
we get 38 − 72 − 6 − 3 = 6480. Now for each case with n-fold symmetry, we overcounted by a
factor of 8/n, so we have a final answer of
3 6 72 6480
+ + + = 834.
1 2 4 8


Exercise. Try solving this problem where the octagon is replaced with a regular 9-gon.
Symmetry will come up often in counting problems that have some kind of visual or geomet-
rical context, so you should always consider symmetry when avoiding overcounting. Symmetry
can also be useful in speeding up casework; if some cases are symmetric, you can group them
together to save time.

§7.4 Homework Problems

Problem 7.1 (2018 AMC 10A). [5]When 7 fair standard 6-sided dice are thrown, the proba-
bility that the sum of the numbers on the top faces is 10 can be written as
n
,
67
where n is a positive integer. What is n?

Problem 7.2 (2019 AMC 8). [5] Alice has 24 apples. In how many ways can she share them
with Becky and Chris so that each of the three people has at least two apples?

Problem 7.3 (2010 AMC 12A). [7]A 16-step path is to go from (−4, −4) to (4, 4) with each
step increasing either the x-coordinate or the y-coordinate by 1. How many such paths stay
outside or on the boundary of the square −2 ≤ x ≤ 2, −2 ≤ y ≤ 2 at each step?

Problem 7.4 (2014 AMC 10B). [7] The numbers 1, 2, 3, 4, 5 are to be arranged in a circle. An
arrangement is bad if it is not true that for every n from 1 to 15 one can find a subset of the
numbers that appear consecutively on the circle that sum to n. Arrangements that differ only
by a rotation or a reflection are considered the same. How many different bad arrangements
are there?

Problem 7.5 (2018 AMC 12A). [8] A scanning code consists of a 7 × 7 grid of squares, with
some of its squares colored black and the rest colored white. There must be at least one square
of each color in this grid of 49 squares. A scanning code is called symmetric if its look does
not change when the entire square is rotated by a multiple of 90◦ counterclockwise around its
center, nor when it is reflected across a line joining opposite corners or a line joining midpoints
of opposite sides. What is the total number of possible symmetric scanning codes?

Problem 7.6 (2012 AMC 10A). [8] A 3 × 3 square is partitioned into 9 unit squares. Each unit
square is painted either white or black with each color being equally likely, chosen independently
and at random. The square is then rotated 90 ◦ clockwise about its center, and every white
square in a position formerly occupied by a black square is painted black. The colors of all other
squares are left unchanged. The probability the grid is now entirely black can be represented
as pq , where p and q are relatively prime positive integers. What is p + q?

10
Everaise Academy (2020) Math Competitions I

Problem 7.7 (2013 AIME I). [7] In the array of 13 squares shown below, 8 squares are colored
red, and the remaining 5 squares are colored blue. If one of all possible such colorings is chosen
at random, the probability that the chosen colored array appears the same when rotated 90
degrees around the central square is n1 , where n is a positive integer. Find n.

Problem 7.8 (2007 AIME I). [10] In a 6 x 4 grid (6 rows, 4 columns), 12 of the 24 squares
are to be shaded so that there are two shaded squares in each row and three shaded squares in
each column. Let N be the number of shadings with this property. Find the remainder when
N is divided by 1000.

Problem 7.9 (2003 AMC 12A). [7] Objects A and B move simultaneously in the coordinate
plane via a sequence of steps, each of length one. Object A starts at (0, 0) and each of its steps is
either right or up, both equally likely. Object B starts at (5, 7) and each of its steps is either to
the left or down, both equally likely. The probability that the objects meet can be represented
as m/2n for positive integers m and n where m is not divisible by 2. What is m + n?
Provide a detailed solution, or explain what progress you made, even if you
didn’t get an answer.

11
8 PIE, States & Recursion

§8.1 Recursion
In this section, we’ll go over recursion. Recursion is a common tool on the AIME that should
be considered for any hard counting problem. This concept is very challenging, so you should
take extra effort to try and understand each step in our example problems.
We’ll start with a motivating example.

Example 8.1 (Classical). Suppose that we have a ten-step staircase. Liam starts at the
bottom, and every second, he may either climb one step or climb two steps. How many
ways can he ascend the staircase? (One way might be to climb two steps, then one step,
then two steps, then one steps five more times.)

Solution. This problem seems annoying to do with standard methods—we might be able to
do casework on the number of two-step climbs Liam attempts, but this is really computational.
To begin solving most math problems, it helps to consider small cases. If we deal with a
one-step staircase, there’s only one way for Liam to climb it. For a two-step staircase, Liam can
either climb one step twice or climb two steps, which makes for two ways. In a similar manner,
we can count the first few cases:

Height of Staircase Ways to Climb

1 1
2 2
3 3
4 5
5 8

(You should confirm for yourself that these numbers are correct!)
At this point, a pattern begins to emerge: the number of ways to climb the staircase appears
to follow the Fibonacci Sequence.
To see why this might be the case, we think about how the Fibonacci sequence works in the
first place: each successive term of the sequence is built on the previous two terms. Suppose
that Liam currently stands on the nth step of the staircase. He must have gotten there from
the (n − 1)th step or the (n − 2)th step, because he can only climb one or two steps at a time.
Therefore, the number of ways for Liam to climb to step n is equal to the number of ways he
can climb to step n − 2 and then climb two steps, added to the number of ways he can climb
to step n − 1 and then climb one step. (If you’re right now asking why he can’t climb to n − 2
and take a single step to n − 1, this case is addressed in the n − 1-case already!) Expressed
mathematically, if Si denotes the way for Liam to climb an i-step staircase, we have

Sn = Sn−1 + Sn−2 .

Thus the number of distinct ways to ascend n steps is equal to the number of ways to ascend
n − 1 steps added to the number of ways to ascend n − 2 steps — which is exactly the Fibonacci

1
Everaise Academy (2020) Math Competitions I

sequence, expressed differently. Now the remaining computation is easy: finishing the table
yields that if the staircase has 10 steps, then there are 89 distinct ways to ascend it. 

It would have been really difficult to do pure casework on the number of two-step ascensions,
or some other method! The idea of recursion here is that the process of climbing some number
n of steps — in this case, 10 — is linked to the process of climbing n − 1 steps or n − 2 steps.
You should try another classical exercise before moving on.
Exercise 8.2 (Classical). How many ways are there to fill in a 2 × 10 grid with 1 × 2 dominos?
Hints: Relate 2 × 10 to 2 × n for some smaller n than 10 by thinking about horizontal or vertical placement at
the end!

Recursion, simply put, is the process of using smaller cases of a problem to build up to a
solution for larger cases. Recursion in combinatorics is similar to having a recursive sequence
(e.g. we found the Fibonacci sequence in the previous example), except the ”sequence” is a
function f (n) that counts the nth case of a problem; for example, f (n) often counts the number
of strings of length n satisfying certain conditions, but it could also count the number of ways
a traveling object can reach a certain location in n steps and a variety of other things.

To solve a problem using recursion, we have to find f (n) in terms of solely f (n − 1), f (n −
2), . . . f (1). Then we can make a table of values and compute values of f (n) recursively. Re-
cursion has many parallels with induction in that we have base cases and we use some kind of
relation to go from the base cases to case n, except we often do actual computation rather than
finding some general formula.

Recursion can also be used in the context of probability, where our recursive function might
be defined as the probability an object has a certain state after n moves, or the probability it
will reach a point P . This often happens with problems involving a moving object with certain
limitations.

Recursion is best illustrated by example. Let’s look at some competition problems.

Example 8.3 (2019 OMO). Susan is presented with six boxes B1 , . . . , B6 , each of which is
initially empty, and two identical coins of denomination 2k for each k = 0, . . . , 5. Compute
the number of ways for Susan to place the coins in the boxes such that each box Bk contains
coins of total value 2k .

Solution. Say the number of ways to solve the problem for n boxes is F (n). We instead solve
the problem for general n (motivated by trying small cases).
We look at the number of ways we can distribute the coins of value 2n−1 . We can either put
both of them in Bn or one of them in Bn and the other in Bn−1 . Notice these are the only two
possibilities since 2n−1 is too big to go anywhere else.
Now notice that in the first case, it remains to fill boxes with capacities 21 , 22 , . . . , 2n−1 with
two coins each of values 20 , 21 , . . . , 2n−2 . In the second case, it’s actually the same, since we
need 2n−1 more to fill box Bn so it is effectively a box with capacity 2n−1 .
Thus, we’ve reduced the n case to the n − 1 case. Since there are two ways to distribute the
coins of value 2n−1 , F (n) = 2F (n − 1).

2
Everaise Academy (2020) Math Competitions I

Obviously F (1) = 1. So F (n) = 2n−1 and F (6) = 32. 

Example 8.4. (2008 AIME I) Consider sequences that consist entirely of A’s and B’s and
that have the property that every run of consecutive A’s has even length, and every run
of consecutive B’s has odd length. Examples of such sequences are AA, B, and AABAA,
while BBAB is not such a sequence. How many such sequences have length 14?

Solution. There are many hints in this problem to use recursion. First of all, it’s a problem
about strings of As and Bs, but the consecutive condition also suggests recursion as what we
can append to a string depends on the previous characters in the string.

So let’s say we have a string of length n satisfying the given conditions. Can we break it
down into a smaller string?
It depends on what the last character of the string is. If the string ends in an A, then we
know that the last two characters are AA since there must be an even number. Then removing
the last two characters should still leave us with a string satisfying the conditions.
This means that every string of length n ending in A is just another string of length n − 2
satisfying the conditions with an AA appended at the end. If the string ends in a B then it
must end in an odd number of Bs. If it only ends in one B, then removing the last character
should give us a string ending in A satisfying the conditions. But if it ends in more than one
B, we have to remove the last 2 characters to ensure the number of consecutive Bs is still odd.
Thus, every string of length n ending in B is either a string of length n − 1 ending in A with
a B appended at the end or a string of length n − 2 ending in B with a BB appended at the
end.

From these observations, it makes the most sense to define the recursive functions a(n) and
b(n) as the number of sequences ending in A of length n and the number of sequence ending in
B of length n, respectively. Then from our earlier statements, we have that
a(n) = a(n − 2) + b(n − 2)
b(n) = a(n − 1) + b(n − 2)
Then if we have the values of a(1), a(2), b(1) and b(2) we should be able to compute up to a(14)
and b(14) recursively. From the problem statement it is easy to see that a(1) = 0, a(2) = 1,
b(1) = 0, and b(2) = 1. Now we compute the rest of the values in a table:
n a(n) b(n) n a(n) b(n)
1 0 1 8 6 10
2 1 0 9 11 11
3 1 2 10 16 21
4 1 1 11 22 27
5 3 3 12 37 43
6 2 4 13 49 64
7 6 5 14 80 92

So the answer is a(14) + b(14) = 80 + 92 = 172 . 

A recursive approach can also be used in the coordinate plane, similar to what we saw in grid
paths yesterday. Here is a fairly standard recursion problem involving probability:

3
Everaise Academy (2020) Math Competitions I

Example 8.5. (2019 AIME I) A moving particle starts at the point (4, 4) and moves until
it hits one of the coordinate axes for the first time. When the particle is at the point (a, b),
it moves at random to one of the points (a − 1, b), (a, b − 1), or (a − 1, b − 1), each with
probability 13 , independently of its previous moves. The probability that it will hit the
coordinate axes at (0, 0) is 3mn , where m and n are positive integers. Find m + n.

Solution. Recursion seems like a promising tool to solve this problem; the particle can only
move left and down, and the problem tells us where the particle can move based on where it
previously was.
Then if we define P (x, y) to be the probability the particle reaches point (0, 0) first if it starts
at (x, y) we have that
1 1 1
P (x, y) = P (x, y − 1) + (x − 1, y) + (x − 1, y − 1)
3 3 3
1
as there is a 3 probability of going one unit up, one unit right, or one unit up and right.
We can make a 2-dimensional grid and recursively compute values (similar to how we counted
grid paths in the previous handout) to get
245 245
P (4, 4) = = 7 ,
2187 3
giving us an answer of 245 + 7 = 252 . 

§8.2 States
States are often used in conjunction with recursions. Essentially, we can use states to represent
locations or situations of an object or character; then the object can move to another state.
Let’s take the basic example: a number line.
Overall, states involve a character moving around a plane, number line, or similar bijection
with probabilities to move to different ’states’. Oftentimes, this is portrayed as a game of some
sorts in the problem statement. Then, we can use a ’state’ to represent where our character is,
either on the number line or their progress in the game, and by doing so we can form equations
out of our states. We first start with a very simple example.

Example 8.6 (2017 AMC 10A). Amelia has a coin that lands heads with probability 13 , and
Blaine has a coin that lands on heads with probability 52 . Amelia and Blaine alternately
toss their coins until someone gets a head; the first one to get a head wins. All coin tosses
are independent. Amelia goes first. Find the probability that Amelia wins.

Solution. In this problem, since Amelia goes first, suppose the probability Amelia win’s given
that she’s tossed k times is pk . We know Amelia has a 1/3 chance to win, but there is also a
2/3 chance she’ll fail to get a heads, so we proceed to the p1 state. Thus,
1 2
p0 = + p1
3 3
Now, we note that at p1 , the only way for Amelia to win is if we can proceed to p2 . This happens
at a 3/5 chance, so we have that
3
p1 = p2 .
5

4
Everaise Academy (2020) Math Competitions I

However, we realize that p2 is exactly the same situation as p0 because it is once again Amelia’s
turn. Hence, going back to our original equation, we have that
1 2 3
p0 = + · p0
3 3 5
1 2
p0 = + p0
3 5
5
p0 =
9
so the probability that Amelia wins is precisely 59 . While this problem doesn’t fully exemplify
the utility of states, it is a nice introductory problem to serve the concepts. 

Now, let’s take a look at a problem with more states.

Example 8.7. (2020 FAIME) Naveen is lost in the maze shown below, where rooms are
labeled with dots, and paths between rooms are labeled with line segments. He is currently
in the room indicated by the arrow. As long as he remains in the maze, Naveen repeatedly
“moves” by randomly choosing one of the paths leading from the room where he is currently
located to the next room. All choices are equally likely and independent of any moves he
has previously made. He stops after he makes a move that leads him out of one of the
maze’s four exits. Compute the expected number of moves Naveen will make before he
leaves the maze.

Solution. At first glance, it seems that states would be clearly be a bad idea. We’re given a
starting point but Naveen could really move anywhere, so are we supposed to use 12 states?
The real trick to this program is utilizing the symmetry to its full extent. As noted before,
we would just have a mess of equations with 12 states. Suppose we break the diagram into 3
groups - the four corners of the large square, the four corners of the middle square, and the
four midpoints of each side. Then, we create 3 states total, one for each group, and this cuts
the amount of equations needed significantly. Before we actually create our states, we need to
know what they are going to represent.
As opposed to the last problem, we want the expected number of moves instead of a proba-
bility. As such, we can make our lives a bit easier by getting rid of the n-th state, and simply
instead let a be the expected number of moves from a center square point, b be the expected
number of moves from a side midpoint, and c be the expected number of moves from a corner
point. Naveen starts at a, so we’ll try to solve for a. In order to do so, we must first set up our
equations.

5
Everaise Academy (2020) Math Competitions I

First, we’ll try to construct the relations between a, b, c. From a, we can only move to 2 other
a points or to a b point. Hence, we can set up the following equation:
2 1
a = 1 + a + b.
3 3
We add a one since we’ve made another move and must add it to our count. Similarly, we can
construct the following equations:
2 1
b=1+ c+ a
3 3
2 1
c=1+ b+ 0
3 3
We let 0 represent the end of the maze, since there are 0 moves left to make once we’ve left the
maze.
Now, we can solve this system of equations without worry. Doing so gives a = 15 , so Naveen
should get out of the maze in 15 moves. 

When working with states, just remember that they are simply another method of recursion
with a more visual component to them.

§8.3 PIE
In combinatorics, we very often have simple brute-force solutions to hard problems that would
work, if not for overcounting. Usually we will have to yield and find another solution, perhaps
proceeding to do tedious casework. However, the Principle of Inclusion-Exclusion gives us
a systematic way of handling overcounting for certain kinds of problems, allowing us to quickly
solve deceptively hard problems.

First, let’s review some set notation. Recall that a set is an unordered collection of objects
(usually numbers) with no repeats.

Definition 8.8. The intersection of two sets A and B, denoted A ∩ B, is the set of elements
they have in common.

Definition 8.9. The union of two sets A and B, denoted A ∪ B, is the set of all elements in
A and B.

Definition 8.10. The cardinality of A, denoted |A|, is the number of elements in A.

Let’s say we have two sets, A and B, and we want to find the cardinality of A∪B. This should
be the number of elements in A, plus the number of elements of B, minus their intersection. So
we have
|A ∪ B| = |A| + |B| − |A ∩ B|.
Now what if we had three sets, A, B, and C? What would the cardinality of A ∪ B ∪ C be?
Let’s draw a Venn diagram to make this clearer.
We can start with considering |A| + |B| + |C|. What do we need to subtract from this? If we
look at the intersections on the Venn diagrams, we see that the elements in A ∩ B, A ∩ C, and
B ∩ C have all been counted twice in this sum since each of these subsets lies in two of A, B,
and C. So we need to subtract out |A ∩ B| + |A ∩ C| + |B ∩ C| to make sure we only count it

6
Everaise Academy (2020) Math Competitions I

once. Now we have |A| + |B| + |C| − |A ∩ B| − |A ∩ C| − |B ∩ C|. But we haven’t considered
A ∩ B ∩ C yet; since these elements lie in all of A, B, C, A ∩ B, A ∩ C, and B ∩ C, we count it
three times, then subtract it out three times, meaning our current sum counts it zero times. So
we need to add in |A ∩ B ∩ C| to get our final answer of
|A| + |B| + |C| − |A ∩ B| − |A ∩ C| − |B ∩ C| + |A ∩ B ∩ C|.
In fact, this pattern continues for more than three sets; if we have n sets A1 , A2 , . . . An , to find
the cardinality of the union of all the sets, we add the cardinalities of all the sets, then subtract
the cardinalities of pairwise intersections, then add the cardinalities of the intersections of all
triplets of sets, then subtract the cardinalities of the intersections of all quadruplets of sets, and
so on (see if you can prove this using the same intuition as in our Venn diagram of three sets).
This result is known as the Principle of Inclusion-Exclusion.

Theorem 8.11 (Principle of Inclusion-Exclusion). If we have n sets A1 , A2 , . . . An , then the

cardinality of A1 ∪ A2 ∪ · · · ∪ An is
n
X X X
|Ai | − |Ai ∩ Aj | + |Ai ∩ Aj ∩ Ak | − · · · + (−1)n−1 |A1 ∩ · · · ∩ An | .
i=1 16i<j6n 16i<j<k6n

If this is confusing to you, just remember that the sum contains the cardinality of the intersec-
tions of every possible combination of any number of the n sets, and the sign of each term is
−1 to the power of the number of sets in the intersection.
We’ll start with a very simple application.

Example 8.12 (Brilliant). At a construction site, Jorge is in charge of hiring skilled workers
for the project. Out of 80 candidates that he interviewed, he found that
• 45 were painters,
• 50 were electricians,
• 50 were plumbers,
• 15 had skills in all three areas, and
• all of them had skills in at least one of these areas.
If he hired everyone who was skilled in exactly 2 areas, how many candidates were hired?

Let A denote set of painters, B denote the electricians, and C denote the plumbers. Then
we want to find |A ∩ B| + |A ∩ C| + |B ∩ C|. Let this sum be x. We know this sum is one of the
terms in PIE, so we can apply that here. We know that
|A| + |B| + |C| − x + |A ∩ B ∩ C| = 80.

7
Everaise Academy (2020) Math Competitions I

We are given the values of |A|, |B|, and |C|, and of |A ∩ B ∩ C|. Plugging them in gives

45 + 50 + 50 − x + 15 = 80

which means that x = 80.

Example 8.13 (2005 AMC 12A). Call a number prime-looking if it is composite but not
divisible by 2, 3, or 5. The three smallest prime-looking numbers are 49, 77, and 91. There
are 168 prime numbers less than 1000. How many prime-looking numbers are there less
than 1000?

Solution. The first thing we want to do is categorize exactly which numbers are prime-looking.
Unfortunately, we don’t know much about prime-looking numbers, but we know what they are
not; numbers that are not prime-looking consist of those that are divisible 2, 3, or 5 or are
prime. So we’ll use complementary counting to find the amount of numbers less than 1000 that
are not prime-looking, then subtract that from 999.

First of all, we’re given that there are 168 prime numbers less than 1000. We also know
that 1 is neither prime nor composite. Then the only remaining not prime-looking numbers are
numbers divisible by 2, 3, or 5. We’ll try counting this. (Note we have to subtract 3 afterwards
since 2, 3, and 5 are primes as well). If we let

S2 = {numbers less than 1000 divisible by 2}

S3 = {numbers less than 1000 divisible by 3}
S5 = {numbers less than 1000 divisible by 5}

then we want to find |S2 ∪ S3 ∪ S5 |. We can use PIE to do this. By PIE,

|S2 ∪ S3 ∪ S5 | = |S2 | + |S3 | + |S5 | − |S2 ∩ S3 | − |S3 ∩ S5 | − |S2 ∩ S5 | + |S2 ∩ S3 ∩ S5 |.

It is quite easy to compute the cardinality of any of these intersections; for example, |S2 ∩ S3 |,
is simply the number of positive integers less than 1000 divisible by 2 × 3 = 6, and |S2 ∩ S3 ∩ S5 |
is simply the number of positive integers less than 1000 divisible by 2 × 3 × 5 = 30. We leave
this computation to the reader, but we should get that

|S2 ∪ S3 ∪ S5 | = 499 + 333 + 199 − 166 − 66 − 99 + 33 = 733.

Subtracting out 3 gives us that there are 730 more not prime-looking numbers that are neither
1 nor prime. So the total number of not prime-looking numbers is 730 + 168 + 1 = 899, giving
us an answer of 999 − 899 = 100. 

That was a pretty standard application of PIE; in fact, you might have done something similar
previously when counting multiples, without even knowing you were using PIE. Make sure you
can rationalize why our use of PIE worked in that problem.

Now let’s move on to a very challenging example that teaches us the importance of being
organized with our sets.

Example 8.14. (Classical) A derangement is a permutation with no fixed points. That is, a
derangement of a set leaves no element in its original place. For example, the derangements

8
Everaise Academy (2020) Math Competitions I

of (1, 2, 3) are (2, 3, 1) and (3, 1, 2), but (3, 2, 1) is not a derangement of {1, 2, 3} because 2
is a fixed point. Let Dn the number of derangements of (1, 2, . . . n). Find Dn .

Solution. Since there’s no easy way to count derangements, we can instead use complementary
counting, and count the number of permutations that have at least one fixed point.
An obvious approach would be to choose one of the n points to fix, then permute the rest of
the n−1 elements. This gives n×(n−1)! = n! permutations. But we are obviously overcounting
here; for example, if a permutation fixes (i.e. does not move) 1 and fixes n, then we’ve counted
it twice.
We can use PIE to deal with this overcounting. We have our n sets A1 . . . An where Aj
consists of all permutations fixing j. Then A1 ∪ A2 · · · ∪ An will gives all permutations that fix
at least one point, which is what we want.

Assume we have a k-tuple of these sets, Ai1 . . . Aik , where i1 , i2 , . . . ik are distinct integers
between 1 and n inclusive. How many elements are in their union?

Well, the union of k such sets would simply be the number of permutations that fix the k
elements i1 . . . ik . Since we are fixing k elements, and permuting
the rest, there are (n − k)!
elements in each intersection of k sets. Furthermore, there are nk different combinations of k
sets. So by PIE, we have that the number of permutations with atleast one fixed point is
     
n n n+1 n
(n − 1)! − (n − 2)! + · · · + (−1) 0!.
1 2 n

Since the total number of permutations is n!, our answer is

     
n n n n
n! − (n − 1)! + (n − 2)! − · · · + (−1) 0!.
1 2 n

Since nk (n − k)! = n!


k! , we can put this more concisely in summation notation as

n
X (−1)k
Dn = n! .
k!
k=0

Remark. Those of you who are familiar with calculus might notice that our answer is simply
n! times the the partial sum of the first n terms of the Maclaurin series expansion of 1e .
Indeed, it turns out that  
n! 1
Dn = + .
e 2
and that
Dn 1
lim = .
n→∞ n! e

In this problem, we used PIE on a variable number of sets; usually competition problems will
be more concrete, and we’ll have a known number of sets to work with. However, this problem
gave us valuable insight; we started by considering a crude approach that obviously overcounted,
then figured out how to use PIE to deal with that overcounting.

9
Everaise Academy (2020) Math Competitions I

§8.4 Homework Problems

Problem 8.1. [3] [2017 AMC 10B] There are 20 students participating in an after-school
program offering classes in yoga, bridge, and painting. Each student must take at least one of
these three classes, but may take two or all three. There are 10 students taking yoga, 13 taking
bridge, and 9 taking painting. There are 9 students taking at least two classes. How many
students are taking all three classes?

Problem 8.2. [4] [2019 AMC 10A] How many positive integer divisors of 2019 are perfect
squares or perfect cubes (or both)?

Problem 8.3. [6] [2018 HMMT] Lil Wayne, the rain god, determines the weather. If Lil Wayne
makes it rain on any given day, the probability that he makes it rain the next day is 75%. If Lil
Wayne doesn’t make it rain on one day, the probability that he makes it rain the next day is
25%. He decides not to make it rain today. Find the smallest positive integer n such that the
probability that Lil Wayne makes it rain n days from today is greater than 49.9%.

Problem 8.4. [7] [2003 AIME II] A bug starts at a vertex of an equilateral triangle. On each
move, it randomly selects one of the two vertices where it is not currently located, and crawls
along a side of the triangle to that vertex. Find the probability that the bug moves to its
starting vertex on its tenth move.

Problem 8.5. [8] [2019 AMC 10B] How many sequences of 0s and 1s of length 19 are there that
begin with a 0, end with a 0, contain no two consecutive 0s, and contain no three consecutive
1s?

Problem 8.6. [9] [2016 PUMaC] The George Washington Bridge is 2016 meters long. Sally
is standing on the George Washington Bridge, 1010 meters from its left end. Each step, she
1
either moves 1 meter to the left or 1 meter to the right, each with probability . What is the
2
expected number of steps she will take to reach an end of the bridge? Hint: Consider symmetry
about the point 1008 meters away from either end.

Problem 8.7. [10] [2019 AIME II]

While watching a show, Ayako, Billy, Carlos, Dahlia, Ehuang, and Frank sat in that order in
a row of six chairs. During the break, they went to the kitchen for a snack. When they came
back, they sat on those six chairs in such a way that if two of them sat next to each other before
the break, then they did not sit next to each other after the break. Find the number of possible
seating orders they could have chosen after the break. Hint: Consider using PIE where each of our
sets contains all configurations where a certain pair is seated together.

Problem 8.8. [8] [2001 AIME I] A mail carrier delivers mail to the nineteen houses on the east
side of Elm Street. The carrier notices that no two adjacent houses ever get mail on the same
day, but that there are never more than two houses in a row that get no mail on the same day.
How many different patterns of mail delivery are possible?
Provide a detailed solution, or explain what progress you made, even if you
didn’t get an answer.

10
9 Binomial Theorem, Pascal’s Triangle,
Combinatorial Identities

§9.1 Pascal’s Triangle

Pascal’s Triangle is a triangle generated by starting with the top row containing just 1. After
that, each number is the sum of the two adjacent numbers in the row above it, with the left
and right end of each row being 1. Above are the first 7 rows of Pascal’s triangle. From now
on, we’ll refer to the top row as row 0, the second row as row 1, and so on.

1
1 1
1 2 1
1
3 3 1
14 6 4 1
1 5 10 10 5 1
1 6 15 20 15 6 1

There are numerous patterns in Pascal’s triangle. For example, you might notice the sequence
1, 2, 3, . . . in a diagonal. After some more observation, you might notice that the sum of the
entries in each row is a power of 2. If you are very astute, you might even notice that if you sum
the numbers in a left-pointing diagonal of any length, this sum is equal to the number directly
below and to the right of where you ended the diagonal. We’ll be exploring these patterns in
this handout.

First, we’ll start with an observation about binomials. When we expand (x + 1)2 , we get
x2 + 2x + 1. When we expand (x + 1)3 we get x3 + 3x2 + 3x + 1. Do you notice anything?
The coefficients of the first expansion are 1, 2, and 1 in order. The coefficients of the second
one are 1, 3, 3, and 1. These are precisely the third and fourth rows of Pascal’s triangle! This
works for (x + 1)1 = x + 1 and (x + 1)0 = 1 as well. It turns out that it is true in the general
case.
So far we have a connection between expansions of (x + 1)n and the triangle. But we may
n


also notice another strange pattern– every

entry
of Pascal’s
triangle is k for some n and k.
For instance, 1 = 40 ,0 4 = 41
, 6 = 42 , 4 = 43 and 1 = 44 . We can conjecture that row n

has all numbers of the form ni for integers i between 0 and n inclusive, from left to right.
Exercise 9.1. Check this for the first 6 rows.
How do we prove this generally? We can use induction and the definition of Pascal’s triangle.
The first row is obvious to check.
Every number is the sum of the two above it. For the inductive step, assume our observation
about binomial coefficients holds for n − 1. We will show it holds for n.

1
Everaise Academy (2020) Math Competitions I

An entry in position k in row n is nk , and the numbers above it are n−1 n−1




k and k−1 . So we
need to prove that      
n n−1 n−1
= +
k k k−1

Proof. Let’s prove this combinatorially. The LHS is the number of ways to choose k objects
out of n objects. Directing our attention to one object, We consider two cases: if this object is
in our group, we need to choose k − 1 objects out of the n − 1. This can be done in n−1


k−1 ways.
If the object isn’t in our group, we can choose k objects out of the other n − 1, giving us the
other n−1


k ways, and we are done. 

Exercise 9.2. Prove this algebraically by expanding out the binomials.

Now we have

a nice description of Pascal’s triangle. Furthermore, the fact that both numbers
n
of the form k and binomial expansions are connected to Pascal’s triangle suggests a connection
between these two objects. This connection is known as the Binomial Theorem.

§9.2 Binomial Theorem

The Binomial Theorem is a simple theorem that can be applied in numerous contexts in algebra
and discrete math.

Theorem 9.3 (Binomial Theorem). For any positive integer n,

       
n n n n−1 n n−2 2 n n
(x + y)n = x + x y+ x y + ... + y .
0 1 2 n

Proof. To prove the Binomial Theorem, we simply expand out the left hand side. That is, we
write it as
(x + y)(x + y) . . . (x + y).
Now say we need to find the coefficient of xn . This is the number of ways we can choose n x’s
from the n brackets. But we can do this in only one way! Hence, the coefficient of xn must be
1.
Now let’s try and find the coefficient of xn−1 y. When we expand the above product, when
will we get a xn−1 y? Of course, we need to choose a y in one of the binomials and an x in the
rest.
 In how many ways can we do this? The number of ways of choosing one y in n brackets is
n
. Looks like we are on the right track. . . can we generalise this argument? Let’s try.
1
Let’s say we want to find the coefficient of xn−k y k . We can get such a term iff we choose y’s
in k binomials and xin the rest. How many ways can we choose k of these binomials out of all
n
n? Of course, in ways.
k
 
n
Thus, once we distribute on the LHS, we are simply adding together terms of each term
k
of the form xn−k y k for all integral k from 0 to n inclusive, giving us the result on the RHS, and
we are done. 

Exercise 9.4. Prove the Binomial Theorem using induction.

2
Everaise Academy (2020) Math Competitions I

Exercise 9.5. What do you get when you substitute x = y = 1 into the Binomial theorem? How
about x = 1, y = −1?

1 7
 
Example 9.6 (1963 AHSME). In the expansion of a − √ , what is the coefficient of
a
1
a− 2 ?

Solution. Since we are expanding out the product of a binomial raised to a high power, the
Binomial Theorem comes to mind. By the Binomial Theorem, each term of the expansion of
1 7 1 7−n
     
7 n 1
a− √ is of the form a −√ . We want the coefficient of a− 2 ; in other words,
a n a
we want  
7 n 1 3
n+ − + = − =⇒ n = 3 =⇒ n = 2.
2 2 2 2
1
Therefore, the coefficient of a− 2 is (−1)5 · 72 = −21.




This problem wasn’t actually a counting problem, but it illustrates how we apply the Binomial
Theorem.

Example 9.7 (2012 AMC 10B). Let (a1 , a2 , . . . a10 ) be a list of the first 10 positive integers
such that for each 2 ≤ i ≤ 10 either ai + 1 or ai − 1 or both appear somewhere before ai
in the list. How many such lists are there?

Solution. When we first look at this problem and test random numbers, we realize that the
number of possibilities depends on what we choose for the first number. This motivates us to
solve this problem using casework.
• The first number in the list is 1. Then, there is only one way to arrange the numbers
(think about why this is true!), namely (1, 2, 3, 4, 5, 6, 7, 8, 9, 10).
• The first number in the list is 2. Then, we must pick one of the remaining spots to be 1,
and the remaining spots will be filled up with 3, 4, 5. . . in that order. We can pick any of
9


the 9 spots (other than the first one) to be the 1, so there are 1 ways.
• The first number in the list is 3. Then, we must pick 2 of the remaining
spots to be 2 and
1 (in that order!). We can pick any 2 of the 9 spots left, so there are 92 ways.
..
.
• The first number in the list is 10. Then, we must pick 9
of the 9 remaining spots to be
the numbers 9, 8, 7, . . . , 2, 1 in that order. There are 99 ways to do this.
In total, we are trying to calculate
     
9 9 9
+ + ··· = 29 = 512.
0 1 9

If you do not understand how we got to the final step, look at Exercise 9.5. 

Example 9.8 (1993 AIME). Let S be a set with six elements. In how many different ways
can one select two not necessarily distinct subsets of S so that the union of the two subsets

3
Everaise Academy (2020) Math Competitions I

is S? The order of selection does not matter; for example, the pair of subsets {a, c} ,
{b, c, d, e, f } represents the same selection as the pair {b, c, d, e, f } , {a, c} .

Solution. It seems really hard to pick both sets at the same time, so instead we will be selecting
each set individually. Suppose that the first set contains x elements. Then, there are x6 ways

to select the first set (we can choose any x elements).

For the second set, we have to choose the 6 − x elements that aren’t in the first set, but we
can choose whether or not to have the x elements in the first set as well. This makes the total
number of ways to pick the second set 2x . Since 0 ≤ x ≤ 6,
6  
X 6 x
2 = (1 + 2)6 = 729.
x
x=0

Make sure you understand why this is true! We may be tempted to put down 729 as our
final answer, but remember that the order of the sets doesn’t matter. We counted every case
twice (other than the case when both sets are equal to {a, b, c, d, e, f }), so our final answer is
729−1
2 + 1 = 364 + 1 = 365. 

The Binomial Theorem is very useful when simplifying calculations. Some things that may
remind us of this theorem include the expansion of an expression to some power and the sum-
mation of a binomial with a fixed top value and a variable on the bottom. If you see any of
these things, the Binomial Theorem should come to mind!
Exercise 9.9. Find another (similar) solution to Example 9.8 that does not involve the Binomial
Theorem.
Exercise 9.10. In the expansion of (ax + b)2000 , where a and b are relatively prime positive
integers, the coefficients of x2 and x3 are equal. Find a + b.

§9.3 Hockey-Stick Identity

The Hockey-Stick Identity is a combinatorial identity that is often used in math competitions.
Even though it may not look particularly special, it can simplify many calculations.

Theorem 9.11 (Hockey-Stick Identity). For nonnegative integers n, r such that n ≥ r,

n    
X i n+1
= .
r r+1
i=r

Proof. We can prove the Hockey-Stick Identity by using a counting argument. We’ll count the
number of ways to put a identical balls in b distinguishable boxes in two different ways. By
stars and bars, we know that there are
 
a+b−1
b−1

ways to do this. However, we can also find this value using casework.
 
a+b−2
• If we put 0 balls into the first box, then there are ways to put the balls in
b−2
the remaining boxes.

4
Everaise Academy (2020) Math Competitions I
 
a+b−3
• If we put 1 ball into the first box, then there are ways to put the remaining
b−2
balls in the remaining boxes.
 
a+b−4
• If we put 2 balls into the first box, then there are ways to put the remaining
b−2
balls into the remaining boxes.
..
.  
a+b−2−a
• If we put a balls into the first box, then there are ways to put the
b−2
remaining balls into the remaining boxes.
We know that we should get the same answer with both methods, so
a    
X k+b−2 a+b−1
= .
b−2 b−1
k=0

If we let r = b − 2, n = a + b − 2, and i = k + b − 2, then the above equation becomes

n    
X i n+1
= .
r r+1
i=r

Remark. The Hockey-Stick Identity gets its name from the fact that the values form a
hockey stick on Pascal’s Triangle. The LHS forms the ”shaft” of the hockey stick and the
RHS forms the ”blade”. For example, with n = 5 and r = 2, the hockey stick would like
like this:

1
1 1
1 2 1
1
3 3 1
14 6 4 1
1 5 10 10 5 1
1 6 15 20 15 6 1

Note that 1 + 3 + 6 + 10 = 20.

Example 9.12 (2006 SMT). Simplify:

2006  
X k
.
10
k=10

Your answer should contain no summations but may still contain binomial coefficients
and/or combinations.

Solution. This problem is a very direct application of the Hockey-Stick identity with n = 2006
and r = 10. The answer is  
2006 + 1
,
10 + 1
which is 2007


11 . 

5
Everaise Academy (2020) Math Competitions I

Example 9.13 (2013 AIME I). Melinda has three empty boxes and 12 textbooks, three of
which are mathematics textbooks. One box will hold any three of her textbooks, one will
hold any four of her textbooks, and one will hold any five of her textbooks. If Melinda packs
her textbooks into these boxes in random order, the probability that all three mathematics
textbooks end up in the same box can be written as m n , where m and n are relatively prime
positive integers. Find m + n.

Solution. You may remember this problem from the first day of combinatorics. This time,
we’ll solve it faster using the Hockey-Stick identity.
There are three possibilities for where the math books go. They could go in any of the three
boxes, and the calculations will be different in each case, so we’ll use casework.
• The math books are all in the smallest box (the one that holds 3 books). There are 33

ways to do this.
• The math books are all in the middle box (the one that holds 4 books). There are 43

ways to do this.
• The math books are all in the largest box (the one that holds 5 books). There are 53

ways to do this.
Of course, we can calculate each value individually, but we can also simplify our calculations
by using Hockey-Stick. This comes to mind because the top numbers of the binomials
are
3
consecutive

integers
and the bottom value is the same as the smallest of the top values. 3 +
4 5 6 6



3 + 3 = 4 = 2 . Our answer is
 
6
number of ways to put all 3 math books in one box 2 15 3
= = = .
total number of ways to put 3 math books in the boxes 12 220 44
2
Our desired value is 3 + 44 = 47. 

We will finish this section with a harder example.

Example 9.14 (2015 AIME I). Consider all 1000-element subsets of the set {1, 2, 3, . . . ,
2015}. From each such subset choose the least element. The arithmetic mean of all of
these least elements is pq , where p and q are relatively prime positive integers. Find p + q.

Solution. This problem may look very intimidating, so let’s think about what it’s asking for.
We are asked to find the average of all the least elements, or
sum of the least elements of all 1000-element subsets
.
total number of distinct 1000-element subsets
The denominator is fairly easy to find, it’s just 2015


1000 , since we are choosing 1000 of the 2015
values to form a set. The numerator is a lot harder to find, because the smallest number varies
with each subset. This motivates us to use casework.
• The smallest element of the subset is 1. There are 2014


999 such subsets. (Why?)
• The smallest element of the subset is 2. There are 2013


999 such subsets (since none of the
other elements can be 1 or 2).
..
.

6
Everaise Academy (2020) Math Competitions I

• The smallest element of the subset is 1016. There are 999

999 such subsets (since none of
the elements can be 1, 2, . . . , 1015, or 1016).
Note that we cannot have a subset with 1017 as its smallest element, since the largest element
we can have is 2015 and we must have 1000 distinct elements. The sum of the least elements of
all 1000-element subsets is
         
2014 2013 2012 2011 999
1· +2· +3· +4· + . . . + 1016 · .
999 999 999 999 999
This value is very hard to compute by itself, but notice that we can split it up as follows:
         
2014 2013 2012 2011 999
= + + + + ...+
999 999 999 999 999
       
2013 2012 2011 999
+ + + + ...+
999 999 999 999
     
2012 2011 999
+ + + ...+
999 999 999
   
2011 999
+ + ...+
999 999
..
.
 
999
+
999
It is very hard to sum everything together, but it becomes much easier when we try to add the
values one row at a time. By the Hockey-Stick Identity,
           
2014 2013 2012 2011 999 2015
+ + + + ... + =
999 999 999 999 999 1000
         
2013 2012 2011 999 2014
+ + + ... + =
999 999 999 999 1000
       
2012 2011 999 2013
+ + ... + =
999 999 999 1000
     
2011 999 2012
+ ... + =
999 999 999
..
.
   
999 1000
=
999 1000
We can simplify the sum to
         
2015 2014 2013 2012 1000
+ + + + ··· .
1000 1000 1000 1000 1000
While this looks better than the first expression we got, it is still very hard to calculate. However,
we can

use the Hockey-Stick Identity one last time to compute the sum, which turns out to be
2016
1001 .
This may look like the answer, but remember that we are not done yet! We are asked for the
expected value, which is
 
2016
1001 2016! 1000!1015! 2016 288
 = · = = .
2015 1001!1015! 2015! 1001 143
1000

7
Everaise Academy (2020) Math Competitions I

Once again, we are dividing by 2015

1000 because that is the total number of distinct 1000-element
subsets. The question asks for the sum of the numerator and the denominator, which is 288 +
143 = 431. 

We can take away a few things from this problem. First, if a problem seems too hard at first,
try simplifying it into something that’s easier or something that you can do. Second, sometimes
it’s worth changing perspectives when trying to simplify a value. Third, remember what you’re
looking for! Don’t do everything correctly only to get it wrong because you didn’t read the
question.
In any problem involving a large sum of binomial coefficients, you should consider using
Hockey-Stick, and see if you can manipulate the expression into a form
where you can apply
Hockey-Stick. When you are summing multiple binomial coefficients nk with constant k and
consecutive n, remember this very important identity!

§9.4 Vandermonde’s Identity

Similar to the Hockey-Stick Identity, Vandermonde’s Identity can often be used to simplify
answers and to get a cleaner solution.

Theorem 9.15 (Vandermonde’s Identity). For nonnegative integers m, n, and r, where

r ≤ m, n,
r     
X m n m+n
= .
k r−k r
k=0

Proof. We will prove this theorem using a combinatorial argument; that is, we will show the
left and right sides of the equation count the same thing.

First, let’s try to figure out what the RHS could count. Obviously, we could say we are
choosing a set of r people out of m + n. What will m and n represent?

Let’s suppose there are m boys and n girls, and we are trying to form a team of r from this
group. The RHSclearly
 describes
 this. How about the LHS? Let’s try looking at a single term
m n
in the sum, say, . What does this mean?
2 r−2
The first term is equivalent to the number of ways of choosing 2 out of m boys, and the
second counts the number of ways to choose r − 2 out of the n girls. This forms a team of r,
but with a specific number of boys and girls. But in the RHS, the number of boys and girls
could be anything.
That’s why we’re summing these single terms over all possible values of k in the LHS! The
first term is the number of ways of choosing a team of r girls, the second is the number of ways
of choosing r − 1 girls and 1 boy, and so on, with the last term being the number of ways of
choosing r boys in the team. In total, this covers each and every way to choose a team of r
from a pool of n girls and m boys, which is exactly what the RHS counts as well, and we are
done. 

Make sure you fully internalize and understand the above proof; it is a classic example of a
combinatorial argument.

8
Everaise Academy (2020) Math Competitions I

Example 9.16. Prove that

n  2  
X n 2n
= .
k n
k=0

Solution. This identity looks extremely similar to Vandermonde’s Identity because of the sum-
mation from 0 to n and the product of two binomials, so we’ll try to apply it. When we substitute
n = m = r into Vandermonde’s, we get
n     
X n n 2n
= .
k n−k n
k=0
   
n n
We know that = , so we are done. 
k n−k

Example 9.17 (2020 AIME I). A club consisting of 11 men and 12 women needs to choose
a committee from among its members so that the number of women on the committee is
one more than the number of men on the committee. The committee could have as few as
1 member or as many as 23 members. Let N be the number of such committees that can
be formed. Find the sum of the prime numbers that divide N.

Solution. Suppose that we pick k men and k + 1 women to be on the committee. k ranges
from 0 to 11, so the total number of ways to do choose a committee is
         
11 12 11 12 11 12
+ + ··· + ,
0 1 1 2 11 12

which is
11    X11    X11   
X 11 12 11 12 11 12
= = .
k k+1 k 12 − (k + 1) k 11 − k
k=0 k=0 k=0

Once again, the product of the binomials reminds us of Vandermonde’s. When we compare
Vandermonde’s Identity to this expression, we see that this is Vandermonde’s when m = 11, n =
12, and r = 11, so the number of ways is
   
11 + 12 23
= = 2 · 7 · 13 · 17 · 19 · 23.
11 11

Thus, the answer is 2 + 7 + 13 + 17 + 19 + 23 = 81. 

Vandermonde’s Identity is used when we are trying to find the sum of the product of two
binomials. In the majority of cases involving the product of two binomials, Vandermonde’s
Identity can make the calculation much easier.

§9.5 Homework Problems

Problem 9.1. [3] Given (x + y)10 , find the coefficient of the x4 y 6 term.
√ √
Problem 9.2 (1975 AHSME). [5] What is the smallest integer larger than ( 3 + 2)6 ?

9
Everaise Academy (2020) Math Competitions I

Problem 9.3
(2001 AIME I). [6] Find the sum of the roots, real and non-real, of the equation
2001
x2001 + 21 − x = 0, given that there are no multiple roots.

Problem 9.4 (1992 AIME). [7] In Pascal’s Triangle, each entry is the sum of the two entries
above it. In which row of Pascal’s Triangle do three consecutive entries occur that are in the
ratio 3 : 4 : 5?

Problem 9.5 (2016 AMC 10A). [7] For some particular value of N , when (a + b + c + d + 1)N
is expanded and like terms are combined, the resulting expression contains exactly 1001 terms
that include all four variables a, b, c, and d, each to some positive power. What is N ?

Problem 9.6 (2008 iTest). [6] Find the sum of the 2007 roots of (x − 1)2007 + 2(x − 2)2006 +
3(x − 3)2005 + · · · + 2006(x − 2006)2 + 2007(x − 2007).

Problem 9.7 (2018 AIME I). [9] The wheel shown below consists of two circles and five spokes,
with a label at each point where a spoke meets a circle. A bug walks along the wheel, starting
at point A. At every step of the process, the bug walks from one labeled point to an adjacent
labeled point. Along the inner circle the bug only walks in a counterclockwise direction, and
along the outer circle the bug only walks in a clockwise direction. For example, the bug could
travel along the path AJABCHCHIJA, which has 10 steps. Let n be the number of paths
with 15 steps that begin and end at point A. Find the remainder when n is divided by 1000.

Hint: Consider a sequence of moves as a string, where an I represents a clockwise movement (in the inner
circle), an O represents a counterclockwise movement (in the outer circle), and an S represents a switch between
the inner and outer circle. What do we know about the number of Is and Os?

Problem 9.8 (2006 AIME II). [9] Seven teams play a soccer tournament in which each team
plays every other team exactly once. No ties occur, each team has a 50% chance of winning
each game it plays, and the outcomes of the games are independent. In each game, the winner
is awarded a point and the loser gets 0 points. The total points are accumulated to decide the
ranks of the teams. In the first game of the tournament, team A beats team B. The probability
that team A finishes with more points than team B is m/n, where m and n are relatively prime
positive integers. Find m + n.
Provide a detailed solution, or explain what progress you made, even if you
didn’t get an answer.

10
10 Probability
Today, we’ll review probability. It won’t be very different from previous handouts this week
because many probability problems are just counting problems in disguise.
We can define probability of a certain event as the number of outcomes where the event
occurs (assuming all outcomes have equal probability) over the total possible number of out-
comes. We can also think of it as the proportion of times an event will occur if we repeat a
certain process many times.
For example, the probability a fair coin flips heads is 21 . There are two outcomes, heads and
tails, with equal probability.

§10.1 Independent/Dependent Events

A pitfall in probability problems is failing to recognize when events are independent and when
they are dependent. To avoid making this mistake, let’s define these two terms and work
through some related problems.

Definition 10.1. Two events are independent if the outcome of one event does not affect the
outcome of the other.

Let X and Y be two events. We let P (X) denote the probability of event X occurring. We
denote the probability that X and Y both occur as P (X ∩ Y ).

Proposition 10.2 (Independent Events). Assume we have two independent events A and
B. Then P (A ∩ B) = P (A) · P (B).

Exercise 10.3. Prove this using the definition of an independent event.

.

Example 10.4. I roll two fair dice numbered 1, 2, 3 . . . , 6. What is the probability that I
roll two 1s?

Solution. In this problem, the outcome of each roll is independent because the result of our
first roll does not affect the result of the second roll. The

1probability that a single die shows 1
1 1 1


is 6 . Thus, the probability that both dice show 1 is 6 6 = 36 . 

This problem was a pretty simple, yet constructive example of independent events. If two
events are independent, we can calculate the probability that each event happens individually
and multiply them to find the probability that they both happen.

Example 10.5. I roll two fair dice numbered 1 through 6. What is the probability that I
roll at least one 1?

Solution. We can use the ideas of complementary counting here. Let’s calculate the probability
that we roll no 1s instead. The probability that one dice does not roll as 1 is 56 . Then the

1
Everaise Academy (2020) Math Competitions I

25
probability that both dice don’t show 1 is 36 . Thus, the probability that at least one die rolls
1 is 1 − 25 11
36 = 36 . 

Example 10.6. We roll 3 dice numbered 1 through 6. What is the probability that exactly
one die rolls 1?

Solution. The probability that one die rolls 1 is 61 , and the probability that the other two
do not roll 1 is 56 . Since there are 3 choices for the die that rolls 1, our desired probability is
3( 61 )( 56 )2 = 216
75
. 

Independent events are relatively straightforward. The general idea is: find the probabilities
of the events who want(or don’t want in the case of complementary probability), and multiply
them together. As we will see, dependent probability is a lot more complicated because, as the
name suggests, events depend on the outcome of other events which adds a whole new wrinkle
in the problem.

Definition 10.7. Two events are dependent if the outcome of one event affects the outcome
of the other event (i.e. change the probability it occurs).

Example 10.8. I shuffle a standard deck of 52 cards and deal them. What is the probability
that the first two cards I deal are spades?

Solution. These events are dependent because after the first spade is dealt, it is removed from
the deck, lowering the probability of drawing a second spade. The probability that the first
card is a spade is 13 1
52 = 4 . Since there are now 12 spades remaining out of the 51 undealt cards,
the probability that the second
card
is a spade is 12 4
51 = 17 . Thus, the probability that the first
two cards dealt are spades is 41 17 4 1
= 17 . 

It is important to note the difference between “with replacement” and “without replacement”.
When we talk about doing something with replacement, it means we are resetting the event
each time. When we talk about doing something without replacement, it’s the opposite: each
event’s change on the outcome of the next event is permanent.

Example 10.9. I have a hat with 10 slips of paper numbered 1 through 10. What is the
probability that I draw a 1?
1. With replacement
2. Without replacement

Solution.
Let’s first consider the problem with replacement. Since we are choosing numbers with replace-
1
ment, we can think of this as rolling a 10 sided die. There is a 10 chance that we draw a one
9 3 729
on any given draw. It follows that the probability we don’t draw a 1 is ( 10 ) = 1000 . Thus, the
271
probability that we draw a 1 is 1000

For the second part, we can use the same idea, but we must remember that we are choosing
9
numbers without replacement. The probability that our first draw is not 1 is 10 . Since we do
not replace the number we drew, the probability that our second draw is not 1 is 89 . Similarly,

2
Everaise Academy (2020) Math Competitions I

7
the probability that our third draw is not 1 is 8. Thus, the probability that we draw a 1 is
9
1 − ( 10 )( 89 )( 78 ) = 1 − 10
7 3
= 10 . 

Now let’s move on to some more complicated problems involving independent and dependent
events.

Example 10.10 (AMC 12). A coin is biased in such a way that on each toss the probability
of heads is 23 and the probability of tails is 31 . The outcomes of the tosses are independent.
A player has the choice of playing Game A or Game B. In Game A she tosses the coin
three times and wins if all three outcomes are the same. In Game B she tosses the coin
four times and wins if both the outcomes of the first and second tosses are the same and
the outcomes of the third and fourth tosses are the same. How do the chances of winning
Game A compare to the chances of winning Game B?

Solution. This problem is wordy, but we are essentially solving for and comparing two different
probabilities. Since we are flipping coins, each outcome is separate from another and thus
independent(the problem also tells us this).

3
3
The probability that we win Game A is just 32 + 13 = 27 9
= 31 .
Calculating the probability that we win Game B is a little trickier. There are 4 possible
winning

outcomes:
HHTT, HHHH,
TTHH,

TTTT. Thus, the probability that we win Game B
2 2 1 2 2 2 2 2 1 2 2 2 1 2 1 2 25





is 3 3 + 3 3 + 3 3 + 3 3 = 81 .
2
Thus, the probability that we win Game A is 81 higher than the probability that we win
Game B. 

Finding the probability that we won Game A was a pretty straightforward application of what
we know about independent events, and it just took some extra casework to find the probability
for Game B.

Example 10.11 (AIME). An urn contains 4 green balls and 6 blue balls. A second urn
contains 16 green balls and N blue balls. A single ball is drawn at random from each urn.
The probability that both balls are of the same color is 0.58. Find N .

Solution. We have two cases that are successful: either both balls drawn are green or both are
blue.

Case 1: Both balls drawn are green

Notice that we are dealing with independent events here because one green ball being drawn
from the first urn does not affect the probability of a green ball being drawn from the second
4
as the urns are separate. The probability that the ball from the first urn is green is 4+6 = 52 .
16
The probability that the ball from the second urn is green is 16+N . Thus, the probability that
   
2 16 32


both balls drawn are green is 5 16+N = 80+5N .

Case 2: Both balls drawn are blue

6
Using similar logic, the probability that the ball from the first urn is blue is 10 = 53 . The
N
probability that the ball from the second urn is blue is 16+N . Thus, the probability that both

3
Everaise Academy (2020) Math Competitions I

3

 N
 
3N

balls are blue is 5 16+N = 80+5N .

Summing up the two cases, we have that the total probability that both balls chosen are the
same color is 32+3N 58 29
80+5N = 100 = 50 . Cross multiplying, we get 1600 + 150N = 2320 + 145N, so
N = 144. 

Now let’s take a look at a problem involving dependent probability.

Example 10.12 (Mathcounts). A 3 × 3 × 3 wooden cube is painted on all six faces, then
cut into 27 unit cubes. One unit cube is randomly selected and rolled. After it is rolled,
5 of the 6 faces of the cube are visible. What is the probability that exactly one of the 5
visible faces is unpainted?

Solution. Observe that when we cut the cube, some unit cubes will have 1 painted face, some
will have 2, some will have 3, and the one cube in the middle will have 0 painted faces. If we
visualize the cube, we find that the 8 corners all have 3 painted faces, the 12 cubes in between
the corners have 2 painted faces, the 6 cubes in the center of each face have 1 painted face, and
the cube in the middle has no painted faces.
Notice that it is impossible for exactly one visible face to be painted with the 0 painted face
cube because it literally has 0 painted faces, and it is also impossible for the cube with 3 painted
faces because at least two visible faces are always painted.

The solutions lie in the other two possible cubes. If we pick a cube with exactly 1 painted
6
face, then the hidden face can not be the originally painted face. Since there is a 27 probability
6 5 5
that we even choose this type of cube in the first place, there is a 27 · 6 = 27 probability of
success in this scenario. If we pick a cube with exactly 2 painted faces, then the hidden face
must be an originally painted face. Since there were 12 of these cubes to start with, we have a
12 2 4
27 · 6 = 27 chance of success in this case.
5 4
Thus, the probability that exactly one of the 5 visible faces is painted is 27 + 27 = 13 . 

Exercise 10.13. I roll three dice. What is the probability that the numbers rolled on two of the
dice sum to the third?
Exercise 10.14. Six people each flip a fair coin. Everyone who flipped tails then flips their coin
again. Given that the probability that all the coins are now heads can be expressed as simplified
fraction m
n , compute m + n.

Exercise 10.15. When rolling a certain unfair six-sided die with faces numbered 1, 2, 3, 4, 5,
and 6, the probability of obtaining face F is greater than 1/6, the probability of obtaining the
face opposite is less than 1/6, the probability of obtaining any one of the other four faces is
1/6, and the sum of the numbers on opposite faces is 7. When two such dice are rolled, the
probability of obtaining a sum of 7 is 47/288. Find the probability of rolling face F .

§10.2 Conditional Probability

Definition 10.16. The conditional probability of an event A given another event B is the
probability that A occurs given that B occurs. We denote this probability as P (A|B).

We also denote the probability that both event A and event B occur as P (A ∩ B).

4
Everaise Academy (2020) Math Competitions I

P (A∩B)
Proposition 10.17 (Conditional Probability). P (A|B) = P (B) .

If this doesn’t make sense at first, think of it in terms of words instead of notation. The
probability that event A occurs given that event B occurs is the probability that both events
A and B occur, divided by the probability that event B occurs.
Compared with traditional probability, the main difference with conditional probability is
that there is usually some restriction given that limits the possible outcomes.

Example 10.18. I roll two fair dice, each numbered 1, 2, . . . , 12. Given that the first die
rolls 11, what is the probability that the other die also rolls a prime number?

Solution. The first die roll does not influence the second die roll. As there are 5 primes (2, 3,
5
5, 7, 11) below 12, the probability that the second die rolls a prime is 12 . 

Note that P (A|B) for independent events A, B is simply P (A) because forcing B to be true
does not do anything for A. This logic can’t apply for dependent events because the outcome
of B affects the outcome of A.
P (A|B) for dependent events A, B can be thought of as the number of ways for A and B to
happen divided by the total number of ways for B.

Example 10.19. I have a fair six sided die. Given that the result is odd, what is the
probability that
a) I rolled a 1
b) I rolled a prime number?

Solution. Let’s consider both situations.

a) We are given that our result was odd, so we must have rolled a 1,3, or 5. All of these are
equally likely, so the probability we rolled a 1 is 31 .
b) Out of the three possible outcomes of 1,3, and 5, both 3 and 5 are prime. Thus, the
probability that we roll a prime is 23 . 

Let’s take a look at another problem continuing with the team of rolling dice.

Example 10.20. I roll two fair six sided dice. Given that at least one of the dice rolls 2,
what is the probability that the sum of the two dice is 5?

Solution. Let’s begin by first counting the number of possible outcomes with the restriction.
There are 6 outcomes for the second die when we roll the first die as 2 and also 6 outcomes
for the first die when we roll the second die as 2. There is also one outcome where we roll two
2s. Thus, by PIE, the number of possible outcomes is 6 + 6 − 1 or 11.
Now, we need to count our successful outcomes. The only possible successes are rolling a 2, 3
2
or a 3, 2. Thus, the desired probability is 11 . 

5
Everaise Academy (2020) Math Competitions I

Exercise 10.21 (AHMSE). A fair standard six-sided dice is tossed three times. Given that the
sum of the first two tosses equal the third, what is the probability that at least one 2 is tossed?

Example 10.22 (AIME). There is a 40% chance of rain on Saturday and a 30% chance of
rain on Sunday. However, it is twice as likely to rain on Sunday if it rains on Saturday
than if it does not rain on Saturday. The probability that it rains at least one day this
weekend is ab , where a and b are relatively prime positive integers. Find a + b.

Solution. Let x be the probability that it rains on Sunday given that there is no rain on
Saturday. Then, we have that
2 3 3 3
(2x) + x = =⇒ x = .
5 5 10 14
So the probability that it doesn’t rain on Sunday given that it doesn’t rain on Saturday is 11 14 .
Since we already know the probability that it doesn’t rain on Saturday, the probability that it
doesn’t rain both days is 11 3 33 37
14 · 5 = 70 . Thus, the probability that it rains at least one day is 70 ,
and our desired sum is 107. 

Here’s another example:

Example 10.23. I’m planning a party. Kevin says he has a 80% chance of showing up.
Russell tells me he has a 30% chance of going if Kevin is going. Otherwise, he has a 75%
chance of showing up. If I see Russell, what is the probability Kevin is also there?

8 3 3
Solution. The probability that both Kevin and Russell are at my party is 10 · 10 = 25 . The
2
probability that just Russell is there is 10 · 34 = 20
3
.
3
Thus, the probability that Kevin is at the part given that Russell is also there is 3
25
+ 3 =
25 20
12
27 . 

Exercise 10.24. 2 percent of the population of America have contracted a certain type of disease.
Fortunately, a screening test for the disease has been developed that claims to be 95 percent
accurate”; that is, the test will return the correct results in 95 percent of the cases and return
incorrect results in the other 5 percent of the cases. What is the probability that someone tested
positive for the virus actually has the virus(calc allowed)?
Let’s move on to a more complicated problem.

Example 10.25 (HMMT). Urn A contains 4 white balls and 2 red balls. Urn B contains
3 red balls and 3 blackballs. An urn is randomly selected, and then a ball inside of that
urn is removed. We then repeat the process of selecting an urn and drawing out a ball,
without returning the first ball. What is the probability that the first ball drawn was red,
given that the second ball drawn was black?

Solution. We first count the probability where a black ball is the second ball drawn and then
find the intersection for when a red ball then black ball is drawn. Dividing our first result by
our second, we will get our desired probability.
To find the probability where a black ball is chosen second, we have three cases: a ball was
chosen from urn A first(color doesn’t matter), a red ball was chosen from urn B, or a black ball
was chosen from urn B.

6
Everaise Academy (2020) Math Competitions I

Case 1: Ball from Urn A first

1 1 1
This probability is 2 ·1· 2 · 2 = 81 .

Case 2: Red ball from Urn B first

1 1 1 3 3
This probability is 2 · 2 · 2 · 5 = 40 .

Case 3: Black ball from Urn B first

1 1 1 2 2
This probability is 2 · 2 · 2 · 5 = 40

1 3 2
Thus, the probability a black ball is chosen second is 8 + 40 + 40 = 41 .
Now, we just have to find the probability that a red ball was drawn first with a black ball
being the second draw. There are two scenarios we need to consider. The first is when a red
ball is drawn out of Urn A first, and the second is when a red ball is drawn out of Urn B first.
The probability is 12 · 13 · 12 · 21 + 12 · 12 · 21 · 35 = 24
1 3
+ 40 7
= 60 .
Thus, the probability the first ball drawn was red given that the second ball drawn was black
7
7
is 60
1 = 15 . 
4

This seemed to be a pretty confusing problem at first glance, but it essentially boiled down
to applying the formula we introduced at the beginning of the section. However, don’t think of
it as a formula, it should come naturally. The probability that event B happens can be thought
of as the possible outcomes and the probability that both occur can be thought of as successful
outcomes. When you think about it, our formula for conditional probability is no different than
traditional probability!

Remark. The Monty Hall Problem is a fun application of conditional probability. The
problem goes:
Suppose you’re on a game show, and you’re given the choice of three doors: Behind one
door is a car; behind the others, goats. You pick a door, say No. 1, and the host, who
knows what’s behind the doors, opens another door, say No. 3, which has a goat. He
then says to you, Do you want to pick door No. 2? Is it to your advantage to switch your
choice?
Think about it!

Exercise 10.26 (AIME). Arnold is studying the prevalence of three health risk factors, denoted
by A, B, and C, within a population of men. For each of the three factors, the probability that
a randomly selected man in the population has only this risk factor (and none of the others) is
0.1. For any two of the three factors, the probability that a randomly selected man has exactly
these two risk factors (but not the third) is 0.14. The probability that a randomly selected
man has all three risk factors, given that he has A and B is 13 . The probability that a man has
none of the three risk factors given that he does not have risk factor A is pq , where p and q are
relatively prime positive integers. Find p + q.
Exercise 10.27 (AMC 10). Two counterfeit coins of equal weight are mixed with 8 identical
genuine coins. The weight of each of the counterfeit coins is different from the weight of each
of the genuine coins. A pair of coins is selected at random without replacement from the 10

7
Everaise Academy (2020) Math Competitions I

coins. A second pair is selected at random without replacement from the remaining 8 coins.
The combined weight of the first pair is equal to the combined weight of the second pair. What
is the probability that all 4 selected coins are genuine?

§10.3 Geometric Probability

In probability, we occasionally are given problems where there’s no way we can count the number
of possibilities; there are infinitely many.. . .
Geometric probability is our solution. Instead of counting individual items, we represent our
counts with geometric regions. Our desired probability is then given as the size of the region
that represents a “success” divided by the size of the total region.
Let’s take a look at an example.

Example 10.28. Segment AB has length 3. Let point C lie on AB with AC = 2. What is
the probability a randomly chosen point is closer to C than A?

Solution. To begin, we have to ask ourselves what the successful region is. Let us denote D
as the midpoint of AC. Notice that all points to the right of C are closer to C than A. DB = 2
and AB = 3.
Thus, the probability the point is closer to C than A is 32 . 

Exercise 10.29. Let x be a real number in the range [-2, 2]. What is the probability x2 < 3?
We can use geometric probability in two dimensions as well.

Example 10.30. Suppose you are playing a game in which you throw a dart at a circle
of diameter 1 inscribed in a 1 × 1 square board. You win only if the dart hits the circle.
What’s the probability you win?

Solution. In this problem, our successful region is the circle and the total region is the square.
The area of the circle is 14 π and the area of the square is 1. Thus, the probability the dart
lands in the circle and we win is 14 π 

Now let’s move on to a classic problem involving geometric probability.

Example 10.31. Anna and Bob are hoping to meet for lunch. They will each arrive at
the restaurant sometime between 2 p.m. and 3 p.m., stay for 15 minutes, and then leave.
What is the probability that they will meet each other at the restaurant?

Solution. There are infinitely many times that Bob and Anna could show up, so it makes sense
to think of geometric probability.
Let’s set up a graph to represent the situation. Take the xy plane. We can let the x-axis
represent the time Anna arrives, and we can let the y-axis represent the time Bob arrives.
Our total area of possibility can be as a square with side lengths of 60(for each minute in the
hour).
Let’s consider the successful region in terms of when Anna arrives. If Anna arrives at 2:00,
Bob has until 2:15 to show up. If Anna arrives at 2:05, Bob has until 2:20 to show up and

8
Everaise Academy (2020) Math Competitions I

he also could be waiting from possibly 2:00. If Anna arrives at 2:15, Bob could have arrived
anytime between 2:00-2:30.
If we graph the successful regions, we notice that the successful region appears to be a hexagon
and the unsuccessful region consists of two right isosceles triangles. The side lengths of both

triangles are 45, so our unsuccessful region has area 2 12 · 45 · 45 = 2025. Thus, our successful


1575 7
region has area 1575 and our desired probability is 3600 = 16 . 

This is a pretty useful strategy for dealing with geometric probability problems. We may not
know the shape at first, but by testing values, we can get a general idea of it and reason out
the actual shape.

Example 10.32 (AMC 10). Chloe chooses a real number uniformly at random from the
interval [0, 2017]. Independently, Laurent chooses a real number uniformly at random from
the interval [0, 4034]. What is the probability that Laurent’s number is greater than Chloe’s
number?

Solution. Similar to the previous problem, we can model this with the xy plane. Let the x-axis
represent Laurent’s number and let the y-axis represent Chloe’s number. We have a rectangle
with length 4034 and height 2017. Since Laurent’s number is the x coordinate of whatever point
is in the rectangle, our successful region is the intersection between the region x > y and the
rectangular region.
2017·2017
After graphing it, we notice that the area of the complement of our region is 2 , and
the total area is 2017 · 4034. Thus, our desired probability is 1 − 41 = 34 . 

In conclusion, we want to use geometric probability when there’s no way to individually count
our successes. However, the basic idea of traditional probability remains: we find the area
representing the possibilities we want, and divide by the total area.
Exercise 10.33. Nate and Francis go to the county carnival at a random time between noon and
6 PM. Nate stays for 2 hours and Francis stays for 3. What is the probability that they meet?
Exercise 10.34 (AMC). What is the probability that if three points are chosen at random on
the circumference of a circle, then the triangle formed by connecting the three points does not
have a side length greater than the radius?

§10.4 Homework Problems

Problem 10.1. [3] Bill is choosing cards out of a standard deck of 52 cards. Given that the
first card is black, the probability that the second card is red can be expressed as m
n where m

9
Everaise Academy (2020) Math Competitions I

and n are relatively prime positive integers. Find m + n

Problem 10.2. [4] I roll one fair die and flip 6 coins. The probability that the number of times
I get heads is equal to the number I roll on the dice can be expressed as mn where m and n are
relatively prime positive integers. Find m + n.

Problem 10.3. [5] Andrew has an urn with 17 marbles, 3 are green, 4 are blue, 5 are red, and
5 are brown. Bill keeps taking marbles out of the urn until there are none left. The probability
the 6th marble he chooses is blue can be expressed as m n where m and n are relatively prime
positive integers. Find m + n

Problem 10.4. [5] Two points are chosen on a unit circle. The probability the distance between
those two points is greater than 1 can be expressed as mn where m and n are relatively prime
positive integers. Find m + n.

Problem 10.5. [5][2020 AMC 10/12B] An urn contains one red ball and one blue ball. A box
of extra red and blue balls lie nearby. George performs the following operation four times: he
draws a ball from the urn at random and then takes a ball of the same color from the box and
returns those two matching balls to the urn. After the four iterations the urn contains six balls.
The probability that the urn contains three balls of each color can be expressed as mn where m
and n are relatively prime positive integers. Find m + n.

Problem 10.6. [6] A bag has 6 red marbles and 4 green marbles, another bag has 5 blue
marbles and 4 green marbles. Jaedon first chooses one of the two bags at random, then chooses
2 marbles from the bag one at a time. If the first of the two marbles he chose was green, the
probability the second marble he chooses is also green can be expressed as mn where m and n
are relatively prime positive integers. Find m + n.

Problem 10.7. [6] Connor flips an unfair coin. The probability that Connor flips exactly 2
heads in 3 flips is equal to the probability that Connor flips √exactly 3 tails in 4 flips. The
p− q
probability the coin flips heads in 1 flip can be expressed as r p and r are relatively prime
and q is not divisible by the square of any prime. Find p + q + r.

Problem 10.8. [7] [2017 AIME II] A triangle has vertices A(0, 0), B(12, 0), and C(8, 10). The
probability that a randomly chosen point inside the triangle is closer to vertex B than to either
vertex A or vertex C can be written as pq , where p and q are relatively prime positive integers.
Find p + q.

Problem 10.9. [7] [2017 AIME I] A circle is circumscribed around an isosceles triangle whose
two congruent angles have degree measure x. Two points are chosen independently and uni-
formly at random on the circle, and a chord is drawn between them. The probability that the
chord intersects the triangle is 14
25 . Find the difference between the largest and smallest possible
values of x.

Problem 10.10. [7] [2017 AIME I] Two real numbers a and b are chosen independently and
uniformly at random from the interval (0, 75). Let O and P be two points on the plane with
OP = 200. Let Q and R be on the same side of line OP such that the degree measures of
∠P OQ and ∠P OR are a and b respectively, and ∠OQP and ∠ORP are both right angles. The
probability that QR ≤ 100 is equal to m
n , where m and n are relatively prime positive integers.
Find m + n.

10
11 Triangles I

§11.1 Introduction to Triangles

A triangle is a three-sided polygon. It is composed of three sides that together form three
angles (hence why it is referred to as ”tri-angle”).

c b

B a C

Throughout this week, in a triangle 4ABC with vertices A, B, C, we will use the notation
a, b, and c for the lengths of segments BC, CA, and AB, respectively. There are several types
of special triangles. Note that a triangle can fit into multiple types here.
1. Equilateral: This one has three sides of equal length, and consequently, three angles of
degree measure 60◦
2. Isosceles: This one has two sides of equal length. There is some debate about whether it
should have exactly two equal sides or just at least two, but that will always be clarified
on competitions. It has two equal base angles and one vertex angle.
3. Scalene: None of the side lengths or angles are equal to each other.
4. Acute: All angles are less than 90◦ . All of these angles are namely acute.
5. Obtuse: One angle is greater than 90◦ . This angle is namely the obtuse angle.
6. Right: One angle is exactly 90◦ . The side opposite this right angle is the hypotenuse,
which is also the longest side. The other two sides are called legs.
It is important to note that the first three types deal with side lengths whereas the last three
deal with angles, thus creating the potential for overlap.
There are some restrictions on the side lengths of a triangle. If I asked you to create a triangle
with lengths 2, 3, and 5, you wouldn’t be able to do it. At best, you would get a straight line
because the smaller two side lengths add to get the third, creating a 180◦ angle. (A triangle like
this is called degenerate). You need a set of lengths that connect to each other and enclose
some non-zero area.

Theorem 11.1 (Triangle Inequality). If a, b, and c are the side lengths of a triangle, then
they must satisfy
a+b>c
a+c>b
b + c > a.

1
Everaise Academy (2020) Math Competitions I

Simply put, any one side has to be less than the sum of the other two. In order to find the
range of acceptable values of the third side given the first two sides, just figure out two values
for which the triangle is degenerate.

Example 11.2. There exists a triangle with side lengths 2, 3, and an unknown side. If I
roll a fair 6 sided die, what is the probability that it can be used as the third side and form
a non-degenerate triangle?

Solution. What values could the 3rd side length have to turn the triangle into a straight line?
Well, if we subtract the lengths, we get 3 − 2 = 1 as a lower bound and if we add them, we get
3 + 2 = 5 as an upper bound. This means that we can have any length in the interval (1, 5)
(excluding the endpoints).
Since 1 and 5 cannot be the side lengths, we can only consider the numbers between them as
1
side lengths, namely 2, 3, and 4. Since these are three numbers on the die, there is a chance
2
that a die roll will give a valid side length. 

Although it may be called the Triangle Inequality, this inequality extends to n-sided polygons in
such a way that the sum of the smallest n − 1 side lengths must be larger than the length of the
longest side. You can use the same intuition about forming a line to understand why.

Example 11.3 (2017 AMC 10A). Joy has 30 thin rods, one each of every integer length
from 1 cm through 30 cm. She places the rods with lengths 3 cm, 7 cm, and 15 cm on a
table. She then wants to choose a fourth rod that she can put with these three to form a
quadrilateral with a positive area. How many of the remaining rods can she choose as the
fourth rod?

Solution. What rod values would make this quadrilateral a straight line?

Let’s start by adding all of the three side lengths together. This gives an upper bound of 25.
Furthermore, if we subtract the smaller two side lengths from the largest one, we get a lower
bound of 15 − 3 − 7 = 5. Thus, the fourth side length must be between 5 and 25 exclusive.
There are 19 such values.
You may think that we are finished here. However, note that although 7 and 15 are part of
this set of acceptable side lengths, they can’t be the fourth side length since they are already
used as one of the first three side lengths. and Joy only has one of each, so no repetitions are
allowed. Thus the answer is 19 − 2 = 17. 

§11.2 Right Triangles

There turns out to be a more specific relationship between the side lengths of a right triangle
beyond the Triangle Inequality.

Theorem 11.4 (Pythagorean Theorem). Given any right triangle with leg lengths a, b and
hypotenuse length c, we have

a2 + b2 = c2 .

2
Everaise Academy (2020) Math Competitions I

There are a multitude of ways to prove this, many of which you’ve probably seen. We’ll ask
you to prove this later in the handout.
A Pythagorean triple is a set of three integer values that form the sides of a right triangle.
For example, 3-4-5 is a Pythagorean triple because

32 + 42 = 52

is true, so there exists a right triangle with legs of length 3, 4 and a hypotenuse of length 5.
It helps to know some smaller Pythagorean triples as because it may save calculation time
and allow you to quickly recognize a right triangle in certain problems.

Example 11.5. Find some other Pythagorean triples with integer values less than 100.

Solution. Through inspection, we find some Pythagorean triples including, but not limited to,
the following:

3−4−5
6 − 8 − 10
5 − 12 − 13
8 − 15 − 17
7 − 24 − 25
9 − 40 − 41
10 − 24 − 26
20 − 21 − 29.

There are a few things worth noting:

• If a-b-c is a Pythagorean triple, then na-nb-nc is one as well, where n is a positive integer.
You can think of this either as scaling up the right triangle by a factor of n, or you can
just plug the scaled values into the pythagorean theorem and see that it works.
2 2
• Some of the Pythagorean triples we found can be expressed in the form n-( n 2−1 )-( n 2+1 ),
where n is an odd integer greater than 1. This pattern continues.
• Some of the Pythagorean triples we found can be expressed in the form 2n-(n2 − 1)-
(n2 + 1), where n is an even integer greater than 1. This pattern also continues.
It is not at all necessary to memorize the last two bullet points, but they are quite interesting
and may come in handy at times. Knowing a few of these common Pythagorean triples is also
quite handy.

3
Everaise Academy (2020) Math Competitions I

There are also some general triangle properties that can be derived from the Pythagorean
Theorem.

Proposition 11.6. For a right triangle 4ABC with D being the perpendicular height from
B with respect to AC,

DB 2 = (AD)(DC).

Proof. For ease of writing, let AD = x, DC = y, and DB = z. From the Pythagorean Theorem,
we have
AB 2 + BC 2 = AC 2
AB 2 + BC 2 = (x + y)2
AB 2 + BC 2 = x2 + y 2 + 2 · x · y.
From the Pythagorean Theorem, we also have the following two equations:
AB 2 = z 2 + x2
BC 2 = z 2 + y 2 .
Substituting these into the previous equation,
z 2 + x2 + z 2 + y 2 = x2 + y 2 + 2 · x · y
2z 2 = 2xy
z 2 = xy.
Thus, DB 2 = (AD)(DC). 

Now, let’s try a problem that utilizes the previous proposition:

Example 11.7 (2009 AMC10A, modified). Triangle ABC has a right angle at B. Point D
is the foot of the altitude from B, AD = 3, and DC = 4. What is BD?

Solution. Applying the previous proposition, we have

BD2 = (AD)(DC)
BD2 = 12
√
BD = 12
√
BD = 2 3.


4
Everaise Academy (2020) Math Competitions I

§11.3 Area of a Triangle

Now let’s investigate the area of a triangle. We’ll assume you already know how to find the
area of a rectangle (base × height).

Theorem 11.8 (Area of a Triangle). A triangle with a base length of length b and a height
of length h has area A given by
1
A = bh.
2

(Remember that height refers to the length of the altitude to the base). Something interesting
to note is that there is no unique choice of base and height. Choosing a different base and
repeating the calculation with the corresponding height will yield the same result.
These simple area formulas with some creativity are often enough to solve problems. Let’s
get into some examples!

Example 11.9 (2016 AMC10A). Find the area of the shaded region.

Solution. We can split the shaded regions into triangles as shown in the diagram below:

By symmetry, we can conclude that the height of each triangle is half of the corresponding
length or width of the rectangle. Thus,
1 1+7 1 1+4
A = (2)( )(1)( ) + (2)( )(1)( )
2 2 2 2
5
=4+
2
13
= .
2


Example 11.10 (Adapted from 2006 AMC 10B). In triangle ABC, cevians AD and BE are
drawn, intersecting at point F and creating three smaller triangles and a quadrilateral.
Given that the area of 4AEF is 3, the area of 4ABF is 7, and the area of 4BF D is 7,

5
Everaise Academy (2020) Math Competitions I

what is the area of the quadrilateral CEF D?

Solution. First, we can divide the quadrilateral in two triangles and label their areas x and y.

Now, compare the areas of triangle that have the same height to compare their bases.
• 4AF B and 4F DB have the same area. Suppose h1 is the perpendicular height from
point B to line segment AD. Then, by the Area formula
1 1
(h1 )(AF ) = (h1 )(F D)
2 2
AF = F D.

• 4AF B and 4AEF have area 7 and 3, respectively. Suppose h2 is the perpendicular
height from A to EB. Then,
1 3 1
(h2 )(EF ) = ( )( )(h2 )(F B)
2 7 2
3
EF = F B.
7

• Suppose h3 is the perpendicular height from point C to line segment AD. Because F D =
AF , we get
1
y = (h3 )(F D)
2
1
y = (h3 )(AF )
2
y = x + 3.

• Finally, we compare the areas of 4CEF and 4CF B. Suppose h4 is the perpendicular
height from C to EB. Then, the area of 4CEF is
1
x = (h4 )(EF ).
2
Finding the area of 4CF B and substituting EF = 73 F B in,

1
y + 7 = (h4 )(F B)
2
1 7
= (h4 )( )(EF )
2 3
7 1
= ( )( )(h4 )(EF )
3 2
7
= x.
3

6
Everaise Academy (2020) Math Competitions I

Now, we have two equations relating the two variables x and y:

y + 7 = 37 x


y =x+3

Using substitution,
7
x+3+7= x
3
7
x + 10 = x
3
4
10 = x
3
15
x= .
2
Then, y = x + 3 = 21
2 . This total area of the quadrilateral, which is equivalent to x + y, is what
we seek. Thus, the answer is 21 15
2 + 2 = 18. 

Remark. As seen in the previous example, comparing base lengths between triangles that
share a common height is a common theme among elementary area problems.

Now, let’s look at some examples that also incorporate the Pythagorean Theorem.

Example 11.11 (2018 AMC 10A). Farmer Pythagoras has a field in the shape of a right
triangle. The right triangle’s legs have lengths 3 and 4 units. In the corner where those
sides meet at a right angle, he leaves a small unplanted square S so that from the air it
looks like the right angle symbol. The rest of the field is planted. The shortest distance
from S to the hypotenuse is 2 units. What fraction of the field is planted?

Solution. This problem can be solved by splitting the large triangle into a clever set of regions
whose areas can be easily computed:

The addition of these lines makes sense because it allows us to utilize the information on the
distance from the top-right corner of the square to each side of the triangle.
Computing the area of the large triangle directly, we get
1 1
A = bh = (3)(4) = 6.
2 2

Then we compute the area of the large triangle as the sum of multiple regions. The 3 regions
include the large triangle with height 2 and hypotenuse as the base, the smaller triangle with

7
Everaise Academy (2020) Math Competitions I

height x and base 4, and the smallest triangle with height x and base 3. Using the Pythagorean
Theorem, we find that the hypotenuse is 5. Thus,
1 1 1
A = (3)x + (4)(x) + (5)(2)
2 2 2
3
6 = x + 2x + 5
2
7
1= x
2
2
= x.
7
4
The area of S is x2 = 49 . Then,
4
49 2
= .
6 147
2 145
of the field is not planted. Therefore, the planted portion is 1 − 147 = 147 .


Exercise 11.12. Prove the Pythagorean Theorem (there are many ways to do so).

Hint: One proof utilizes the diagram above and areas.

§11.4 Triangle Similarity and Congruence

Similarity and congruence are concepts we can use to find more information about triangles
using what we already know.
If two triangles are congruent, then they are identical with respect to the three angles and
the three side lengths. Even if one triangle of a congruent pair is rotated or flipped, congruency
remains. There are several ways to determine if two triangles are congruent:
1. Side-Side-Side: If all three pairs of corresponding sides are equal, then the angles are
determined and the triangle is congruent.
2. Side-Angle-Side: If two sides and their including angle are the same across two triangles,
then the third side is determined and congruency is determined.
3. Angle-Angle-Side: If you are given two angles, the third is determined. If any side is
the same as in another triangle, then congruency is determined.
4. Hypotenuse-Leg: If both of these sides are known to be of same length of corresponding
parts of other right triangles, then the third leg is determined and congruence is proven.
Often, congruence is not found on its own in geometry problems. Instead, you will see it
coupled with the concept of similarity.

8
Everaise Academy (2020) Math Competitions I

If two triangles are similar, then their angles are identical. They must be the same shape,
but not necessarily the same size. Here are some ways to determine similarity:
1. Side-Side-Side: If ratios of all three pairs of corresponding sides are all equal, then the
triangles are similar. However, this is not very common to spot.
2. Side-Angle-Side: If the corresponding angles are the same and both pairs of its neigh-
boring sides are proportional with the same ratio, then the triangles are similar.
3. Angle-Angle: This is the most common indicator. If two corresponding pairs of angles
are the same, the third angle should also be the same, thus determining similarity.
4. Hypotenuse-Leg: This only applies to right triangles. If the ratio of hypotenuse to one
leg is the same, then the other leg is determined for both triangles and the angles are
determined as well.

Theorem 11.13. If two triangles ABC and DEF are similar ABC ∼ DEF (in that
specific order of angles), then

∠A = ∠D, ∠B = ∠E, ∠C = ∠F

and
AB BC AC
= = = k.
DE EF DF
We call k the scale factor.

Remark. This property isn’t confined to sides either. For example, if triangle ABC has
circumradius R and triangles ABC and DEF are similar with scale factor k, then triangle
DEF has circumradius R k . Or, if the distance between A and the orthocenter of triangle
ABC is x, then the distance between D and the orthocenter of triangle DEF is xk .
For area, the scale factor is squared: if the area of triangle ABC is S, then the area of
triangle DEF is kS2 . You can prove this using our area formulas.

We will demonstrate the similarity principle below:

Example 11.14 (Adapted from AHSME 1986). In 4ABC, AB = 8, BC = 7, CA = 6 and

side BC is extended to a point P so that 4P AB is similar to 4P CA. What is the length
of P C?

Solution. As always, draw a diagram and assign variables to unknown lengths:

9
Everaise Academy (2020) Math Competitions I

y C

A 8 B

Since we are given that 4P AB ∼ 4P CA, we can form the following proportions and con-
clusions:
PA PC y x
= =⇒ = =⇒ 4x = 3y
AB CA 8 6
PB PA x+7 y
= =⇒ = =⇒ 8y = 6x + 42
AB CA 8 6
At this point, we have two equations and two unknown variables, which we know how to solve!
We get that P C = x = 9, and this is our answer.


Now we’ll move on to a problem that gives insight about areas.

Example 11.15 (Adapted from 1984 AIME). A point P is chosen in the interior of 4ABC
such that when lines are drawn through P parallel to the sides of 4ABC, the resulting
smaller triangles t1 , t2 , and t3 formed by the lines have areas 4, 9, and 49, respectively.
Find the area of 4ABC.

Solution. As always with geometry problems, draw a diagram.

t1 t2

t3

B C

These smaller triangles are similar to each other and to 4ABC due to angle similarity.
Convince yourself why this is true.
Looking back at the remark regarding similarity, we see that by taking the square root of each
of these areas, we get their relative side length proportions: 2, 3, and 7. Now, take the bottom
side of each smaller triangle and call them 2x, 3x, and 7x. This means that BC = 12x. This is
because the existence of parallelograms in the figure means that opposite sides are congruent.
Make sure you agree with this line of reasoning. Now, we get that the side length proportions

10
Everaise Academy (2020) Math Competitions I

(including 4ABC at the end) are 2 : 3 : 7 : 12. Square each of these to get back to area
proportions, and we get that the area of 4ABC = 144.


Now for a trickier one. The following problem also highlights an important strategy: If you
can’t see many congruent/similar triangles in a diagram, draw some lines and make some up!

Example 11.16 (2014 AMC 12B). Let ABCD be a square of side length 1. Points E and
F lie on AB and CD respectively such that CF EB is a rectangle. Points K, H, G, and
J lie on EF , F D, DA, and AE respectively such that the rectangles JKHG and EBCF
are congruent. What is BE?

Solution. First things first: draw a diagram. Call BE = x. Now we should draw some lines.

Specifically, we can draw a line from H that intersects AB at a right angle. Call this point L.
Now, draw line HJ. Note that
4HKJ ∼ = 4HLJ
because of the Hypotenuse-Leg congruence theorem and

4KEJ ∼
= 4GDH

via the ASA congruence theorem. (Can you figure out why?) Now, since corresponding parts
of congruent triangles are congruent, JL = 1. Now,
1
BE = JL, EJ = HD = LA =⇒ BE + EJ = JL + LA =⇒ BJ = JA =⇒ BJ = JA =
2
√
1 3
Since we know that GJ = 1 and JA = , GA = because of the Pythagorean Theorem.
√ 2 2
2− 3
Thus, KE = DG = . Now, focusing on right triangle KEJ, we can solve for x using
2
the Pythagorean Theorem:
√ !2  2
2− 3 1
(KE)2 + (EJ)2 = (KJ)2 =⇒ + − x = x2 .
2 2
√
Through some algebra, we get that x = 2 − 3. 

11
Everaise Academy (2020) Math Competitions I

Example 11.17 (2019 AIME II). Triangle ABC has side lengths AB = 120, BC = 220, and
AC = 180. Lines `A , `B , and `C are drawn parallel to BC, AC, and AB, respectively, such
that the intersection of `A , `B , and `C with the interior of 4ABC are segments of length
55, 45, and 15, respectively. Find the perimeter of the triangle whose sides lie on `A , `B ,
and `C .

Solution. Label points as in the diagram on the next page. Note that triangle AA1 A2 is similar
to triangle ABC. Thus,
AB BC AC 120 220 180
AA1 = A1 A2 = AA2 =⇒ AA1 = 55 = AA2 .

This means that AA1 = 30 and AA2 = 45. By the same reasoning, we find BB1 = 55,
BB2 = 30, CC1 = 22.5, and CC2 = 27.5 — make sure to confirm for yourself that this works!
Now, by subtracting lengths, we get that A1 B2 = 60, B1 C2 = 137.5, and C1 A2 = 112.5.
But there are more similar triangles to work with — one example being triangles AA1 A2 and
B2 A1 Z. Pause for a moment and see if you can find the others.
Because of this, we have ZB ZA1 A1 B2
AA2 = A1 A2 = A1 A , from which we find ZA1 = 110 and ZB2 = 90.
2

From the other similar triangle pairs, we deduce that Y A2 = 137.5, Y C1 = 75, XB1 = 112.5,
and XC2 = 75. We now have all of the lengths to complete the problem!
Our answer is

XY + Y Z + ZX = ZA1 + A1 A2 + A2 Y + Y C1 + C1 C2 + C2 X + XB1 + B1 B2 + B2 Z
= 110 + 55 + 137.5 + 75 + 15 + 75 + 112.5 + 45 + 90
= 715

and we are done. 

§11.5 Homework Problems

Problem 11.1 (Adapted from 2011 AMC 8). [3] Let A be the area of the triangle with sides
of length 25, 25, and 30. Let B be the area of the triangle with sides of length 25, 25, and 40.
A
What is ?
B
Problem 11.2 (Adapted from 2002 AMC 10B). [5] A regular octagon ABCDEF GH has sides
√
of length two. The area of 4ADG can be expressed in the form a + b c The square of the area
of the triangle with sides a, b, and c can be expressed ad d (which can be 0 if the triangle is
impossible or degenerate). Find 10000d.

12
Everaise Academy (2020) Math Competitions I

Problem 11.3 (2018 AMC 8). [4] In 4ABC, a point E is on AB with AE = 1 and EB = 2.
Point D is on AC so that DE k BC and point F is on BC so that EF k AC. The ratio of the
area of CDEF to the area of 4ABC? can be expressed as ab . Find a + b.

Problem 11.4 (2017 CMIMC). [3] Let ABC be a triangle with ∠BAC = 117◦ . The angle
bisector of ∠ABC intersects side AC at D. Suppose 4ABD ∼ 4ACB. Compute the measure
of ∠ABC, in degrees.

Problem 11.5 (2000 AMC 12). [5] Through a point on the hypotenuse of a right triangle, lines
are drawn parallel to the legs of the triangle so that the triangle is divided into a square and
two smaller right triangles. The area of one of the two small right triangles is m times the area
of the square. If the ratio of the area of the square to the area of the other small right triangle
can be expressed as am where a is an integer, what is a?

Problem 11.6 (2007 AMC 10B). [5] Right 4ABC has AB = 3, BC = 4, and AC = 5. Square
XY ZW is inscribed in 4ABC with X and Y on AC, W on AB, and Z on BC. If the side
length of the square can be expressed as the common fraction m
n , what is m + n?

Problem 11.7 (2017 AMC 10B). [6] The diameter AB of a circle of radius 2 is extended to a
point D outside the circle so that BD = 3. Point E is chosen so that ED = 5 and the line ED
is perpendicular to the line AD. Segment AE intersects the circle at point C between A and E.
If the area of 4ABC can be expressed as ab , where a and b are positive integers, what is a + b?

Problem 11.8 (Adapted from 2017 AMC 10A). [6] A square with side length x is inscribed
in a right triangle with sides of length 3, 4, and 5 so that one vertex of the square coincides
with the right-angle vertex of the triangle. A square with side length y is inscribed so that one
side of the square lies on the hypotenuse of the triangle. The ratio xy , when simplified, gets the
common fraction ab . What is a2 − b2 ?

Problem 11.9 (2020 AMC 10B). [6] In square ABCD, points E and H lie on AB and DA,
respectively, so that AE = AH. Points F and G lie on BC and CD, respectively, and points
I and J lie on EH so that F I ⊥ EH and GJ ⊥ EH. Triangle AEH, quadrilateral BF IE,
quadrilateral DHJG, and pentagon F CGJI each has area 1. F I 2 can be expressed in the form
√
a − b c. What is a + b + c?

Problem 11.10 (2006 AIME II). [7] Square ABCD has sides of length 1. Points E and F are
on BC and CD, respectively, so that 4AEF is equilateral. A square with vertex B has sides
that are parallel
√ to those of ABCD and a vertex on AE. The length of a side of this smaller
square is a−c b , where a, b, and c are positive integers and b is not divisible by the square of any
prime. Find a + b + c.

Problem 11.11 (2018 AIME I). [8] In 4ABC, AB = AC = 10 and BC = 12. Point D lies
strictly between A and B on AB and point E lies strictly between A and C on AC so that
p
AD = DE = EC. Then AD can be expressed in the form , where p and q are relatively prime
q
positive integers. Find p + q.

Problem 11.12 (2015 AIME I). [7]

ABCD is a square. Point E is the midpoint of AD. Points F and G lie on CE, and H and
J lie on AB and BC, respectively, so that F GHJ is a square. Points K and L lie on GH, and
M and N lie on AD and AB, respectively, so that KLM N is a square. The area of KLM N is
99. Find the area of F GHJ.

13
Everaise Academy (2020) Math Competitions I

Provide a detailed solution, or explain what progress you made, even if you
didn’t get an answer.

14
12 Circles

In this handout, we’ll go over many important topics related to circles.

§12.1 Arcs

Proposition 12.1 (Prerequisites). Given a circle with radius r and a point A outside the
circle,
• The area of the circle is πr2 .
• The circumference of a circle is 2πr.
• A line passing through A that is tangent to the circle is perpendicular to the radius
passing through the point of tangency.

For the sake of simplicity, we will now include most of the properties of arcs below. These
properties are fairly simple, but they can be used to prove many stronger statements.

Proposition 12.2 (Properties of Arcs). Let XY

¯ denote the length of arc XY . Given a circle
with center O, radius r, and four points A, B, C, D on this circle,
• (Inscribed Angle Theorem) ∠ABC = 12 ∠AOC.
• AB = CD if and only if AB
˜ = CD.
˜
• Given that θ is ∠AOB measured in radians and θ0 is ∠AOB measured in degrees,

˜ = rθ = 2πr θ0
AB
360
In other words, in a circle, the length of an arc is proportional to the central angle.

Now that we know have this information, we are ready to tackle some problems! Make sure you
understand what the statements above are saying, because that will really help with solving
problems about arcs and arc lengths!

Example 12.3 (2011 AMC 12B). Two tangents to a circle are drawn from a point A. The
points of contact B and C divide the circle into arcs with lengths in the ratio 2 : 3. What
is the degree measure of ∠BAC?

Solution. The first thing we do is draw a diagram and label the important points.
Let the center of the circle be O. Then, we see that the question is asking for an angle when
we are only given the length of an arc. This is where Part 3 of Proposition 12.2 comes in.
However, we can’t use that directly, as we need to simplify the expression first. We know
that ∠BOC = 2θ for some θ. We also know 2θ + 3θ = 360◦ =⇒ θ = 72◦ . This means that
∠BOC = 144◦ .
We know what ∠BOC is now, but we want to relate that to ∠BAC. How do we do that? If
we look at the diagram for some time, we see that

1
Everaise Academy (2020) Math Competitions I

ABOC is a quadrilateral, so ∠ABO + ∠BOC + ∠OCA + ∠CAB = 360◦ =⇒ 90◦ + 144◦ +

90◦ + ∠CAB = 360◦ =⇒ ∠CAB = ∠BAC = 36◦ . 

Example 12.4 (2007 AMC 12A). A star-polygon is drawn on a clock face by drawing a
chord from each number to the fifth number counted clockwise from that number. That
is, chords are drawn from 12 to 5, from 5 to 10, from 10 to 3, and so on, ending back at
12. What is the degree measure of the angle at each vertex in the star polygon?

Solution. Once again, we start with drawing a diagram.

It should be obvious that every angle is the same, so let’s look at the angle between 1, 6, and
11. Call the point at 1 A, the point at 6 B, the point at 11 C, and the center of the clock O. We
want to calculate the angle of an inscribed angle, so the Inscribed Angle Theorem. However,
we can’t use that just yet, as we need to simplify first.

2
Everaise Academy (2020) Math Competitions I

˜ is 1 of the circumference of the clock (see diagram!), so ∠AOC =

We know that AC 6
1
6 · 360◦ =
60 . By the Inscribed Angle Theorem, ∠ABC = 21 ∠AOC = 21 · 60◦ = 30◦ .
◦ 

Example 12.5 (2002 AMC 12A). A 45◦ arc of circle A is equal in length to a 30◦ arc of
circle B. What is the ratio of circle A’s area and circle B’s area?

Solution. As always, we label points first. Let rA be the radius of circle A and let rB be the
radius of circle B. Once again, we want to relate the central angle to the length of an arc, so we
will probably use Part 3 of Proposition 12.2.
45 30 1 1 3
2πrA · = 2πrB · =⇒ rA · = rB · =⇒ rB = · rA .
360 360 8 12 2
The ratio of A’s area to B’s area is
πrA 2
,
π( 3r2A )2
which is ( 32 )2 = 49 . Make sure you don’t misread the question and forget to square 23 ! 

Arcs may seem a bit intimidating at first, but they’re really not that bad. Once you get used
to them and are used to using Proposition 12.2, they should become a piece of cake!
Exercise 12.6. Prove all parts of Proposition 12.2.

§12.2 Cyclic Quadrilaterals

Definition 12.7. A cyclic quadrilateral is a quadrilateral that can be inscribed in a circle.

Theorem 12.8 (Cyclic Quadrilaterals). Given a convex quadrilateral ABCD, the following
statements are all equivalent:
1. ABCD is cyclic.
2. ∠ABD = ∠ACD.
3. ∠ABC + ∠CDA = 180◦ .

This theorem may seem insignificant, but it can be applied in many different ways. This is
arguably the most important theorem when trying to angle chase. In most cases, the problem
may not seem to have a cyclic quadrilateral, but if you look hard enough (don’t literally stare
at a piece of paper!), oftentimes you will find one that will help you solve a problem.
Exercise 12.9. Prove that a trapezoid is cyclic if and only if it is isosceles.

Example 12.10 (COMC 2019). In triangle KLM , let points G and E be on segment LM
so that ∠M KG = ∠GKE = ∠EKL = α. Let point F be on segment KL so that GF
is parallel to KM . Given that KF EG is an isosceles trapezoid and that ∠KLM = 84◦ ,
determine α.

3
Everaise Academy (2020) Math Competitions I

Solution. When we are trying to find angles, the first thing we should do is try to find cyclic
quadrilaterals. We are given that KF EG is an isosceles trapezoid, so it is also cyclic. This
means that ∠F GE = ∠F KE = α (By Part 2 of Theorem 12.8). It may seem like we are stuck,
until we read the question again and realize that we are given that GF is parallel to KM , so
∠GF L = ∠KM L = 3α.
We also know that the sum of the three angles of 4GF L is 180◦ , so 3α + α + 84◦ = 180◦ =⇒
4α = 96◦ =⇒ α = 24◦ . 

This problem taught us that if we are stuck at some point in a problem, it’s never a bad idea
to read the problem statement again and see what condition we haven’t used yet.
While Theorem 12.8 is very powerful, when solving computational problems, it is often not
enough. Here is another useful theorem that will make solving problems much easier.

Theorem 12.11 (Ptolemy’s Theorem). Given any four points,

AB · CD + BC · DA ≥ AC · BD,

where equality holds if and only if ABCD is a cyclic quadrilateral.

This theorem can seem intimidating at first, but it is very useful and worth remembering.

Example 12.12 (2004 AMC 10B). In triangle ABC we have AB = 7, AC = 8, BC = 9.

Point D is on the circumscribed circle of the triangle so that AD bisects angle BAC. What
AD
is the value of CD ?

Solution. Let CD = a and AD = b (labelling the points!).

4
Everaise Academy (2020) Math Competitions I

We know that we want to use Ptolemy’s because the question deals with lengths, not just
angles. AD bisects ∠BAC, so ∠BAD = ∠DAC =⇒ BD ˜ = DC ˜ =⇒ BD = a. By Ptolemy’s
Theorem,

AB · CD + AC · BD = AD · BC =⇒ 7a + 8a = 9b =⇒ 15a = 9b.
AD b 1 15
To find CD = a we simply multiply both sides of the equation by 9a to get 9 = 35 . 

Of course, there are many other solutions, and you don’t have to Ptolemy’s Theorem. In fact,
try finding a solution that does not involve Ptolemy’s!

Example 12.13 (2013 AMC 12B). In triangle ABC, AB = 13, BC = 14, and CA = 15.
Distinct points D, E, and F lie on segments BC, CA, and DE, respectively, such that
AD ⊥ BC, DE ⊥ AC, and AF ⊥ BF . The length of segment DF can be written as m n,
where m and n are relatively prime positive integers. What is m + n?

5
Everaise Academy (2020) Math Competitions I

Solution. We first solve for AD, because it is fairly easy to find and it may make the problem
much simpler. Using Heron’s Formula, we know that the area of 4ABC is
1 p
· 14 · AD = (21)(6)(7)(8) = 84,
2
so AD = 12.
We can look at the diagram and try to find cyclic quadrilaterals. After eyeballing some
quadrilaterals and trying to prove whether or not they are cyclic, we find that AF DB is cyclic
because ∠AF B = ∠ADB = 90◦ , so ∠ABF = ∠ADF =⇒ 4ABF ∼ 4ADE ∼ 4ACD (by
AA).
By the Pythagorean Theorem, DC = 9, so
AB AC 13 15 13 · 3
= =⇒ = =⇒ BF = .
BF CD BF 9 5
Similarly,
AB AC 13 15 13 · 4
= =⇒ = =⇒ AF = .
AF AD AF 12 5
Now, we see that we have almost every side of the cyclic quadrilateral (other than DE, the side
we want). This is a sign that Ptolemy’s Theorem may come in handy. By Ptolemy’s,
13 · 4 13 · 3
AB · DF + AF · BD = AD · BF =⇒ 13 · DF + · (14 − 9) = 12 · .
5 5
When we divide both sides by 13, we get
36 16
DF + 4 = =⇒ DF = .
5 5
We are asked to find m + n = 16 + 5 = 21. 

This problem looks intimidating at first, but it’s not too bad once you have some lengths down.
In the future, if there is a problem with a bunch of unknowns and you’re stuck, it’s not a bad
idea to try and find some random lengths and see where that gets you (of course, only do this
is the lengths are really easy to find!). Cyclic quadrilaterals are something one needs to take a
while to get used to, but once you are familiar with them they can be very helpful!
Exercise 12.14. Two different points, C and D, lie on the same side of line AB so that 4ABC
and 4BAD are congruent with AB = 9, BC = AD = 10, and CA = DB = 17. The intersection
of these two triangular regions has area m
n , where m and n are relatively prime positive integers.
Find m + n.

§12.3 Power of a Point

Theorem 12.15 (Power of a Point). This theorem has two parts.

• Fix two chords AB, CD of a circle. Let them intersect at a point P inside the circle.
Then P A · P B = P C · P D.
• Fix two lines r, s and a circle ω, none of which are parallel to each other. r and s
pass through a point P outside of the circle.
Let r intersect ω at A1 and B1 and s intersect ω at A2 and B2 . Then P A1 · P B1 =
P A2 · P B2 .

6
Everaise Academy (2020) Math Competitions I

We let the power of each point be the product in both parts of this theorem. The power of
points on the circle is 0.

Proof. We prove each part of the theorem separately.

• We claim that 4P AC ∼ 4P BD. Note that ∠AP C = ∠BP D because of vertical angles,
and ∠CAP = ∠BDP and ∠ACP = ∠DBP from inscribed angles. Therefore, 4P AC ∼
4P DB, and it follows that P A/P C = P D/P B =⇒ P A · P B = P C · P D.
• We claim that 4P A1 A2 ∼ 4P B2 B1 . Note that ∠P A1 A2 = 180◦ − ∠B1 A1 A2 = ∠P B2 B1 .
It’s quite obvious that ∠A1 P A2 = ∠B1 P B2 as well, so by AA similarity we have that
4P A1 A2 ∼ 4P B2 B1 . Our length ratios follow immediately.


Note. If the circle is tangent, we sometimes say that it intersects the same point twice. Therefore,
it follows that if a line from P is tangent to the circle at C, then the power of P is P C 2 .

Example 12.16 (ARML). In a circle, chords AB and CD intersect at R. If AR : BR = 1 : 4

and CR : DR = 4 : 9, find the ratio AB : CD.

Solution. Let AR = x, BR= y. Wethen have that AB = 5x, CD = 13y. It suffices to find
5x 5 x
, which is equivalent to . From Power of a Point, we have that x · 4x = 4y · 9y,
13y 13 y

7
Everaise Academy (2020) Math Competitions I

so it follows that

4x2 =36y 2
 2
x
=⇒ =9
y
x
=⇒ =3.
y
5 15
Hence, the desired ratio is ·3= . 
13 13

Example 12.17 (2013 AMC 10A). In 4ABC, AB = 86, and AC = 97. A circle with center
A and radius AB intersects BC at points B and X. Moreover BX and CX have integer
lengths. What is BC?

Solution. Let our circle meet line AC at points Y, Z, such that Y is closer to C. It follows
that AY = AZ = 86. Therefore, CY = 97 − 86 = 11 and CZ = 11 + 86 · 2 = 183. Hence, from
Power of a Point, BC · CX = 11 · 183.
As BX and CX must be integral, so must BC. Therefore, we need two integers a, b such
that a > b and ab = 3 · 11 · 61. Because a is equal to BC, we must also have a + 86 > 97 and
that a < 86 + 97 = 183. Trying all solution pairs, it can be found that a = 61 is the only valid
solution. 

Example 12.18 (2019 AIME I). In convex quadrilateral KLM N side M N is perpendicular
to diagonal KM , side KL is perpendicular to diagonal LN , M N = 65, and KL = 28.
The line through L perpendicular to side KN intersects diagonal KM at O with KO = 8.
Find M O.

Solution. Let the foot of the altitude from L to KN be P . Note that N P OM is cyclic
because ∠OP N = ∠OM N = 90◦ . Also note that (N P L) is tangent to KL. Hence, we
2
have that KO · KM = 8 · KM = KP · KN = KL2 = 282 . Hence, KM = 288 = 98, so
M O = 98 − 8 = 90. 

§12.4 Homework Problems

Problem 12.1 (2007 AMC 10B). [3] The point O is the center of the circle circumscribed
about 4ABC, with ∠BOC = 120◦ and ∠AOB = 140◦ , as shown. What is the degree measure
of ∠ABC?

8
Everaise Academy (2020) Math Competitions I

Problem 12.2 (2020 AMC 12B). [4] In unit square ABCD, the inscribed circle ω √intersects
CD at M, and AM intersects ω at a point P different from M. Given that AP = ba , where
this fraction is simplified, what is a + b?

Problem 12.3 (1971 Canada). [4] DEB is a chord of a circle such that DE = 3 and EB = 5.
Let O be the center of the circle. Join OE and extend OE to cut the circle at C. Given EC = 1,
find the radius of the circle.

√
Problem 12.4 (2020 AMC 12B). [6] Let AB be a diameter in a circle of√radius 5 2. Let CD
be a chord in the circle that intersects AB at a point E such that BE = 2 5 and ∠AEC = 45◦ .
What is CE 2 + DE 2 ?

Problem 12.5 (2005 AMC 12A). [6] Let AB be a diameter of a circle and C be a point on
AB with 2 · AC = BC. Let D and E be points on the circle such that DC ⊥ AB and DE is
a second diameter. If the ratio of the area of 4DCE to the area of 4ABD is ab , where this
fraction is simplified, what is a + b?

Problem 12.6 (2005 AIME I). [6]

Triangle ABC has BC = 20. The incircle of the triangle evenly trisects the median AD. If
√
the area of the triangle is m n where m and n are integers and n is not divisible by the square
of a prime, find m + n.

Problem 12.7 (2017 AMC12B). [7] The diameter AB of a circle of radius 2 is extended to a
point D outside the circle so that BD = 3. Point E is chosen so that ED = 5 and line ED
is perpendicular to line AD. Segment AE intersects the circle at a point C between A and E.
Given that the area of 4ABC is ab , where this fraction is simplified, what is a + b?

Problem 12.8 (2014 AMC 12B). [10] Let ABCDE be a pentagon inscribed in a circle such
that AB = CD = 3, BC = DE = 10, and AE = 14. The sum of the lengths of all diagonals of
ABCDE is equal to m
n , where m and n are relatively prime positive integers. What is m + n?

9
Everaise Academy (2020) Math Competitions I

Problem 12.9 (2008 iTest). [8] Points C and D lie on opposite sides of line AB. Let M and
N be the centroids of 4ABC and 4ABD respectively. If AB = 841, BC = 840, AC = 41,
AD = 609, and BD = 580, find the sum of the numerator and denominator of the value of M N
when expressed as a fraction in lowest terms.
Provide a detailed solution, or explain what progress you made, even if you
didn’t get an answer.

10
13 Triangles II

In this handout, we’ll build on our knowledge of triangles and tackle some more advanced
concepts.

§13.1 Triangle Centers

In this section, we describe the properties of the most common triangle centers, i.e. special
points associated with each triangle. The proofs are just as important to understand as the
properties themselves, and provide useful intuition, so we strongly urge you try to do every
exercise in the section.

As usual, and especially with geometry, its important to have pencil and paper handy. It’s
almost impossible (or at least very ineffective) to go through examples without drawing the
diagrams yourself to follow along.

§13.1.1 Circumcenter

Definition 13.1. The perpendicular bisector of a segment AB is the locus (i.e. collection)
of points C such that AB = AC.

Exercise 13.2. Show that the perpendicular bisector of a segment AB is the line perpendicular
to AB passing through its midpoint.
Hint: Why does the midpoint M always lie on the perpendicular bisector? Then show that any other
point on the perpendicular bisector lies on the line perpendicular to AB and passing through M using congruent
triangles.

Theorem 13.3 (Circumcenter). In any triangle 4ABC, the perpendicular bisectors of its
sides concur at a point O, the circumcenter of the triangle.

B C

Exercise 13.4. Prove that the perpendicular bisectors concur at a point O.

Hint: Let two perpendicular bisectors meet at O, and then show that the third one must pass through O as
well.

1
Everaise Academy (2020) Math Competitions I

Exercise 13.5. Show that there exists a circle with center at O passing through A, B, and C,
4ABC’s circumcircle.
Hint: The radius of the circle should be AO. Why must this length also be the distance from O to B and
O to C?

Exercise 13.6. Show that this implies that three points uniquely define a circle; that is, show
that there exists exactly one circle passing through any three given points.
Hint: Use a proof by contradiction. A circle is different from another circle if and only if they either have
different centers, radii, or both, so assume that there exists two circles passing through some points A, B, and
C, and show that they must be the same.

§13.1.2 Centroid

Definition 13.7. The median from a vertex A of a triangle 4ABC is the line from A to the
midpoint of BC.

P N
G

B M C

Definition 13.8. The medial triangle of a triangle 4ABC is the triangle formed by the
midpoints of its sides.

Proposition 13.9. In a triangle 4ABC, if the midpoints of BC, CA, and AB are M, N,
and P, respectively, then 4M N P ∼
= 4AN P ∼= 4BP M ∼ = 4CM N, and all four triangles
are similar to 4ABC.

Exercise 13.10. Prove that 4M N P ∼

= 4AN P.
Exercise 13.11. Prove that 4M N P ∼ 4ABC.

Theorem 13.12 (Centroid). In a triangle 4ABC, the medians of the triangle meet at a
point G, the centroid of the triangle.

Proof. Let BE intersect CF at G. Since 4AF E ∼ 4ABC, F E k BC. Thus 4BCG ∼ 4EF G
with a ratio of BC BG
EF = 2, so GE = 2.
BG0
Similarly let BE intersect AD at G0 . Repeating the above yields G0 E = 2. Thus G and G0
are the same point, and the medians are concurrent.

2
Everaise Academy (2020) Math Competitions I

F E
G

B C
D

Corollary 13.13. Let M be the midpoint of BC. Then the centroid G divides AM into a
2 : 1 ratio.

Exercise 13.14. Prove that the medians divide 4ABC into 6 triangles with equal area.
Hint: Triangles 4BGM and 4CGM have the same height and the same base, so their areas must be equal.
On the other hand, 4AM C must have triple the area of 4CGM due to the centroid ratio: their bases are the
same but by similar triangles, the ratio of the height from A to BC to the height from G to BC is equal to the
ratio of AM to GM.

§13.1.3 Incenter

Definition 13.15. Let A, B, and C be points in the plane. The internal angle bisector of
angle ∠BAC is the locus of points D such that D lies inside the angle ∠BAC (i.e. B and C
are on opposite sides of line AD) and the distances from D to AB and AC are equal.

We refer to the internal angle bisector when we say angle bisector unless otherwise specified
(there also exists an external angle bisector, which we will not talk about in this handout).
Exercise 13.16. Show that every point D lying on the angle bisector of an angle ∠ABC must
satisfy the relation ∠BAD = ∠DAC.
Hint: Congruent triangles are helpful.

Exercise 13.17. Show that the angle bisector of an angle ∠BAC must be a line.
Hint: To show that an angle bisector of ∠ABC is a line, fix a point D on it and show that any other point
on the angle bisector must lie on line AD.

Theorem 13.18 (Incenter). In a triangle 4ABC, the internal angle bisectors of ∠BAC,
∠ACB, and ∠CBA meet at a point I, the incenter of the triangle.

E
F
I

B D C

3
Everaise Academy (2020) Math Competitions I

Definition 13.19. Let the incircle of 4ABC touch BC at D. Then D is the A-intouch point,
and the triangle formed by the A-intouch point and the similarly defined B- and C-intouch
points is the contact triangle or intouch triangle of 4ABC.

Lemma 13.20. Let the incircle of 4ABC be tangent to BC, CA, and AB at D, E, and
F , respectively. Then AE = AF = s − a, BF = BD = s − b, and CD = CE = s − c.

Exercise 13.21. Prove that AE = AF = s − a.

Hint: Using the fact that the tangents from a point to a circle are equal, set x = AE = AF, y = BF =
BD, z = CD = CE, and solve the resulting system of 3 equations for x, y, z.

§13.1.4 Orthocenter

Definition 13.22. Let A, B, C be points in the plane. The A-altitude of 4ABC is the line
passing through A perpendicular to BC. We call the foot of the A-altitude the point where the
A-altitude intersects BC.

E
F
H

B D C

Definition 13.23. In 4ABC, if D, E, and F are the feet of the altitudes from A, B, and C,
respectively, then 4DEF is the orthic triangle of 4ABC.

Proposition 13.24. Let D, E, and F be the feet of the altitudes from A, B, and C, respec-
tively. Then BCEF, CAF D, and ABDE are cyclic quadrilaterals.

Exercise 13.25. Prove that BCEF is cyclic; the other two follow from symmetry.

Theorem 13.26. In a triangle 4ABC, the A-altitude, B-altitude, and C-altitude meet at
a point H, the orthocenter of the triangle.

Exercise 13.27. Let the feet from the A-, B-, and C-altitudes be D, E, and F, respectively. If
the B- and C-altitudes meet at H, then prove that AEHF is cyclic.
Exercise 13.28. Prove that the altitudes concur at a point H.
Hint: At this point it should sound like a broken record. Let the B- and C- altitudes meet at H, then use
the previous exercise along with some angle chasing to show that A, H 0 , D must be collinear.

4
Everaise Academy (2020) Math Competitions I

§13.1.5 Examples

We begin with a nice lemma that is often a convenient shortcut.

Lemma 13.29. In triangle 4ABC, let O, H, and I be the circumcenter, orthocenter, and
incenter, respectively. Then
• ∠BOC = 2∠BAC if ∠BAC is acute and ∠BOC = π − 2∠BAC if ∠BAC is obtuse.
• ∠BHC = π − ∠BAC
• ∠BIC = π
2 + ∠BAC.

Proof.

• This follows immediately from the Inscribed Angle Theorem.

• If E and F are the feet of the altitudes from B and C, this follows immediately from the
fact that AEHF is cyclic.
• Since ∠IBC = 12 ∠ABC and ∠ICB = 21 ∠BCA, we find that
π  π
∠BIC = π − ∠IBC − ∠ICB = π − − ∠BAC = + ∠BAC.
2 2


Next, we present two applications of the centroid ratio.

Example 13.30 (2018 AMC 10B). Line segment AB is a diameter of a circle with AB = 24.
Point C, not equal to A or B, lies on the circle. As point C moves around the circle, the
centroid (center of mass) of 4ABC traces out a closed curve missing two points. To the
nearest positive integer, what is the area of the region bounded by this curve?

Solution. Let O be the center of the circle and G be the centroid of 4ABC. Since O is also
the midpoint of AB and thus lies on CG, we’re motivated to make use of the 2 : 1 ratio. CO
always has length 12, so it follows that GO always has length 4. This means that the locus of
G is a circle with center O and radius 4 by definition, so the area is 16π ≈ 50.

A O B

Example 13.31 (2020 AMC 10A). Triangle AM C is isoceles with AM = AC. Medians
M V and CU are perpendicular to each other, and M V = CU = 12. What is the area of
4AM C?

Solution. Let M V and CU meet at the centroid G and let T be the midpoint of BC. Since
4M T C and 4AM C share the same base and GT : AT = 1 : 3, it suffices to find the area of

5
Everaise Academy (2020) Math Competitions I

4M T C and multiply it by 3. But we know that it is a right triangle and that M G = CG =

2 8·8
3 · 12 = 8, so [AM C] = 3[M T C] = 3 · 2 = 96.

U V

M C

Exercise 13.32. In a triangle 4ABC, show that the B- and C-medians are perpendicular if and
only if b2 + c2 = 5a2 . Hint: Use the Pythagorean Theorem.

Example 13.33 (Brazil 2007). Let ABC be a triangle with circumcenter O. Let P be
the intersection of straight lines BO and AC and ω be the circumcircle of triangle AOP .
Suppose that BO = AP and that the measure of the arc OP in ω, that does not contain
A, is 40◦ . Determine the measure of the angle ∠OBC.

Solution. Note that BO = AO = AP, so 4AOP is isosceles. Thus ∠OAP = 20◦ , implying
∠AOC = 140◦ , and ∠AOP = ∠AP O = 80◦ , implying that ∠AOB = 100◦ . So ∠OBC =
360◦ − 140◦ − 100◦ = 120◦ , or ∠OBC = 30◦ .

B P C

§13.2 Length Chasing

Often while solving problems we want to find the lengths of unknown line segments in triangles.
We do this using a variety of methods.

Theorem 13.34 (Angle Bisector Theorem). In triangle ABC, let AD be either angle bisector
of ∠BAC where D lies on BC. Then we have the following proportion:
c BD
= .
b CD

6
Everaise Academy (2020) Math Competitions I

Proof. We prove the theorem for the internal angle bisector only since the external bisector is
very similar. First, let’s create some similar triangles. Draw a line through B parallel to AD
and let its intersection with line AC be E. Then, by AA similarity, we have that EBC is similar
to ADC. Because these triangles are similar, we know the following:
a EC
= .
DC b
Note that a = BD + DC, and EC = EA + b. Then we can rewrite the proportion as the
following:
BD EA
1+ =1+ .
DC b
This is starting to look familiar, but... How do we show that EA = c? Let’s find some angles.
By parallel lines, we have that ∠DAC = ∠BEC. Further, we have that ∠BAD = ∠EBA. Thus
since AD is the angle bisector of ∠BAC, ∠AEB = ∠ABE so EA = c. Thus when we substitute
it in,
BD c
= .
CD b


ac
Remark. Notice that with a little further manipulation, we find that BD = b+c and
ab
CD = b+c . This is very useful too!

The Angle Bisector Theorem is a simple yet instrumental tool in solving geometry problems.
Let’s try a few examples.

Example (2016 AMC 12A). In 4ABC, AB = 6, BC = 7, and CA = 8. Point D lies on

BC, and AD bisects ∠BAC. Point E lies on AC, and BE bisects ∠ABC. The bisectors
intersect at F . What is the ratio AF : F D?

Solution. By the Angle Bisector Theorem applied to 4ABD, the desired ratio is equal to the
ratio of AB to BD. This is a very common trick! Any time you see an angle bisector and a
length ratio, the Angle Bisector Theorem should immediately jump out at you.
The application of the theorem will rarely be on the main triangle, so you should always check
all the segments the angle bisector intersects before deciding whether or not you should switch
ac 7·6
to another length chasing method. To finish the problem, we know that BD = b+c = 8+6 , so
AB 6
F D = 7·6 = 2, giving us the answer 2 : 1. 
8+6

Example (2011 AIME). In triangle ABC, AB = 20 11 AC. The angle bisector of ∠A intersects
BC at point D, and point M is the midpoint of AD. Let P be the point of the intersection
of AC and BM . Find the ratio of CP to P A.

Solution. Okay, there’s an angle bisector, so we know that by the Angle Bisector theorem
AB BD 20
= = .
AC DC 11
We can apply this ratio in other ways by applying similar triangles – draw a line through D
parallel to BP so that it intersects AC at X. Then DCX is similar to BCP through AA
similarity.

7
Everaise Academy (2020) Math Competitions I

BC PC
This means that DC = XC , which in truth means that

BC BD 20 PC PX 31
=1+ =1+ = =1+ = .
DC DC 11 PX XC 11
Further, through AA similarity again, we have that AM P is similar to ADX. Thus, since M
is the midpoint of AD, P must be the midpoint of AX. Thus AP = P X, so the desired ratio
PC PC
P A = P X , which is equivalent to

1 1 31
XC
= 11 = 20 .
1 − PC 1 − 31

Exercise 13.35. Let ABC be a right triangle with ∠ACB = 30 and ∠ABC = 90. Let the angle
bisector of ∠BAC intersect BC at D. What is the length of DC, if AC = 4?
One of the most important formulas to come up in computational geometry may be Stewart’s
theorem.

Theorem 13.36 (Stewart’s Theorem). In triangle ABC, let D lie on BC such that AD
is a cevian with length d, dividing BC into line segments BD = m, CD = n. Then the
following holds: b2 m + c2 n = a(d2 + mn).

Remark. Stewart’s theorem is commonly remembered mnemonically in the form

man + dad = bmb + cnc,

which corresponds to “a man and his dad put a bomb in the sink.”

As you can imagine, this definitely comes in handy while lengths chasing in problems.

Example (2013 AMC 10A). In 4ABC, AB = 86, and AC = 97. A circle with center A and
radius AB intersects BC at points B and X. Moreover BX and CX have integer lengths.
What is BC?

Solution. Since BX and CX have integer lengths, it may be possible to use Stewart’s theorem
on this problem. Note that AX = AB = 86, the radius of the circle, and AC = 97. Then let
BX = b and CX = c. Then by Stewart’s theorem,

bc(b + c) + 862 · (b + c) = 972 · b + 862 · c.

When we simplify and combine like terms, we get the following:

b2 c + c2 b = (972 − 862 )b.

Since b is nonzero, we can divide both sides by b to get

bc + c2 = c(b + c) = 972 − 862 = (97 + 86)(97 − 86) = 11 · 61 · 3.

By the triangle inequality, BC < AC + AB, BC < 86 + 97, so BC < 183. Since c < b + c, c
must equal 33 and b + c = 61. Thus BC = 61, as desired. 

8
Everaise Academy (2020) Math Competitions I

Exercise 13.37. In triangle 4ABC, let M be the midpoint of BC. Prove that
r
2c2 + 2b2 − a2
AM = .
4

Exercise 13.38. In triangle 4ABC, let D be the intersection of the angle bisector of ∠BAC
with BC. Prove that s
(a + b + c)(b + c − a)bc
AD =
(b + c)2
ac ab
Hint: Use BD = b+c
and CD = b+c
.

Finally, we have a lesser-used, but still important, result:

Theorem 13.39 (Ceva’s Theorem). Given a triangle 4ABC, let points D, E, F be on lines
BC, CA, AB, respectively. If we further require that either all three points lie inside the
triangle or two points lie outside the triangle and one lies inside, then AD, BE, CF concur
at a single point if and only if
AF BD CE
· · = 1.
F B DC EA

E
F
P

B D C

Exercise 13.40 (AoPS). Suppose AB, AC, and BC have lengths 13, 14, and 15, respectively. If
AF 2 CE 5
F B = 5 and EA = 8 , find BD and DC.

Exercise 13.41. Show that the medians, angle bisectors, and altitudes of a triangle all concur
at three respective points (the centroid, incenter, and orthocenter) using Ceva.
Exercise 13.42. Let the incircle of 4ABC touch BC, CA, and AB at D, E, and F, respectively.
Prove that AD, BE, CF concur at a point known as the Gergonne Point of 4ABC.
The other two cases for where the points D, E, F lie with respect to the triangle 4ABC are
covered by the following analogous theorem.

Theorem 13.43 (Menelaus’s Theorem). Given a triangle 4ABC, let D, E, F be on lines

BC, CA, AB, respectively. If we further require that all three points lie outside the triangle
or two points lie inside the triangle and one lies outside, then D, E, F are collinear if and
only if
AF BD CE
· · = 1.
F B DC EA

9
Everaise Academy (2020) Math Competitions I

E
A
F

B C
D P

§13.3 Mass Points

Mass points involves assigning weights or a “mass” to points using the ratio of lengths relating
the vertices. If we imagine the line segments as having a certain ratio, then the ratio between
the weights will be inversely proportional.
Consider a seesaw along line segment AB such that the fulcrum is at point C where AC = 3,
BC = 7. Then, since the ratio of AC to BC is 3 to 7, we can assign the shorter line segment
with weight 7 at point A and the longer line segment with weight 3 at point B. Hence, points
A and B have a “mass” that represents the ratio AC : BC. Further, since C is the fulcrum, or
balancing point, C has a mass of the other two points: 3 + 7 = 10.
Here’s the rigorous definition of mass points.

Theorem 13.44 (Mass Points). Consider segment XY with P on XY. Then assign masses
Y
X, Y to points X, Y such that XP
Y P = X .

X P Y

AP
Now consider cevians AD, BE, CF of 4ABC that concur at some point P. Then PD =
B+C
A .
This means that for P on XY, we can define P = X + Y.

E
F
P

B D C

From here we can proceed to derive many properties. This technique can prove to be ex-
tremely useful with geometry problems involving the intersection of several line segments.
As a simple example, we’ll present the centroid.

Example (Centroid). Assign masses to 4ABC, its midpoints, and its centroid.

Solution. Note A = B = C. Without loss of generality, let A = 1.

Then note that since X + Y = P for P on segment XY, D = B + C = 2. Similarly,
E = F = 2, and G = A + D = 1 + 2 = 3.

10
Everaise Academy (2020) Math Competitions I

2 2
3

1 2 1

Example (2019 AMC 8). In triangle ABC, point D divides side AC so that AD : DC =
1 : 2. Let E be the midpoint of BD and let F be the point of intersection of line BC and
line AE. Given that the area of 4ABC is 360, what is the area of 4EBF ?

B F C

Solution. Since we already know the area of triangle ABC and the ratios of certain lengths, it
makes to express the area of EBF, [EBF ] in terms of [ABC]. First, note that EBF and ABF
share the same base, BF. Thus
EF
[EBF ] = [ABF ].
AF
Further, since ABF and ABC share the same altitude,
BF
[ABF ] = [ABC].
BC
Thus, substituting in our expression for [ABF ], we have that
EF BF
[EBF ] = · [ABC].
AF BC
This gives us some motivation to proceed with mass points.
Let’s start by considering AC. Since AD : DC = 1 : 2, we can assign A with a mass of 2 and
C with a mass of 1. Then the mass of D is 1 + 2 = 3. Further, since E is the midpoint of BD,
the mass of B is equal to the mass of D, which we already know is 3. Thus, the mass of E is
3 + 3 = 6. Further, since the mass of B is 3 and the mass of C is 1, we know that F has a mass
of 3 + 1 = 4.
Hence, the ratio of BF : BC is 41 . Further, since E has a mass of 6, A has a mass of 2, and
F has a mass of 4, the ratio of EF : AF is 13 . Substituting these values in, we have that

EF BF 1 1
[EBF ] = · [ABC] = · · 360 = 30.
AF BC 3 4


11
Everaise Academy (2020) Math Competitions I

Example (2011 AIME). In triangle ABC, AB = 20 11 AC. The angle bisector of ∠A intersects
BC at point D, and point M is the midpoint of AD. Let P be the point of the intersection
of AC and BM . Find the ratio of CP to P A.

Solution. You may recognize this problem from the previous section. This time, we’ll use mass
points in order to arrive at a faster solution.
Again, we use the Angle Bisector theorem to arrive at the conclusion that
AB BD 20
= = .
AC DC 11
This means that we can assign B a mass of 11 and C a mass of 20, with D having a mass of
11 + 20 = 31. Then, since M is the midpoint of AD, we have that A has the same mass as D,
31.
Thus we know, when considering the line segment AC, that A has a mass of 31 and C has
a mass of 20. Therefore, the ratio of CP to P A is equal to the mass of A over the mass of C,
which is 31
20 . 

This next problem is another instance in which we can apply both the Angle Bisector theorem
and mass points.

Example (2009 AIME I). Triangle ABC has AC = 450 and BC = 300. Points K and L
are located on AC and AB respectively so that AK = CK, and CL is the angle bisector
of angle C. Let P be the point of intersection of BK and CL, and let M be the point on
line BK for which K is the midpoint of P M . If AM = 180, find LP .

Solution. First, by the Angle Bisector theorem, we have that

BL AL
= ,
CB CA
so when we substitute in CB = 300 and CA = 450,
BL AL
= ,
300 450
so BL = 23 AL. Thus, we can use mass points and assign A, B, and L weights of 2, 3, and 5
respectively. Further, since K is the midpoint of AC, C has the same mass, 2, as A. Then, the
mass of P must be the combined masses of C and L, which is 2 + 5 = 7. This means that LP,
our desired length, is equivalent to
2
LP = CP.
5
There’s just one last step: finding the length of CP. Thankfully, since K is the midpoint of
both AC and P M, we know that M KA is congruent to P KC by SAS congruence. This means
that M A = CP = 180, so
2
LP = · 180 = 72.
5


Exercise 13.45 (2016 AMC 12A). In 4ABC, AB = 6, BC = 7, and CA = 8. Point D lies

on BC, and AD bisects ∠BAC. Point E lies on AC, and BE bisects ∠ABC. The bisectors
intersect at F . What is the ratio AF : F D?
Exercise 13.46 (2004 AMC 10B). In 4ABC points D and E lie on BC and AC, respectively.
AT
If AD and BE intersect at T so that DT = 3 and BT CD
ET = 4, what is BD ?

12
Everaise Academy (2020) Math Competitions I

§13.4 Area Formulas Revisited

Area formulas are a simple yet effective way for calculating the area of triangles using the
property of circumscribed and inscribed circles.

Theorem 13.47. For a triangle ABC, [ABC] is equal to r · s where r denotes the radius
of the inscribed circle and s denotes the semiperimeter (half of the perimeter).

Proof. Consider a triangle ABC with a point O that is the center of its inscribed circle. By
the definition of an inscribed circle, the perpendicular segments from AB, AC, and BC will
intersect at point O. Furthermore, the distance from the point to each segment will be equal
to the radius of the inscribed circle.
If we let the radius of the inscribed circle be r, [ABO] is equal to AB · 2r . Similarly, [ACO] is
equal to AC · 2r and [BCO] is equal to BC · 2r . Therefore, the total area of the triangle equals
(AB+BC+AC)
2 · r. 

Example. If there exists an equilateral triangle ABC such that all three sides have a length
of 6, what would be the radius of the inscribed circle? What relation does the radius have
with the area of the triangle?

Solution. We can √ solve this problem using the formula for the area of the equilateral triangle,
which gives us 9 √ 3. The perimeter of the triangle is 18. Plugging this in, the radius of the
inscribed circle is 3. The area of the triangle can be noted as the product of the radius and
the perimeter divided by 2. 

Example (2004 AMC 10). A triangle with sides of 5, 12, and 13 has both an inscribed and
a circumscribed circle. What is the distance between the centers of those circles?

Solution. We can first note that the triangle is a right triangle as (5,12,13) is a pythagorean
triple. Let us denote the triangle as ABC with AB, AC, BC respectively being 5, 13, and 12.
The area of the triangle is equal to 5 · 12
2 = 30. Using the area formula, we can find the radius
30
of the inscribed circle to be 15 = 2.

Let the center of the circumcircle of ABC be point O1 and the center of the inscribed circle
be O2 . Additionally, let us denote the points of intersection between the inscribed circle and
AB, AC, BC to be X, Y, and Z, respectively. We can calculate the length of segment of AX by
using the property of an inscribed circle. If we denote its length as p, the length of BX is 5 − p
and BY as well.
Then, CY and CZ are 12 − (5 − p) = p + 7. Now, we can calculate the length of AZ with
the equation of p + 7 + p = 13, giving us p = 3.

√ of ZO1 is√ 6.5 − 3 = 3.5. Using the Pythagorean theorem on 4O1 ZO2 leads
Lastly, the length
to an answer of 16.25 or 265 . 

Exercise 13.48 (2008 AMC 12). Triangle ABC has AC = 3, BC = 4, and AB = 5. Point D is
on AB, and CD bisects the right angle. The inscribed circles of 4ADC and 4BCD have radii
ra and rb , respectively. What is ra /rb ?

13
Everaise Academy (2020) Math Competitions I

Another simple yet efficient way of calculating the area of a triangle involves the sine function,
where for two sides a, b and an angle ∠C between them, b · sin ∠C denotes the height of the
triangle relative to the side A. Thus, it can be noted that this theorem is a variation of the 21 bh
formula where b is the base and h is the height.

Theorem 13.49. For a triangle ABC, [ABC] is equal to

1
ab · sin C.
2

Example. For a triangle ABC with sides AB,√ BC, AC respectively having lengths of 7, 8,
and 9. The area of the triangle is given to be 720. Find sin ∠A + sin ∠B.

Solution. Using the previous theorem, we can calculate sin ∠A and sin ∠B.
9 · 7 sin ∠A √
= 720
2
√
720
sin ∠A =
63
We can similarly create an expression for sin ∠B as the following:
8 · 7 sin ∠B √
= 720
2
√
720
sin ∠B =
56
Adding the two expressions up, we have our answer of
√
17 720
504


Now we will cover another convenient way of calculating the area of a triangle using the
properties of its circumscribed circle. We can rewrite the previous theorem of [ABC] = ab
2 sin ∠C
abc c
to 4R with sin ∠C = 2R .

Theorem 13.50. For a triangle ABC, [ABC] = A·B·C 4R where A, B, C are three sides of the
triangle and R is the radius of the triangle’s circumscribed circle.

Remark. This theorem is usually paired with the extended law of sines, which we’ll go over
tomorrow. It will expand on this formula.

Example. For a triangle ABC with AB, BC, and AC respectively having lengths of 3, 8
and 5, develop a generic expression for the radius of circumscribed circle of the triangle
using angles ∠A, ∠B, and ∠C.

14
Everaise Academy (2020) Math Competitions I

Solution. Referring back to the previous theorem, we can note that [ABC] = ab 2 sin ∠C where
b represents side AC and a represents BC. Then we can use the Law of Sines to express
b = asin
sin ∠B
∠A (if you don’t fully understand this step, we’ll go over the Law of Sines in detail
tomorrow).
2
Substituting this in gives us [ABC] = a sin2 sin
∠B sin ∠C
∠A . Applying the circumradius formula
yields
a2 sin ∠B sin ∠C
R=
240 sin ∠A


Theorem 13.51 (Heron’s formula). The area of a triangle 4ABC is

p
s(s − a)(s − b)(s − c),

where s denotes the semiperimeter.

Remark. This can be proven by splitting the triangle into right triangles, then bashing
using the Pythagorean theorem.

Example (2010 AMC 10). A triangle has side lengths 10, 10, and 12. A rectangle has width
4 and area equal to the area of the triangle. What is the perimeter of this rectangle?

Solution. We can compute √ the semiperimeter to be 16. Using Heron’s Formula, we find that
the area of the triangle is 16 · 6 · 6 · 4. Since this value equals 48, for the area of the rectangle
to equal that of the triangle, its length must be 12. Thus, the perimeter of the rectangle is
2 ∗ (4 + 12) = 32. 

Exercise 13.52. Find the area of a triangle with side lengths 13, 14, and 15 using Heron’s. What
is the circumradius of this triangle? The inradius?

§13.5 Homework Problems

Problem 13.1 (2019 AMC 10A). [4] Let 4ABC be an isosceles triangle with BC = AC and
∠ACB = 40◦ . Construct the circle with diameter BC, and let D and E be the other intersection
points of the circle with the sides AC and AB, respectively. Let F be the intersection of the
diagonals of the quadrilateral BCDE. What is the degree measure of ∠BF C?

Problem 13.2 (1985 AIME). [4] As shown in the figure, triangle ABC is divided into six
smaller triangles by lines drawn from the vertices through a common interior point. The areas
of four of these triangles are as indicated. Find the area of triangle ABC.

15
Everaise Academy (2020) Math Competitions I

84

35

40 30

A B

Problem 13.3 (2001 AIME). [5] Triangle ABC has AB = 21, AC = 22 and BC = 20. Points
D and E are located on AB and AC, respectively, such that DE is parallel to BC and contains
the center of the inscribed circle of triangle ABC. Then DE = m/n, where m and n are
relatively prime positive integers. Find m + n.

Problem 13.4 (2013 AMC 10B). [5] In triangle ABC, medians AD and CE intersect at P ,
P E = 1.5, P D = 2, and DE = 2.5. What is the area of AEDC?

Problem 13.5 (2016 AMC 12A). [5] Let ABCD be a square. Let E, F, G and H be the
centers, respectively, of equilateral triangles with bases AB, BC, CD, and DA, each exterior to
the square.
√ We can express the ratio of the area of square EF GH to the area of square ABCD
as a+c b for positive integers a, b, and c where b is squarefree. What is a + b + c?

Problem 13.6 (2003 AIME II). [6]

Find the area of rhombus ABCD given that the radii of the circles circumscribed around
triangles ABD and ACD are 12.5 and 25, respectively.

Problem 13.7 (2016 AMC 10A). [6] In rectangle ABCD, AB = 6 and BC = 3. Point E
between B and C, and point F between E and C are such that BE = EF = F C. Segments
AE and AF intersect BD at P and Q, respectively. The ratio BP : P Q : QD can be written
as r : s : t where the greatest common factor of r, s and t is 1. What is r + s + t?

Problem 13.8 (2016 AMC 10B). [7] Rectangle ABCD has AB = 5 and BC = 4. Point E lies
on AB so that EB = 1, point G lies on BC so that CG = 1. and point F lies on CD so that
DF = 2. Segments AG and AC intersect EF at Q and P , respectively. We can write the value
of PEF
Q
as a fraction pq in lowest terms. What is p + q?

Problem 13.9 (2016 AIME I). [7] In 4ABC let I be the center of the inscribed circle, and let
the bisector of ∠ACB intersect AB at L. The line through C and L intersects the circumscribed
circle of 4ABC at the two points C and D. If LI = 2 and LD = 3, then IC = pq , where p and
q are relatively prime positive integers. Find p + q.
Provide a detailed solution, or explain what progress you made, even if you
didn’t get an answer.

16
14 Analytic Geometry + Trigonometric
Techniques

§14.1 Coordinate Bashing

You are probably familiar with the Cartesian coordinate plane and its algebraic applications.
You know the formulas for line equations, circle equations, conditions for perpendicularity,
conditions for parallel lines, etc. Usually, we see these objects in the context of algebra, but
seem closely related to many geometrical concepts.
This naturally leads to the question, “Can we use our algebraic techniques in the Cartesian
coordinate plane to solve geometry problems?” The answer is in fact yes, and this is the idea
behind coordinate bashing, or turning a geometry problem into an algebraic one.
We first present three formulas that will help greatly when coordinate bashing:

Proposition 14.1 (Distance Formula). The distance between two points (x1 , y1 ) and (x2 , y2 )
is q
(x1 − x2 )2 + (y1 − y2 )2 .

Exercise 14.2. Prove this!

Proposition 14.3 (Point to Line Distance Formula). The distance between point (x1 , y1 ) to
line ax + by + c = 0 is
|ax1 + by1 + c|
√ .
a2 + b2

Proposition 14.4 (Shoelace Formula). Suppose an n-sided convex polygon has vertices at
(x1 , y1 ), (x2 , y2 ), · · · , (xn , yn ), where the vertices are in counterclockwise or clockwise order.
Its area is given by
1
|(x1 y2 + x2 y3 + · · · + xn y1 ) − (x2 y1 + x3 y2 + · · · x1 yn )|.
2

The following diagram shows why it’s called the Shoelace Formula. Take half of the absolute
value of the products joined by the blue lines and subtract the products joined by the red lines
to get the area, as shown in the diagram.

1
Everaise Academy (2020) Math Competitions I

x1 y1

x2 y2

x3 y3
..
.
xn−1 yn−1

xn yn

Let’s see it in action.

Example 14.5 (2017 AMC 10). Let ABC be an equilateral triangle. Extend side AB
beyond B to a point B 0 so that BB 0 = 3AB. Similarly, extend side BC beyond C to
a point C 0 so that CC 0 = 3BC, and extend side CA beyond A to a point A0 so that
AA0 = 3CA. What is the ratio of the area of 4A0 B 0 C 0 to the area of 4ABC?

Solution. Since this problem is asking for ratios, let’s just assume that 4ABC has side length
of 1. First, we need to find A, B, and C on the coordinate plane. To make life easier, we like
to get as many points as possible on the coordinate axes, since at least one of the coordinates
will become 0 (we’ll see how this reduces calculations). Set one of these points at the origin,
say, A. We can then get another point on the coordinate axis: √
let B be at (1, 0). Finally, since
3
ABC is an equilateral triangle, C will be at a distance of 2 away  from the midpoint of AB.
√ 
0+1 0+0 1 1 3



The midpoint of AB is given by 2 , 2 , or 2 , 0 , so C is at 2 , 2 ,

Now how do we have A0 , B 0 , and C 0 ? Well, for each segment, we can phrase √
it in terms of “rise
and run.” For example, if we consider AC, we can consider it a “rise” of 23 and a “run” of
1
2 . However, if we consider CA, which goes in the opposite direction, we consider it a “rise” of
√
3
− 2 and a “run” of − 12 . Thus, if we extend AC past A to A0 so that AA0 = 3CA, we can think
√
of it as extending CA past A for 3 “rise and runs”, or changing the y-coordinate of A by − 23
three times and changing the x coordinate of A by − 12 three times. Thus, our coordinates for
A0 are   √ !! √ !
1 3 3 3 3
0+3 − ,0 + 3 − = − ,−
2 2 2 2

2
Everaise Academy (2020) Math Competitions I

Applying similar ideas to find B 0 and C 0 ,

B 0 : (1 + 3(1), 0 + 3(0)) = (4, 0)

  √ √ !! 
1 1 3 3 √ 
C0 : +3 − , +3 = −1, 2 3
2 2 2 2

Remark. The concept of CA and AC being different due to direction is known as directed
lengths.

All we need to do now is to find the area of A0 B 0 C 0 . This is where we see one of the benefits of
coordinate bashing: finding areas is extremely easy once you know the points of the shape. To
Shoelace, set your points up as follows:

(4, 0)
√
3 3 3
(− , − )
2 √2
(−1, 2 3)
(4, 0)

Think about why we chose to start with (4, 0). Because we’re multiplying a lot, we want as
many zeroes as possible to reduce calcuations. Because we put the first point again at the end,
we will ideally want the first point to have at least one coordinate be 0. Applying the formula,
we get our area to be
√ !   ! √ ! !
√  √
 
1 3 3 3 3 3 3
| (4) − + − 2 3 + (−1)(0) − (0) − + − (−1) + (2 3)(4) |
2 2 2 2 2
√
1 √ √ 3 3 √
= |−6 3−3 3− − 8 3|
2 2
√ ! √
1 37 3 37 3
= =
2 2 4
√ √
So our area for A0 B 0 C 0 is 374 3 . Since the area of ABC is 3
4 , the ratio of the areas of A0 B 0 C 0
to ABC is simply 37 : 1 . 

Some takeaways: the “rise and run” idea is extremely important, and this will come up over
and over again (this is especially true for problems involving reflections). Also, staying organized
when coordinate bashing is extremely important! One mistake with a sign or calculation can
be your downfall.
Now, let’s look at another advantage that coordinate bashing provides:

5
Example 14.6. (2011 AIME I) Let L be the line with slope 12 that contains the point
A = (24, −1), and let M be the line perpendicular to line L that contains the point
B = (5, 6). The original coordinate axes are erased, and line L is made the x-axis, and line
M the y-axis. In the new coordinate system, point A is on the positive x-axis, and point
B is on the positive y-axis. The point P with coordinates (−14, 27) in the original system
has coordinates (α, β) in the new coordinate system. Find α + β.

3
Everaise Academy (2020) Math Competitions I

Solution. How can we interpret these new coordinates? That is, what does (1, 2) mean in the
Cartesian coordinate plane? We know it’s the point that is 1 to the right of the y axis, and 2
above the x axis, i.e. it is has (signed) distance 1 from the y axis and a distance of 2 from the
x axis.
So what is (α, β) really asking for? It’s asking for the distance from the new rotated x and
y axes! Of course, we need to remember positive and negative is significant, but that won’t be
too difficult as we just need to check if the point is above/below the new axes.
This is where coordinate bashing provides a significant advantage, because we literally have
a formula for finding the distance between points and lines. So let’s compute this!

First, we need to find our equations for our lines in standard form (ax + by + c = 0). Using
point-slope,
5
L : (y + 1) = (x − 24) ⇐⇒ 5x − 12y − 132 = 0
12
12
M : (y − 6) = − (x − 5) ⇐⇒ 12x + 5y − 90 = 0
5
Using our point line distance formula, we can find the distance between P and each of L and
M ! Let the distance between point X and line L be denoted as d(X, L):

|5(−14) − 12(27) − 132| 526

d(P, L) = √ =
52 + 122 13

|12(−14) + 5(27) − 90| 123

d(P, M ) = √ =
2
5 + 12 2 13

So we now know that P is of the form ± 123 526

13 , ± 13 and it suffices to figure out the signs.
Remember that the distance from the x axis is the absolute value of the y-coordinate, and vice
versa! Don’t mix these up!
Let’s find the x coordinate first: if A is on the positive x axis, then anything on the same
side of the y axis, M , will have a positive x coordinate. To check which side A is on compared
to M , we can easily see that 12(24) + 5(−1) − 90 > 0. Thus, if we get a positive value when we
plug in P into the equation, then the x coordinate will be positive, and vice versa:

12(−14) + 5(27) − 90 = −168 + 135 − 90 < 0

Since our value is negative, P lies on the opposite side of M as A, meaning the x coordinate is
negative.

4
Everaise Academy (2020) Math Competitions I

We follow a similar process to find the sign of the y coordinate. First, plug in B into the L
to get that 5(5) − 12(6) − 132 < 0. Thus, if we plug in P and get a negative value, P will lie on
the same side of L as B, meaning the y coordinate is positive, and vice versa:

5(14) − 12(27) − 132 < 0

This means our y coordinate is positive, so our point P is − 123 526

13 , 13 , meaning our answer is
− 123 526 403
13 + 13 = 13 = 31 . 

Remark. For those of you preparing for the AIME, this problem appeared very early and
is very difficult for
its position. If you are running out of time and you have gotten to
P = ± 123 13 , ± 526
13 , you can take advantage of the fact that answers for the AIME are
positive integers, so you just need to check 526+123
13 and 526−123
13 , of which only the latter
gives an integer.

Now let’s finish with one more example, which shows an application of using coordinate
bashing to find points and solve equations, rather than straight computation.

Example 14.7. (2018 AMC 10A) Farmer Pythagoras has a field in the shape of a right
triangle. The right triangle’s legs have lengths 3 and 4 units. In the corner where those
sides meet at a right angle, he leaves a small unplanted square S so that from the air it
looks like the right angle symbol. The rest of the field is planted. The shortest distance
from S to the hypotenuse is 2 units. What fraction of the field is planted?

Solution. You’ve seen this example before, but this time we’ll do it with coordinates.
In order to find the fraction of the field that is planted, we need to find the part that is
unplanted. Since this is just a square, all we need is the side length, s. How can we set this up
with coordinates?
We want to take advantage of the fact that its a right triangle: because the x and y axes are
perpendicular, its natural to place the right triangle in the plane such that its vertex is at the
origin (0, 0) and the other two points are on the axes (so that the legs coincide with each axis).
Let these points be (0, 3) and (4, 0). Remember, we want to get as many zeroes as possible!

4
y
3.5
B
3

2.5

1.5

1
(s, s)
0.5
A C x
−0.5 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5
−0.5

5
Everaise Academy (2020) Math Competitions I

So how could we find the side length of the square? Let’s use the piece of information we haven’t
used yet: the shortest distance from the square S to the hypotenuse is 2 units. This means that
the point diagonally opposite the vertex will be a distance 2 away from the hypotenuse.
Aha! If we can represent this point in terms of variables, we can just plug it into the distance
formula to solve for the point! This is where the square comes into play: let the side length
of the square be s. The point diagonally opposite the vertex then is simply (s, s) (convince
yourself of this!).
−3
To use the distance formula, we need our line equation. The slope is just 4 , and we know
the y-intercept is (0, 3), so our equation is
3
y = − x + 3 ⇐⇒ 3x + 4y − 12 = 0
4
Plugging in (s, s) into our distance formula and setting the value equal to 2,
|3s + 4s − 12| 22 2
√ = 2 =⇒ |7s − 12| = 10 =⇒ s = ,
2
3 +4 2 7 7
Which value of s do we use? Well, 22 7 > 3, which means that part of the square will actually lie
outside of the triangle, which is not right (essentially, it’s too big). Thus, s = 27 , or the area of

2
our triangle is 27 = 49 4
. Since the area of our triangle is 3·4
2 = 6, the fraction of the square that
4 2 2 145
is unplanted is 49·6 = 147 , so the fraction of the square that is planted is 1 − 147 = . 
147

Hopefully these examples give you a good sense of the power coordinate bashing provides.
However, use this technique sparingly! It can get computationally intense, and is prone to
mistakes. In general, use this technique only if there are lots of perpendicular lines/right angles
or triangles (points of quadrilaterals can get difficult to find), and if the question asks to find
distances or areas.
If you are familiar with equations of circles, feel free to use this as well! Although we won’t
cover it in this handout, we will list some things to keep in mind with circles:
• Generally, try to keep it to one circle: you should never go over two.
• Tangency to lines is always nice, because the radius will be equal to the distance from
the point to the respective line, so using one of the axes as the line will give you a free
coordinate
• Remember that if you intersect a line with a circle, you will (usually) have two intersection
points: never forget about the other!
• If you decide to use two circles and have a problem dealing with the line through the
two intersections, you can actually find the equation of the line by subtracting the circle
equations (although this isn’t common, it’s extremely useful when it does show up)
Exercise 14.8 (2015 AMC 10A). The isosceles right triangle ABC has right angle at C and area
12.5. The rays trisecting ∠ACB intersect AB at D and E. What is the area of 4CDE?
Exercise 14.9 (2019 AMC 10B). Rectangle ABCD has AB = 5 and BC = 4. Point E lies on
AB so that EB = 1, point G lies on BC so that CG = 1, and point F lies on CD so that
DF = 2. Segments AG and AC intersect EF at Q and P , respectively. What is the value of
PQ
EF ?

§14.2 Law of Sines

6
Everaise Academy (2020) Math Competitions I

Theorem 14.10. In triangle 4ABC with circumradius R, we have

a b c
= = = 2R.
sin ∠CAB sin ∠ABC sin ∠BCA

Proof. Let M be the midpoint of BC and let O be the circumcenter of 4ABC. We will just
a b c
show that sin ∠CAB = 2R, since we can prove that sin ∠ABC = 2R and sin ∠BCA = 2R in the
exact same way. If ∠BAC is right, the circumcenter is the midpoint of the hypotenuse, so
2R = a and cos ∠CAB = cos π2 = 1, which immediately implies sin ∠CAB
a
= 2R. If ∠BAC is
acute, by the Inscribed Angle Theorem, we must have ∠BOC = 2∠CAB, so by definition,
a
BM
sin ∠CAB = sin ∠BOM = = 2,
BO R
a
which becomes sin ∠CAB = 2R as desired. In the case where ∠CAB is obtuse, we instead have
∠BOC = 2π − 2∠CAB, so sin ∠BOM = sin(π − ∠CAB) = sin ∠CAB, which gives the same
result.
A

A0
O

B C

Example 14.11 (2005 AIME II/14). In triangle ABC, AB = 13, BC = 15, and CA = 14.
Point D is on BC with CD = 6. Point E is on BC such that ∠BAE ∼ = ∠CAD. Given that
p
BE = q where p and q are relatively prime positive integers, find q.

Solution. This is very straightforward. We find that

AB sin ∠BAE
BE = (Law of Sines on 4ABE)
sin ∠AEB
AB sin ∠BAE
= (∠AEB = π − ∠AEC)
sin ∠AEC
AB · CE sin ∠BAE
= (Law of Sines on 4AEC)
AC sin ∠CAE
AB · CE sin ∠BAE
= (∠CAE = ∠BAD)
AC sin ∠BAD
AB 2 · CE sin ∠BAE
= (Law of Sines on 4BAD)
AC · BD sin ∠BDA
AB 2 · CE sin ∠BAE
= (∠BDA = π − ∠CDA)
AC · BD sin ∠CDA
AB 2 · CE · CD sin ∠BAE
= (Law of Sines on 4CAD)
AC 2 · BD sin ∠CAD
2
13 · (15 − BE) · 6
= ,
142 · 9
132 ·15
so BE = 463 =⇒ q = 463 . 

7
Everaise Academy (2020) Math Competitions I

We pretty much solved this problem just by continuously substituting sines with either the
Law of Sines or the given angle equalities, with little geometric or even algebraic cleverness
involved. The main reason why this worked was because we already had most of lengths in
the problem and thus substituting sines didn’t result in new unknowns, but usually this is not
the case. In the next example, we will have to construct an auxiliary point, though it is rather
easily motivated.

Example 14.12 (2019 PUMAC). Let 4ABC be a triangle with circumcenter O and ortho-
center H. Let D be a point on the circumcircle of ABC such that AD ⊥ BC. Suppose
area(4ABC)
that AB = 6, DB = 2, and the ratio area(4HBC) = 5. Then, if OA is the length of the
circumradius, then OA can be written in the form m
2
n , where m, n are relatively prime
positive integers. Compute m + n.

Solution. Let AO = R as usual. Since O isn’t explicitly connected to anything in the problem,
it’s probably a good idea to use the Law of Sines to find R. By cyclic quadrilaterals, ∠BCA =
∠BDA, so
6 6
2R = = ,
sin ∠BCA sin ∠BDA
and thus it suffices to find sin ∠BDA. Well, we know that AD ⊥ BC, so we are motivated to let
D0 be the foot of the altitude from A to BC. By the given area ratio and orthocenter reflections,
we can set HD0 = DD0 = x and AH = 4x. Now we just want x, which we can find with the
Pythagorean Theorem:
r r
4 8
62 − (5x)2 = BD02 = 22 − x2 =⇒ x = , BD0 = .
3 3
We’ve found all the lengths we need, and at this point the problem is just computation:
 2 !2
 2  2 r
2 6 3BD 6 27
R = = = q  = ,
2 sin ∠BDA BD0 8 2
3

so 29 is our desired answer. 

Although the symmetric sine equality is certainly pretty useful as we saw in our first example,
the Law of Sines really shines when we want to find the circumradius; any time you see the
circumradius even mentioned, you should immediately consider the Law of Sines. However,
the Law of Sines obviously can’t solve every circumradius problem (efficiently at least), so try
not to go down a trig bash rabbit hole. Using the Law of Sines, we can also prove a generalization
of the Angle Bisector Theorem.

Lemma 14.13 (Ratio Lemma). In 4ABC, let D be a point on A. Then

BD c sin ∠CAD
= · .
CD b sin ∠BAD

Proof. By the Law of Sines on 4ABD, 4ACD, we know that

BD c CD b
= , = ,
sin ∠BAD sin ∠ADB sin ∠CAD sin ∠ADC

8
Everaise Academy (2020) Math Competitions I

respectively. Dividing the first equation by the second, we get

BD sin ∠BAD c sin ∠ADB
· = ·
CD sin ∠CAD b sin ∠ADC
BD c sin ∠ADB sin ∠CAD
= · ·
CD b sin(π − ∠ADB) sin ∠BAD
BD c sin ∠CAD
= ·
CD b sin ∠BAD

as desired.
A

B D C

§14.3 Law of Cosines

Theorem 14.14 (Law of Cosines). In triangle ABC, we have that

c2 = a2 + b2 − 2ab cos ∠ACB.

Proof. Let the foot of the altitude from A to BC be H. Then note that A = b sin C, CH =
b cos C, and BH = |a−b cos C|. (The absolute value is because ∠B can either be acute or obtuse.)
Then note by the Pythagorean Theorem, (b sin C)2 + (a − b cos C)2 = a2 + b2 − 2ab cos C = c2 .
A

C H B

We can use the Law of Cosines to prove a very handy formula from the previous handout!

Example. Prove Stewart’s theorem using Law of Cosines.

Solution. We start off with triangle ABC such that D lies on BC with AD a cevian with
length d, dividing BC into line segments BD = m, CD = n. Then we can apply Law of Cosines
to both ACD and ABD. In ACD, AC 2 = DC 2 + AD2 − 2 · AD · DC cos ADC corresponds to

b2 = d2 + n2 − 2dn cos ADC.

9
Everaise Academy (2020) Math Competitions I

Similarly, we apply the same method to ABD to obtain

c2 = d2 + m2 − 2dm cos ADB.
Note that cos ADB = − cos ADC because ∠ADB and ∠ADC are supplementary. Then, solving
for cos ADC in two ways, we have that
d2 + n2 − b2 c2 − d2 − m2
cos ADC = = .
2dn 2dm
Once we cross multiply and simplify a bit, we have that
c2 n + b2 m = m2 n + n2 m + d2 m + d2 n.
Factoring the right hand side of the equation yields mn(m+n)+d2 (m+n), so we can substitute
a = m + n to finally obtain
c2 n + b2 m = a(mn + d2 ),
as desired.
A

C D B

Now let’s apply Law of Cosines to some competition questions.

Example (2012 AMC 12A). Circle C1 has its center O lying on circle C2 . The two circles
meet at X and Y . Point Z in the exterior of C1 lies on circle C2 and XZ = 13, OZ = 11,
and Y Z = 7. What is the radius of circle C1 ?

Solution. We want to find r, the radius of C1 . Note that XO = OY = r. We could either apply
Law of Cosines on Y ZO or OZX, except for one issue: how do we find the cosine of ∠Y ZO or
∠OZX, respectively?
One of the key things to note with this problem is that XOY Z is a cyclic quadrilateral. It’s
important to note that ∠Y ZO = ∠Y XO, and we can find cos Y XO due to XOY being isoceles
– all we need to obtain is the ratio of
1
2 XY
.
r
Thus, we can try to use Ptolemy’s theorem to help find our desired ratios. By Ptolemy’s theorem
applied to XOY Z,
11 · XY = 7r + 13r = 20r,
20
so XY = 11 r. Thus,
1
2 XY 10
cos Y ZO = cos Y XO = = .
r 11
Success! Now we can apply Law of Cosines to obtain our desired length:
10
r2 = 112 + 72 − 2 · 11 · 7 · = 30,
11
√
so r = 30. 

10
Everaise Academy (2020) Math Competitions I

Example (2012 AIME I). Let 4ABC be a right triangle with right angle at C. Let D and E
be points on AB with D between A and E such that CD and CE trisect ∠C. If DE 8
BE = 15 ,
then find tan B.

AC
Solution. First, let’s note that tan B = BC . Since this problem is concerned with the ratio of
lengths and not lengths-chasing, without loss of generality we can set BC = 1. Since CD and
CE trisect ∠C, we also have angle bisectors – which is important because we can then apply the
Angle Bisector Theorem. Another great thing about the trisectors of this right angle is that it
splits ∠C into angles that we can easily calculate the cosine of – 90, 60, 30, etc. Since DE 8
BE = 15 ,

8 8
DC = · BC = .
15 15
Now we can apply Law of Cosines to BDC :
 2
2 8 8 169
BD = +1−2· · 1 · cos 60 = .
15 15 225
13
Thus BD = 15 . Now we can obtain the cosine of ∠B, and from there, use that value to obtain
sin B
the sine of ∠B. Since tan B = cos B , we can then get our desired answer. We apply Law of
Cosines again to BDC :
64 169 13
= 12 + −2·1· · cos B.
225 225 15
Solving for
1 + 169 64
225 − 225
cos B = 26
15
11
yields us 13 . Then we can solve for
r √
p 121 4 3
sin B = 1 − cos2 B = 1− = .
169 13
Finally, √
sin B 4 3
tan B = = .
cos B 11


Exercise 14.15. Let ABCDEF be a regular hexagon with side length 4. Let M be the midpoint
of side AB. Then find the length of M C.
Exercise 14.16 (2006 AMC 12B). Isosceles 4ABC has a right angle at C. Point P is inside
4ABC,√such that P A = 11, P B = 7, and P C = 6. Legs AC and BC have length s =
p
a + b 2, where a and b are positive integers. What is a + b?

11
Everaise Academy (2020) Math Competitions I

§14.4 Homework Problems

Problem 14.1 (2015 PUMaC B Division). [3] Find the distance CF in the diagram below
where ABDE is a square and angles and lengths are as given:
E D
8
F
15

15
C
8
A B
√
The length CF is of the form a b for integers a, b such that no integer square greater than 1
divides b. What is a + b?

Problem 14.2 (2001 AIME I). [4] In triangle ABC, angles A and B measure 60 degrees and
45 degrees, respectively. The bisector of angle A intersects BC at T , and AT = 24. The area of
√
triangle ABC can be written in the form a + b c, where a, b, and c are positive integers, and
c is not divisible by the square of any prime. Find a + b + c.

Problem 14.3 (2020 AMC 10A). [5] Quadrilateral ABCD satisfies ∠ABC = ∠ACD =
90◦ , AC = 20, and CD = 30. Diagonals AC and BD intersect at point E, and AE = 5.
What is the area of quadrilateral ABCD?

Problem 14.4 (1971 AHSME). [5] Quadrilateral ABCD is inscribed in a circle with side AD,
a diameter of length 4. If sides AB and BC each have length 1, then side CD has length pq ,
where p and q are positive coprime integers. What is the value of p + q?

Problem 14.5 (2017 AIME II). [7] Rectangle ABCD has side lengths AB = 84 and AD = 42.
Point M is the midpoint of AD, point N is the trisection point of AB closer to A, and point O
is the intersection of CM and DN . Point P lies on the quadrilateral BCON , and BP bisects
the area of BCON . Find the area of 4CDP .

Problem 14.6 (2010 AIME II). [7] Let ABCDEF be a regular hexagon. Let G, H, I, J, K,
and L be the midpoints of sides AB, BC, CD, DE, EF , and AF , respectively. The segments
AH, BI, CJ, DK, EL, and F G bound a smaller regular hexagon. Let the ratio of the area of
the smaller hexagon to the area of ABCDEF be expressed as a fraction m
n where m and n are
relatively prime positive integers. Find m + n.

Problem 14.7 (2013 AIME II). [7] In 4ABC, AC = BC, and√point D is on BC so that
CD = 3 · BD. Let E be the midpoint of AD. Given that CE = 7 and BE = 3, the area of
√
4ABC can be expressed in the form m n, where m and n are positive integers and n is not
divisible by the square of any prime. Find m + n.

Problem 14.8 (2019 AIME I). [10] Triangle ABC has side lengths AB = 4, BC = 5, and
CA = 6. Points D and E are on ray AB with AB < AD < AE. The point F 6= C is a point of
intersection of the circumcircles
√ of 4ACD and 4EBC satisfying DF = 2 and EF = 7. Then
BE can be expressed as a+bd c , where a, b, c, and d are positive integers such that a and d are
relatively prime, and c is not divisible by the square of any prime. Find a + b + c + d.

12
Everaise Academy (2020) Math Competitions I

Problem 14.9 (2018 AIME II). [7] Octagon ABCDEF GH with side lengths AB = CD =
EF = GH = 10 and BC = DE = F G = HA = 11 is formed by removing four 6 − 8 − 10
triangles from the corners of a 23 × 27 rectangle with side AH on a short side of the rectangle,
as shown. Let J be the midpoint of HA, and partition the octagon into 7 triangles by drawing
segments JB, JC, JD, JE, JF , and JG. Find the area of the convex polygon whose vertices
are the centroids of these 7 triangles.

G F

H E

A D

B C

Provide a detailed solution, or explain what progress you made, even if you
didn’t get an answer.
x1 +x2 +x3 y1 +y2 +y3


Hint: The centroid of the triangle with vertices (x1 , y1 ), (x2 , y2 ), (x3 , y3 ) is 3
, 3
.

13
15 3D Geometry
Until now, we’ve mostly only looked at geometry in the Euclidean plane. Today, we’ll explore
geometry in three dimensions, which requires more visual intuition.

§15.1 3D Shapes
There are many shapes that can be drawn in 3D space. The ones that we’ll focus on, as well as
the most frequently appearing ones on contests, are planes, prisms, polyhedra, spheres, cones,
and cylinders.

§15.1.1 3D Coordinates

Before we begin, we first introduce a coordinate system for 3D shapes – which is quite similar
to our 2D coordinate system.

Definition 15.1. 3D coordinate space is the set of all points (x, y, z) for real numbers
x, y, z. The number x represents the displacement from the yz-plane, while y represents the
displacement from the xz-plane and z is the displacement from the xy-plane.

You can think of this as taking a point in the familiar xy-plane then moving it so it levitates
above or below with a certain displacement z away from the plane.

§15.1.2 Planes

Definition 15.2. A plane is a flat, 2D surface in 3D space.

···
···

Example 15.3. Remember that it takes 2 points in 2D (or 3D) space to determine a line.
How many points does it take to determine a plane?

Solution. It takes 3 points. Note that a plane is just a collection of lines stacked onto each
other. We can choose such a line (2 points determine a line), and then another point that the
plane must contain. As planes must be flat surfaces, there is only one such plane that can
contain the point and the line. 

Remark. These three points must not be collinear; otherwise, they simply form a line,
which could be contained by infinitely many planes.

Exercise 15.4. Two non-parallel lines intersect at a point. How do two intersecting planes
intersect? Must two arbitrary planes intersect?
Exercise 15.5. Let lines s, t intersect at P . Must s and t both be contained in some plane?

1
Everaise Academy (2020) Math Competitions I

§15.1.3 Prisms and Polyhedra

Definition 15.6. A prism is a 3D solid figure with two congruent polygonal bases, either of
which can be translated to the other (they are at the “top” and “bottom”). The faces which
are not bases are called sides. All sides have equal height, and the edges exlusive to side faces
are perpendicular to the edges which form the bases.

You can think of a prism as taking a polygon (which will be the bases), then stacking it on
top of itself repeatedly until it is no longer flat and is three-dimensional.

Definition 15.7. A right prism is a prism where the sides are perpendicular to both bases.

Unless otherwise noted, prisms are generally assumed to be right.

Definition 15.8. A rectangular prism is a prism with a rectangular base.

Note that a rectangular prism does not necessarily have to be right, though in competition
math rectangular prisms will be assumed to be right unless otherwise specified.

Example 15.9. The volume of a right rectangular prism is lwh, where l is the length, w is
the width, and h the height of the rectangular prism.

Definition 15.10. A regular prism is a prism with a regular polygon as a base.

Definition 15.11. A parallelepiped is a prism where all faces are parallelograms.

Note that the two bases are the side facing us and the side opposite of that.
Exercise 15.12. Describe to the best of your ability the cross-section of a parallelepiped parallel
to one of its faces.
We next talk about polyhedra.

Definition 15.13. A polyhedron is a 3D shape with all faces as polygons.

Two types of polyhedra are most of interest: pyramids and regular polyhedra.

2
Everaise Academy (2020) Math Competitions I

Definition 15.14. A pyramid is a polyhedron formed by connecting all the vertices of a

polygon to a point not in the same plane as the polygon. This point is known as the apex.
Like with prisms, a pyramid is regular if it has a regular polygon as a base. A right pyramid
is a regular pyramid whose apex is directly above the center of the base.

The volume of a pyramid is 31 bh, where b is the area of the base and h is the height of the
pyramid — the distance between the base and the top vertex.

Example 15.15 (AMC 10). In the rectangular prism shown, AB = 3, BC = 1, and CG = 2.

Point M is the midpoint of F G. What is the volume of the rectangular pyramid with base
BCHE and apex M ?

H G
M
E
F 2

D C
1
A 3 B

Solution. Although it’s quite easy to find the base of said pyramid, we would need to coordinate
bash to find the height of the pyramid. As coordinate bashing was the topic of the previous
section, we leave this as an exercise for the reader.
Another method to find the volume is quite similar to complementary counting in combina-
torics. Note that plane BCHE splits prism ABCDEF GH up into two equal halves. As the
volume of ABCDEF GH is 6, the half that contains pyramid M BCHE must have volume 3.
The two regions in this half that are not contained in M BCHE are triangular pyramids
BM F E and CM GH. Use the formula stated above to show that the volumes of BM F E and
3 · 2 · 21
CM GH are both = 1. Thus, the volume of M BCHE is 2. 
3

Remark. Another name for a triangular pyramid is a tetrahedron. Some other -hedra to
know the names of:

Polygon name Number of faces

tetrahedron 4 faces
hexahedron 6 faces
octahedron 8 faces
dodecahedron 12 faces
icosahedron 20 faces

These polyhedra are the only polyhedra that can be regular. (The regular hexahedron
is the cube.) Note that the name depends on the number of faces.

Exercise 15.16 (AMC 10). The centers of the faces of the right rectangular prism with dimen-
sions 3×4×5 below are joined to create an octahedron. What is the volume of this octahedron?

3
Everaise Academy (2020) Math Competitions I

§15.1.4 Spheres, Cones, and Cylinders

Definition 15.17. A sphere is the set of points in 3D space equidistant (with distance r) from
a point O. We call O the center and r the radius.

4πr3
The volume of a sphere is V = , as you may remember.
3
Definition 15.18. A cylinder is the resulting solid after rotating a parallelogram about one
of its edges. The cylinder is right if this parallelogram is a rectangle.

h
h

r r

A right cylinder. An oblique cylinder.

Definition 15.19. A cone is the set of points on the lines from a point P to a circle ω NOT
on the same plane as the point. ω is the base and P is a cone.
Let O be the center of ω. The cone is right of OP ⊥ OX for all X on ω.

h h

r r

A right cone. An oblique cone.

Like in prisms and pyramids, cylinders and cones are assumed to be right unless otherwise
stated. We can imagine cylinders as “circular prisms” and cones as “circular pyramids”.
Therefore, the volume of a cylinder is just bh. But we know b = πr2 where r is the radius.
Thus, we have that the volume is πr2 h.
πr2 h
The volume of a cone, similarly, is from similar reasoning with cones.
3

4
Everaise Academy (2020) Math Competitions I

Example 15.20 (AMC 10). A right circular cylinder with its diameter equal to its height
is inscribed in a right circular cone. The cone has diameter 10 and altitude 12, and the
axes of the cylinder and cone coincide. Find the radius of the cylinder.

Solution. The key strategy for this problem is taking the cross-section perpendicular to a
diameter, as shown below. Let the radius of the cone be r and the midpoint of BC be M and
the midpoint of DE be N . We then have the following diagram.

We now use similar triangles. We know from AA similarity that 4AM B ∼ 4AN D. Therefore,
5 r
= . This just turns into algebra, which we leave as an exercise to the reader. Our
12 12 − 2r
30
final answer is . 
11
The key concept here was taking 2D cross-sections. In fact, most such 3D problems can be
solved by taking clever 2D cross-sections, then using our knowledge of 2D geometry.
Exercise 15.21 (AMC 8). In the cube ABCDEF GH with opposite vertices C and E, J and I
are the midpoints of segments F B and HD, respectively. Let R be the ratio of the area of the
cross-section EJCI to the area of one of the faces of the cube. What is R2 ?
E

F
H
G
J
A I
B
D

5
Everaise Academy (2020) Math Competitions I

Exercise 15.22 (AMC 12). A sphere with center O has radius 6. A triangle with sides of length
15, 15, and 24 is situated in space so that each of its sides is tangent to the sphere. What is the
distance between O and the plane determined by the triangle?

§15.1.5 Dihedral Angles

Definition 15.23. A dihedral angle is the angle between two intersecting planes. Two planes
are perpendicular if their dihedral angle is 90◦ .

In the diagram below, α and β are planes and the angle between them is the dihedral angle.

When we take cross-sections, we can think of dihedral angles as regular angles in 2D geometry.

Exercise 15.24 (AMC 12). Triangle P AB and square ABCD are in perpendicular planes. Given
that P A = 3, P B = 4, and AB = 5, what is P D?
Exercise 15.25 (AIME). A right prism with height h has bases that are regular hexagons with
sides of length 12. A vertex A of the prism and its three adjacent vertices are the vertices of
a triangular pyramid. The dihedral angle (the angle between the two planes) formed by the
face of the pyramid that lies in a base of the prism and the face of the pyramid that does not
contain A measures 60 degrees. Find h2 .

§15.2 Volume and Surface Area

Definition 15.26. The lateral surface area of an object is the area of all the sides excluding
the area of its base and top.

Proposition 15.27. The volume of an object is proportional to its lengths cubed, while the
surface area of an object is proportional to its length squared.

As a result, the volume ratio between two similar objects is equal to the cube of the ratio
of their respective lengths, and the surface area ratio between two similar objects is equal
to the square of the ratio of their respective lengths.

6
Everaise Academy (2020) Math Competitions I

Example 15.28. What is the lateral surface area of a cone with radius r and slant height
s?

Solution. The slant height of a cone is the distance between the apex and a point on the
circumference of the circular base.
Observe that we may unwrap the cone into a 2-D shape. More specifically, it unwraps into a
sector of a larger circle with radius s. The arc length of the sector is the circumference of the
base of the cone: 2πr.

Hence, the sector is 2πr

2πs of the whole circle, meaning that the area of the sector, and thus the
lateral surface area of the cone, is πs2 · 2πr
2πs = πrs. 

Example 15.29 (AMC 10). The centers of the faces of the right rectangular prism shown
below are joined to create an octahedron, What is the volume of the octahedron?

Solution. Note that the octahedron is two pyramids attached at the base. As a result, it
suffices to find the volume of one such pyramid and multiply by two. The base of the pyramid
is highlighted in green.

7
Everaise Academy (2020) Math Competitions I

Observe that the shaded figure is actually a rhombus, and its diagonals are parallel to the sides
of the prism and hence perpendicular to each other. Thus, the area of the shaded rhombus is
1
· 4 · 5 = 10.
2
Finally, the height of one of the pyramids is half of the height of the octahedron, which is the
height of the prism: 3. Hence, each prism has volume
1 3
· 10 · = 5,
3 2
meaning the volume of the final octahedron is 2 · 5 = 10. 

Example 15.30 (AIME). A cylindrical barrel with radius 4 feet and height 10 feet is full
of water. A solid cube with side length 8 feet is set into the barrel so that the diagonal of
the cube is vertical. The volume of water thus displaced is v cubic feet. Find v 2 .

Solution. The volume of water displaced is the volume taken up by the cube in the cylinder.
The portion of the cube in the water is a right triangular pyramid, as it is the corner of the cube.

Note that the condition, “the diagonal of the cube is vertical” is important here. A direct
corollary is that the line determined by the center of the cube and the center of the cylinder is
the diagonal of the cube. (Convince yourself that this is true before continuing.)

As a result, the three legs of the tetrahedron are equal because of rotational symmetry. Hence,
the desired tetrahedron has three isosceles right triangle faces and one equilateral triangle face,
resembling something like below, where the three isosceles faces meet at pairwise right angles:

8
Everaise Academy (2020) Math Competitions I

Let the legs of each of √ the isosceles right triangle faces have length x, and thus the other
three edges have length x 2. With this knowledge, we are ready to tackle the computation.
As usually helpful with 3D problems, we should consider cross-sections in order to convert this
problem into a 2D problem. Namely, the cross-section between√ the cube and cylinder is the
equilateral triangle face of our tetrahedron with length x 2!

Moreover, note that the circumcircle of this equilateral triangle is the base of the cylinder,
which is a circle with radius 4. Hence,
√ it suffices to find x such that the circumradius of the
equilateral triangle with length x 2 is 4.

This is simple enough and can be done in several ways. We present two methods.

Method 1 (Triangle Centers): Observe that the length of the circumradius of an equilateral
triangle is the distance from a vertex to the circumcenter. Recall that the centroid, orthocenter,
and circumcenter of an equilateral triangle are the same point. Hence, by centroid ratios, the
circumradius, R, is 2/3 of the altitude. Thus, we have
√
3 √ 3
R = (x 2) · ,
2 2
√
Giving that when R = 4, x = 2 6.

Method 2 (Extended Law of Sines): Recall that by the extended law of Sines,

a b c
2R = = =
sin A sin B sin C
Hence, when R = 4 in an equilateral triangle, we have
√
x 2
8= ,
sin 60◦
√
Solving for x gives x = 2 6, as well.

Of course, we are not finished! We must find the volume of the tetrahedron. However, this
is quite simple, for since the three isosceles right triangle faces are pairwise perpendicular, then
the volume is simply  
1 1 2 1
v= · x · x = x3 ,
3 2 6
√ √
whence for x = 2 6, v = 8 6. Finally, we are left with v 2 = 384. 

Exercise 15.31 (AMC 10). A pyramid with a square base is cut by a plane that is parallel to
its base and is 2 units from the base. The surface area of the smaller pyramid that is cut from
the top is half the surface area of the original pyramid. What is the altitude of the original
pyramid?
Exercise 15.32 (AIME). Two congruent right circular cones each with base radius 3 and height
8 have axes of symmetry that intersect at right angles at a point in the interior of the cones
a distance 3 from the base of each cone. Find the volume of the maximal sphere that can lie
inside both cones.

9
Everaise Academy (2020) Math Competitions I

§15.3 Homework Problems

Problem 15.1. [3] A sphere with radius r has volume 36π. If a cone with radius r and height
h has the same volume as the sphere, find h.

Problem 15.2. [4] Let A denote the region of points in space with a distance at most 1 unit
away from a 3 × 3 × 3 cube. If the volume of A can be represented as a + bπ
c where a, b, c are
integers and b, c are relatively prime, find a + b + c.

Problem 15.3 (2007 AMC 12A). [6] Corners are sliced off a unit cube so that the six faces
each become regular
√ octagons. If the total volume of the removed tetrahedra can be expressed
a−b c
in the form d , where c is square free and gcd(a, b, d) = 1, then find abcd.

Problem 15.4 (2014 AMC 12A). [6] A 4 × 4 × h rectangular box contains a sphere of radius
2 and eight smaller spheres of radius 1 (diagram below). The smaller spheres are each tangent
to three sides of the box, and the larger sphere is tangent to each of the smaller spheres. If h
√
can be expressed in the form a + b c, find a + b + c.

Problem 15.5. [6] Let ABCD be a tetrahedron, and let planes ACD and BCD intersect at
a 60◦ angle, and 4BCD is an isosceles right triangle with ∠BCD = 90◦ . Let F be the foot of
the perpendicular from A to BD. Let E be the point on DE such that BC||EF and both lines
are coplanar. If AD = 5 and DF = 3, find AE 2 .

Problem 15.6 (2005 HMMT). [8] A cube with side length 2 is inscribed in a sphere. A second
cube, with faces parallel to the first, is inscribed between the sphere and one face of the first
cube. THe side length of the smaller cube can be expressed in the form m/n where m, n are
relatively prime. Find m + n.

Problem 15.7 (2012 AIME I). [10] Cube ABCDEF GH, labeled as shown below, has edge
length 1 and is cut by a plane passing through vertex D and the midpoints M and N of AB
and CG respectively. The plane divides the cube into two solids. The volume of the larger of
the two solids can be written in the form pq , where p and q are relatively prime positive integers.
Find p + q.

10
Everaise Academy (2020) Math Competitions I

Problem 15.8 (1983 AIME). [7] The solid shown has a square base of side length s. The
upper√edge is parallel to the base and has length 2s. All other edges have length s. Given that
s = 6 2, what is the volume of the solid?

Provide a detailed solution, or explain what progress you made, even if you didn’t
get an answer.

11
16 Basics & Arithmetic Functions

§16.1 Basics
Number theory is the study of integers, and is often considered one of the most beautiful
subjects in math because of its depth. In this handout, we’ll introduce the most basic concepts
in number theory and begin exploring arithmetic functions, which will be very useful later this
week.

Definition 16.1. A positive integer p is prime if it has only 2 distinct positive divisors, 1 and
itself.

Definition 16.2. A positive integer is composite if and only if it is not prime and not equal
to 1.

You can think of a composite number as being built out of more than one (not necessarily
distinct) prime factors.

Definition 16.3. We say that n is a multiple of m if there exists an integer a such that
n = am.

Definition 16.4. We say that m is a divisor of n if n is a multiple of m; that is, there exists
an integer a such that n = am. We write m | n if m divides n and m - n if m does not divide n.

Definition 16.5. We say that n and m are relatively prime if they share no common factors
besides 1. That is, gcd(m, n) = 1.

Note that with respect to the idea of division, primes are the ”building blocks” of integers;
we can decompose any integer into a product of powers of primes. This is intuitive because
primes are the only integers that cannot be further divided, in the sense that they do not have
factors distinct from themselves.
Exercise. Prove that if a prime p divides ab where a and b are integers, then p divides a or p
divides b.

Example 16.6. Show that 37 is prime.

Solution. A prime number has no other divisors other than 1 and itself. So, let us check all
the integers between 1 and p, exclusive. We can see that 37 is not even, so 2 does not divide it.
We can also check that 3 does not divide 37 through a quick check. 4 clearly does not divide
37 since, again, 37 is not even...
This is taking quite a while. In fact, using this approach, we would have to check the
divisibility of 35 integers! Let us find some optimizations for checking if a number is prime.
First of all, note that if an integer k divides p, then kp also divides p. That is, factors come in
√
pairs. Moreover, if k < p, then
p p √
> √ = p,
k p

1
Everaise Academy (2020) Math Competitions I
√
and vice versa. This means that we only have to test the first b pc positive integers for
divisibility. If we cannot find such a divisor, the number is prime. The number cannot only
√ √
have factors greater than p, since a factor greater than p would have a factor pair less than
√
p.
This significantly reduces work, as now we only have to check for divisibility by 2, 3, 4, 5, 6,
which we can quickly do. However, what if the number was larger? We can still optimize this
process further. Notice how if, say 6 | n, then 2 | n, and 3 | n. If 2 - n, then 6 could not possibly
divide n. This leads us to see that we only have to check for divisibility of the factors’ “building
blocks”. That is, we only have to check for divisibility of prime factors. Thus, we reduce the
list of factors 2, 3, 4, 5, 6 to 2, 3, 5, which we can very easily check to determine that 37 is indeed
prime. 

By the same argument we used here, in general for an integer n, we only need to test primes
√
less than or equal to n to see if n is prime.
Taking the idea that primes are building blocks further, we can form the following definition:

Definition 16.7. A prime factorization of an integer n is a representation of n as products

of powers of primes:
n = p1 k1 p2 k2 . . . pm km .

Theorem 16.8 (Fundamental Theorem of Arithmetic). Every integer has a unique prime
factorization up to the ordering of the prime powers.

This theorem is what makes prime factorizations so important; they are unique representa-
tions of integers that give us more information on their divisors.

Example 16.9. Prime factorize 60.

Solution. Begin by checking for prime factors starting from 2 and going up. When you find
a prime factor, divide by the prime and repeat to find the highest power of the prime dividing
the integer. In this case, we first test 2 and see that 2 indeed divides 60. Now we write down
2 and test 60/2 = 30. We again see that 2 divides 30, so now write down another 2 and test
30/2 = 15. 2 does not divide 15, so we move on to 3, which divides 15. We write down 3. Now
we are left with 5, which is a prime. Checking the prime factors we’ve written down, we see the
prime factorization is 2 · 2 · 3 · 5 = 22 · 3 · 5. 

Example 16.10. Prime factorize 213 − 103 .

Solution. We could calculate 213 − 103 and prime factorize that, but there is no need to. We
quickly see that the expression is a difference of cubes. Since we can factor a difference of cubes,
and our goal is to completely factor the expression, this helps quite a lot. Applying difference
of cubes, we have

213 − 103 = (21 − 10)(212 + 21 · 10 + 102 )

= 11 · (441 + 210 + 100)
= 11 · 751

2
Everaise Academy (2020) Math Competitions I

We see that 11 is clearly prime,

√ and we can check that 751 is prime by checking divisibility
of all the primes from 2 to b 751c = 27, inclusive. Thus, our answer is 11 · 751. 

Remark. The format of the expression in the problem made it pretty obvious that we were
dealing with a difference of cubes. However, this may not always be the case. For instance,
it maybe a little more difficult to see that 9919 is a difference of squares (1002 − 92 ). The
factorization may even be a little more complicated. For example,

100064 = 104 + 4 · 24
= (102 − 2 · 10 · 2 + 22 )(102 + 2 · 10 · 2 + 22 )

is factorable by Sophie-Germain’s Identity. When coming across a special-looking number

to factor, keep an eye out for these common factorizations!

Example 16.11 (1998 AHSME). If 1998 is written as a product of two positive integers
whose difference is as small as possible, then the difference is
(A) 8 (B) 15 (C) 17 (D) 47 (E) 93

Solution. We could find all the factors of 1998, then find the factor pair with the smallest
difference, but it is a bit easier to prime factorize 1998, then separate the factors into two
groups of close products.
After some work similar to example 16.9, we can see that

1998 = 2 · 33 · 37.

We note that 37 is noticeably bigger than the other factors, so we put that aside and calculate
2 · 33 = 54. This gives a difference of 54 − 37 = 17.
We might wonder if it is possible for the factors to be even closer together. To answer this,
note that one of the two factors must have 37 as a factor. Say it is 37k, where k is an integer
greater than 2. The other factor in the pair is 1998
37k . Now, note that k ≥ 2 implies that 37k ≥ 74,
so we must have
1998 1998
≤ = 27.
37k 74
This difference is greater than 17, so (C) 17 is our answer. 

Example 16.12. Find the largest perfect square that divides 9!.

Solution. A perfect square is a number that can be written in the form k 2 for some positive
integer k. If the prime factorization of k is

p1 k1 p2 k2 . . . pm km ,

then the prime factorization of k 2 is

 2
p1 k1 p2 k2 . . . pm km
p1 2k1 p2 2k2 . . . pm 2km .

3
Everaise Academy (2020) Math Competitions I

In other words, a number is a perfect square if and only if the exponents in its prime factorization
are even.
We first prime factorize 9!:

9! = 9 · · · 2 · 1
= (32 ) · (23 ) · · · (22 ) · 3 · 2 · 1
= 27 · 34 · 51 · 71 .

We wish to find that largest perfect square that divides 9!, so we will take the largest power
with an even exponent of each prime factor. Doing so gives an answer of

26 · 34 · 50 · 70 = 5184.

Example 16.13 (1983 AIME). What is the largest 2-digit prime factor of the integer n =
200


100 ?

Solution. It is not clear what the prime factors of a binomial coefficient are; it is certainly
200


easier to find the prime factors of a factorial. Rewriting 100 with factorials gives
 
200 200!
= .
100 100! · 100!

There are two 100!’s in the denominator, which seems pretty nice. Assume the highest power
of a prime p dividing 100! is pk for some integer k. Then we know that we have a total of 2k
factors dividing the denominator, so we need at least 2k + 1 factors of p in 200!.
The most logical first case to consider is when k = 1, since we want our prime to be as big as
possible. Note that this implies that p ≥ 53 (why?). Then we know

p3 | 200!.

Since 200! is the product of every integer from 1 to 200 inclusive, if p3 | 200!, p must divide
at least 3 of these integers; we can’t have an integer with repeated factors of p since p ≥ 53.
A positive integer is divisible by p if it can be written in the form pk for some integer k. The
lowest positive values pk can take are p, 2p, 3p. Therefore, we must have 3p ≤ 200 in order for
p to appear 3 times in the product of the integers from 1 to 200.
The largest prime p that satisfies 3p ≤ 200 is 61 by trial and error. We can indeed check that
613 | 200! since

200! = 200 · 199 · · · 184 · 183 · 182 · · · 127 · 126 · 125 · · · 61 · · · 2 · 1

= 200 · 199 · · · 184 · (3 · 61) · 182 · · · 127 · (2 · 61) · 125 · · · (1 · 61) · · · 2 · 1
= 613 · (· · · )

We can also

check that 6161divides 100! once, which implies that 612 divides 100! · 100! once.
200 3
Thus, 100 is divisible by 612 = 61. 

Exercise. Prime factorize 216 − 1.

4
Everaise Academy (2020) Math Competitions I

Exercise (2001 AMC 12). Four positive integers a, b, c, and d have a product of 8! and satisfy:

ab + a + b = 524
bc + b + c = 146
cd + c + d = 104

What is a − d?

§16.1.1 Number and Sum of Divisors

Definition 16.14. The number of divisors function, denoted τ (n), is the number of positive
integer divisors of n, or X
1
d|n

Proposition 16.15. Assume n has prime factorization

p1 k1 p2 k2 . . . pm km .

Then
τ (n) = (k1 + 1)(k2 + 1) . . . (km + 1).

Proof. Consider the prime factorization any divisor of n. It can only have primes in the prime
factorization of n, and they must be to a power less than or equal to the power of that prime
in the factorization of n.
For example, assume d | n and that the maximum power of p1 dividing d is p1 r . Then we
know that 0 ≤ r ≤ k1 , so there are k1 + 1 possibilities for r.
The same logic holds for any prime pi , so let’s consider constructing a divisor of n. We first
choose what power of p1 it has from the range 0 to k1 , inclusive, then choose what power of p2
it has from the range 0tok2 inclusive, and so on. For each prime pi there are ki + 1 possibilities,
so we get
τ (n) = (k1 + 1)(k2 + 1) . . . (km + 1).


Example 16.16. Find τ (360).

Solution. Since 360 = 23 · 32 · 51 ·, we get σ(360) = (3 + 1)(2 + 1)(1 + 1) = 24. 

Example 16.17 (2005 AMC 10A). How many positive cubes divide 3! · 5! · 7!?

Solution. σ(3! · 5! · 7!) does not seem like much help, as the problem asks for factors that are
cubes. What can we do now?
First, let us find the prime factorization of the number. We have

3! · 5! · 7! = 3 · 2 · 5 · 4 · · · = 28 · 34 · 52 · 7.

5
Everaise Academy (2020) Math Competitions I

Now that we have the prime factorization, we should consider how to count the number of
factors that are cubes. We take inspiration from the derivation of the formula for the number
of divisors function.
A perfect cube is an integer that can be written as k 3 for some positive integer k. This means
that the exponents in the prime factorization of a perfect cube are divisible by 3. Given this,
how many possibilities are there for the power of 2 in the prime factorization of a cubic factor?
We can either choose 20 , 23 , or 26 , since the power of 2 in the prime factorization of 3! · 5! · 7!
is 28 . What about for 3? The power of 3 in the prime factorization of of 3! · 5! · 7! is 34 , so we
have 2 choices, 30 and 33 . Similarly, we have 1 choice for 5 and 1 choice for 7.
Our answer is the product of the number of choices for each prime factor, which is

3 · 2 · 1 · 1 = 6. 

Example 16.18 (1990 AIME). Let n be the smallest positive integer that is a multiple of
n
75 and has exactly 75 positive integral divisors, including 1 and itself. Find 75 .

Solution. Let us first ignore the multiple of 75 condition. We prime factorize 75 as 3 · 52 . Since
we wish for n = pe00 pe11 · · · to have 75 factors, we need

σ(n) = (e0 + 1)(e1 + 1) · · ·

= 75
= 3 · 52 .

Of course, it is possible to have something like k = 274 so that σ(k) = 75, but this number is
incredibly large. In order to make n small, we would want to distribute the factors more evenly.
We should also note that the exponents should be arranged in decreasing order. For example,
n = 24 · 34 · 52 is better than n = 22 · 34 · 54 since arranging the exponents in decreasing order
minimizes n. Similarly, n = 20 · 32 · 51 is greater than n = 22 · 31 · 50 , so we want to choose the
minimal exponent bases, so long as our resulting n satisfies the problem condition that 75 | n.
With these facts, we consider possible distributions. Splitting 75 = 3 · 5 · 5 gives n = 25−1 ·
35−1 · 53−1 = 24 · 34 · 52 . This is, in fact, the least we can do. If you are not convinced, you can
try things like 75 = 5 · 15, but you will see that exponential growth just inflates numbers too
quickly. Note that there may be some instances in which the prime bases grow faster; that is,
it is beneficial to split 75 into fewer factors. However, this is not one of them.
n 24 ·34 ·52
We note that n = 24 · 34 · 52 is luckily divisible by 75 = 3 · 52 , so our answer is 75 = 3·52
=
432. 

Remark. If 75 did not divide n, we might have to redistribute the factors to satisfy the
constraint. This is one of those exceptions where it is possible for the best solution to not
have decreasing exponents.

Here’s a final example that demonstrates how knowing small values of τ (n) can give us infor-
mation on what n looks like:

Example 16.19 (2019 AIME I). Let τ (n) denote the number of positive integer divisors of
n. Find the sum of the six least positive integers n that are solutions to τ (n)+τ (n+1) = 7.

6
Everaise Academy (2020) Math Competitions I

Solution. The first thing you should notice is that τ (n) and τ (n+1) are integers, so they could
be (1, 6), (2, 5), (3, 4) and permutations. However, we can actually get rid of the first case, since
τ (n) ≥ 2 (always have 1 and n as a divisor). Thus, we have four cases:
• Case 1a: τ (n) = 2, τ (n + 1) = 5. If τ (n) = 2, then it must be prime. What about
τ (n + 1) = 5? Well, the product of the exponents for each prime incremented by one
is the formula for τ , so that means that the exponents incremented by one must be of
the form (5, 1, 1 · · · 1) in some order of primes. Decreasing by one to get the exponent,
that gives (4, 0, · · · 0), so that means n + 1 must be some perfect fourth power of a prime
number! Let n = p for some prime p and n + 1 = q 4 for some prime q. We then have

n + 1 = q 4 =⇒ p + 1 = q 4 =⇒ p = q 4 − 1 = (q 2 + 1)(q 2 − 1) = (q 2 + 1)(q + 1)(q − 1)

Since p is prime, we must have two of the factors be 1 and the other be p. However,
q 2 + 1 ≥ q + 1 > q − 1, so the smallest two numbers can’t both be one since they are
always different. Thus, we have 0 solutions in this case.
• Case 1b: τ (n) = 5, τ (n + 1) = 2. We know from case 1 that this means n = q 4 for some
prime q and n + 1 = p for some prime p. This means

n + 1 = p =⇒ q 4 + 1 = p

If q = 2, we get p = 17, which works (so n = 16). However, if q is odd, then p must be
even, meaning p = 2. This corresponds to q = 1, contradiction of q is prime, so n = 16 is
the only solution in this case.
• Case 2a: τ (n) = 3, τ (n + 1) = 4. What does τ (n) = 3 mean? Well, let’s go through a
similar process for when we analyzed τ (n) = 5. Since 3 is prime, the set of all exponents
of primes incremented by 1 is (3, 1, · · · , 1), so the exponents must be (2, 0, · · · , 0). Thus,
n must be some prime squared: let n = p2 for some prime p.
What does τ (n + 1) = 4 mean? Performing a similar process, we can see that this means
the set of exponents incremented by one are either (4, 1, · · · , 1) or (2, 2, 1 · · · , 1), so the
set of exponents could be (3, 0, · · · , 0) or (1, 1, 0, · · · , 0). Thus, n + 1 is either some prime
cubed, or the product of two different primes. Suppose the latter option for now, and let
n + 1 = q 3 for some prime q. We then have

n + 1 = q 3 =⇒ p2 + 1 = q 3 =⇒ p2 = (q − 1)(q 2 + q + 1)

Note that q 2 + q + 1 > q − 1, so we must have that q − 1 = 1 and q 2 + q + 1 = p2 , giving

q = 2 =⇒ p2 = 7, a contradiction. This means we have no solution for when n + 1 is a
perfect cube of a prime. Suppose now that n = qr for some primes q and r. We then have

n + 1 = qr =⇒ p2 + 1 = qr

If p = 2, we have no solutions for q and r that give both of them are prime, so henceforth
assume p is odd. Then, qr is even, meaning either q or r is equal to 2. If both are equal
to 2, that gives p2 = 3, contradiction, so at most one can be equal to 2. WLOG, let r = 2.
We have
p2 + 1 = 2q
At this point, we can just manually find some small values of p and q. It turns out
that we have (p, q) = (3, 5), (5, 13), (11, 61), (19, 181) are some of the smallest solutions,
corresponding to n = 9, 25, 121, 361.

7
Everaise Academy (2020) Math Competitions I

• Case 2b: τ (n) = 4, τ (n + 1) = 3. For now, assume n is a perfect cube of a prime, or

n = p3 for some prime p and n + 1 = q 2 for some prime q. This gives

p3 + 1 = q 2

If p = 2, we have q = 3, a valid solution, so n = 8 works. Otherwise, p is odd, meaning q

even or q = 2. This yields a contradiction, since that means p3 = 3, so there are no other
solutions when n is of the form p3 .
Now suppose n = pq for some primes p and q, and n + 1 = r2 for some prime r. Then we
have
pq + 1 = r2 =⇒ pq = (r + 1)(r − 1)
If both p and q are odd, then r is even, meaning r = 2. This can be seen to give no
solutions, so WLOG q = 2, meaning 2p = (r + 1)(r − 1). Since r is odd, r + 1 and r − 1
are both even, or (r + 1)(r − 1) = 2p is divisible by 4. The only way for this to happen is
for p = 2, but this gives an invalid solution for r, so there are no solutions in this case.
In summary, our six smallest solutions for n are 8, 9, 16, 25, 121, 361, so our answer is 540. 

Make sure to understand this! It is extremely important to be able to figure out what form n
is in when you know τ (n). Also, another takeaway should be that analyzing parity, or whether
or not it is even or odd, is extremely useful when dealing with primes since the only even prime
is 2.

Definition 16.20. The sum of divisors function, denoted σ(n), is the sum of all positive
integers dividing n, or X
d
d|n

Proposition 16.21. Assume n has prime factorization

p1 k1 p2 k2 . . . pm km .

Then
kj
m X
Y
σ(n) = pi = (1 + p1 + · · · + p1 k1 )(1 + p2 + · · · + p2 k2 ) . . . (1 + pm + ... + pm km ).
j=1 i=0

Example 16.22. Find σ(360).

Solution. Since 360 = 23 · 32 · 51 , we get that the sum of the divisors is

σ(360) = (1 + 2 + 4 + 8)(1 + 3 + 9)(1 + 5) = 15 · 13 · 6 = 1170.

Example 16.23. Find the sum of the even factors of 360.

8
Everaise Academy (2020) Math Competitions I

Solution. An even factor must have a power of 2 greater than 0 in its prime factorization.
This leaves a lot of cases though, so we will instead calculate the sum of the odd factors, then
subtract that from the sum of all the factors.
So, how do we calculate the sum of the odd factors? First, note that since 2 does not divide
an odd factor, the factors of 2 are irrelevant. That is, if an odd factor is a factor of 2n for some
integer n, it is also a factor of n. We can also see
this through prime factorization. Thus, the
sum of all the odd factors of 360 is equal to σ 360
2k
, where k is the exponent of 2 in the prime
factorization of 360. By prime factorizing, we see that 360 = 23 · 32 · 5, so 360
2k
= 45. We can now
compute σ(45) = (1 + 3 + 9)(1 + 5) = 78.
Calculating the sum of the divisors of 360 is straightforward; it is just σ(360) = 1170.
Thus, the sum of the even divisors of 360 is 1170 − 78 = 1092. 

Exercise (Brilliant). Find the number of odd (positive) divisors of 7!

Exercise (2017 AMC 12B). The number 21! = 51, 090, 942, 171, 709, 440, 000 has over 60, 000
positive integer divisors. One of them is chosen at random. What is the probability that it is
odd?
Exercise. Prove Proposition 16.21. Hint: Expand the product given and see what you get.

§16.2 Euler Totient Function

Definition 16.24. The Euler totient function, denoted φ(n), is the number of integers from
1 to n, inclusive, that are relatively prime to n.

Proposition 16.25. If n has prime factorization

p1 k1 p2 k2 . . . pm km .

Then
m       
Y 1 1 1 1
φ(n) = n · 1− =n 1− 1− ··· 1 −
pi p1 p2 pm
i=1

or similarly
m  
pki i − pki i −1 = (p1 k1 − p1 k1 −1 )(p2 k2 − p2 k2 −1 ) · · · (pm km − pm km −1 )
Y
φ(n) =
i=1

Proof. We will count the fraction of integers from 1 to n, inclusive, that are relatively prime to
n. Then, we can simply multiply by n to get the number of integers. Note that the fraction of
integers from 1 to n relatively prime to n is equivalent to the probability of randomly selecting
an integer relatively prime to n.
In order for a number to be relatively prime to n, it must be relatively prime to each of its
distinct prime factors, p1 , p2 , · · · , pm . Note that being relatively prime to one of these primes
is independent of being relatively prime to any of the others, so the probability of a number
being relatively prime to all of these primes is simply the product of the probabilities of being
relatively prime to each prime separately.
Well, what’s the probability a number is relatively prime to a prime pi ? Let’s use comple-
mentary counting, and find the probability a number is divisible by pi (convince yourself this

9
Everaise Academy (2020) Math Competitions I

is equivalent to being not relatively prime to pi ). The number of integers from 1 to n that are
divisible by n is simply pni , since pi divides n. This means the probability of randomly selecting
 
n
one of these numbers is n = p1i . Since this is the complement, the probability of selecting an
pi

integer relatively prime to pi in the range 1 through n is simply 1 − p1i .

To get our final answer, we just need to multiply each of the probabilities and then multiply
by n to get the number of integers:
    
1 1 1
φ(n) = n · 1 − 1− ··· 1 −
p1 p2 pm
which is exactly what we claimed. 

Example 16.26. Find φ(60).

Solution. Since 60 = 22 · 31 · 51 , we get that the number of integers from 1 to 60, inclusive,
that are relatively prime to 60 is φ(60) = 60(1 − 12 )(1 − 31 )(1 − 15 ) = 60 · 12 · 23 · 45 = 16. 

Example 16.27. Find φ(p), φ(p2 ), φ(pk ) for prime p and positive integer k.

Solution. The prime factorization of any prime number p is just p1 , so φ(p) = p·(1− p1 ) = p−1.
Similarly, p2 is already prime factorized, so φ(p2 ) = p2 · · · (1 − p1 ) = p · (p − 1). In general,
φ(pk ) = pk · (1 − p1 ) = pk−1 · (p − 1). 

k
Example 16.28. Let n be a positive integer. For how many integers 1 ≤ k ≤ n is n
simplified?

Solution. A fraction ab is simplified if a and b do not share any common factors. We can also
state this as ab is simplified if gcd(a, b) = 1, that is, a and b are relatively prime. But this is
what the Euler totient function counts! Thus, we wish to find the number of integers k from 1
to n inclusive that are relatively prime to n, which is just φ(n). 

Example 16.29 (1985 AIME). How many of the first 1000 positive integers can be expressed
in the form
b2xc + b4xc + b6xc + b8xc,
where x is a real number, and bzc denotes the greatest integer less than or equal to z?

Solution. If possible, we would like to minimize the expression inside the floor functions so
that we can deal with fewer cases. That is, if we can take out an integer part of an expression
inside the floor function, it will be easier to handle. We can do this by noting that x = bxc+{x},
since {x}, the fractional part of x, equals x − bxc by definition. This will allow us to take out
that bxc out of the floor function, since we already know that it is an integer. Doing so gives
b2xc + b4xc + b6xc + b8xc = b2(bxc + {x})c + · · · + b8(bxc + {x})c
= (2bxc + b2{x}c) + · · · + (8bxc + b8{x}c)
= 20bxc + b2{x}c + b4{x}c + b6{x}c + b8{x}c.

10
Everaise Academy (2020) Math Competitions I

Now, we think about how the expression changes as x changes. If the integer part of x changes
by n, then the expression changes by 20n because of that first term.
The fractional part is a bit harder. We think about when bk{x}c changes. From our knowledge
of the floor function, we see that at k{x} = 1, 2, . . . , the expression will increase by 1. This
implies that
1 2
{x} = , , . . .
k k
are important points for us to consider.
Going back to the problem, we see that we have k = 2, 4, 6, 8 in our expression. This means
that
1 1 2 3 1
{x} = , , , , , . . .
2 4 4 4 6
are points to consider. However, we note that there may be overlap when the fraction can be
simplified. If a fraction can be simplified, this means that its numerator and denominator are
not relatively prime, which brings us to Euler!
We see that φ(2) + φ(4) + φ(6) + φ(8) = 9 is the number of fractions that cannot be simplified.
However, we need to take into account 26 = 13 and 46 = 23 since while they can be simplified,
they do not overlap with any of the other floor functions. Also, we need to include 0 for
another case. This brings us to 9 + 2 + 1 = 12 nonoverlapping changes in the expression
b2{x}c + b4{x}c + b6{x}c + b8{x}c for 0 ≤ {x} < 1.
Finally, we note that the question asks for integers between 1 and 1000, inclusive. Thus,
we can have 20bxc = 0, 20, 40, . . . , 980 for a total of 50 cases for bxc. Note that although
we can have bxc = 50, {x} = 0, we cannot have bxc = 0, {x} = 0, which is included in
20bxc = 0, 20, 40, . . . , 980. This adds and subtracts a case, canceling itself out.
Now, by doing some counting, there are 50 possibilities for bxc and 12 for {x}, leading to
50 · 12 = 600 possible integers. 

Exercise. If you know that φ(n) = k, where n is some positive integer, then what is φ(n2 ) in
terms of n and k?
Exercise. Prove that the sum of all positive integers relatively prime to and less than an integer
n > 1 is n2 φ(n). Hint: Consider pairing up integers k and n − k.

§16.3 Multiplicativity
Euler’s Totient Function has a lot of interesting applications, especially in modular arithmetic,
which you will learn about in a few days. Let’s look at a separate property of the totient
function:

Proposition 16.30. If x and y are relatively prime integers, φ(x)φ(y) = φ(xy).

Exercise. Prove this! Use the fact that if x and y are relatively prime, the sets of their prime
divisors are disjoint.
Exercise. Try to come up with a formula for φ(xy) that covers instances when x and y are not
relatively prime as well. Prove it if you want to.
We can consider φ(n) to be a multiplicative function over relatively prime integers. A
multiplicative function f satisfies the relation f (ab) = f (a)f (b) (where a and b are part of some

11
Everaise Academy (2020) Math Competitions I

defined domain). An easy example would be f (x) = x. Surprisingly, we now know of two other
multiplicative functions!

Proposition 16.31. τ (n) and σ(n) are multiplicative over relatively prime integers.

Exercise. Prove this! The proof is extremely similar to proving that φ(n) is multiplicative over
relatively prime integers.
This can be helpful for calculating different values.

Example 16.32. Let x = 2a and y = 3b for positive integers a, b. Verify that the totient
function is multiplicative in this case; i.e. Proposition 16.30 holds.

Solution. First, we note that x and y are already in prime factorized form. Thus, applying
the formula for the totient function gives us
 
a 1
φ(x) = 2 · 1 −
2
= 2a−1 .

Similarly,
 
b 1
φ(y) = 3 · 1 −
3
= 2 · 3b−1 .

Thus, φ(x) · φ(y) = 2a · 3b−1 . Now, since x and y are clearly relatively prime, we can calculate
φ(xy) and see if it is equal to φ(x) · φ(y). Since xy = 2a · 3b ,
1 1
φ(xy) = 2a · 3b · (1 − ) · (1 − )
2 3
1 2
= 2a · 3b · ( )( )
2 3
= 2a · 3b−1 .

Thus, Proposition 16.30 has been verified for this particular case of x, y. 

Example 16.33. Given that φ(p) = 8 and φ(q) = 18, where p, q are relatively prime, find
φ(pq).

Solution. By Proposition 16.30, we know that the totient function is multiplicative. Thus,
φ(pq) = φ(p) · φ(q) = 8 · 18 = 144. 

Example 16.34. Compute φ(9!).

Solution. As usual, we first prime factorize 9!, giving

9! = 9 · 8 · · · 2 · 1
= 27 · 34 · 5 · 7.

12
Everaise Academy (2020) Math Competitions I

Seeing these products, we use Proposition 16.30 to get

φ(9!) = φ(27 · 34 · 5 · 7)
= φ(27 ) · φ(34 ) · φ(5) · φ(7).

Finally, we use the fact that φ(pk ) = pk−1 (p − 1) to get

φ(9!) = φ(27 ) · φ(34 ) · φ(5) · φ(7)

= (26 · 1) · (33 · 2) · 4 · 6
= 82944.

§16.4 Homework Problems

Problem 16.1 (2007 PuMAC). [4] If you multiply all positive integer factors of 24, you get
24x . Find x.

Problem 16.2. [5] Given that n = 36019 and φ(n) = 35640, find the sum of the 2 distinct
prime factors of n.

Problem 16.3 (2013 AMC 10B). [6] A positive integer n is nice if there is a positive integer m
with exactly four positive divisors (including 1 and m) such that the sum of the four divisors is
equal to n. How many numbers in the set {2010, 2011, 2012, . . . , 2019} are nice?

Problem 16.4 (2011 PUMaC). [6] The only prime factors of an integer n are 2 and 3. If the
sum of the divisors of n is 1815, find n.

Problem 16.5 (Brilliant). [6] Let φ(n) be the Euler phi function. If 1 ≤ n ≤ 1000, what is the
smallest integer value of n that minimizes φ(n)
n ?

a
Problem 16.6 (2019 SMT). [6] How many rational numbers can be written in the form b such
that a and b are relatively prime positive integers and the product of a and b is (25!)?

Problem 16.7. [6] Find the product of all primes p for which 17p + 9 is a perfect square.

Problem 16.8 (2014 AIME I). [6] Find the number of rational numbers r, 0 < r < 1, such
that when r is written as a fraction in lowest terms, the numerator and the denominator have
a sum of 1000.

2
Problem 16.9 (2012 PuMAC). [8] How many factors of 2012 less than 2012 are not factors
of 2012 ?

Problem 16.10 (2016 AMC 10B). [8] Let f (x) = 10

P
k=2 (bkxc − kbxc), where brc denotes the
greatest integer less than or equal to r. How many distinct values does f (x) assume for x ≥ 0?
Provide a detailed solution, or explain what progress you made, even if you didn’t
get an answer.

13
17 Euclid’s Algorithm, GCD and LCM

§17.1 GCD and LCM

Definition 17.1. The greatest common divisor (GCD) of two integers m, n is the largest
integer d such that d divides both m and n.

Definition 17.2. The least common multiple (LCM) of two integers m, n is the smallest
integer q such that both m and n divide q.

Definition 17.3. Two integers a, b are called relatively prime or coprime iff gcd(a, b) = 1.

There are several ways of finding the GCD of two integers. Namely, Euclid’s algorithm,
discussed next section, does so.
However, the simplest way is to use prime factorizations. Let the prime factorizations of m
and n be defined as
∞
Y Y∞
m= pai i and n= pbi i ,
i=1 i=1

Where pi is the ith prime and ai , bi ≥ 0 for all i.

Remark. Defining ai , bi as nonnegative rather than positive allows us to define m and n

as such. They are not infinitely large; every exponent will be zero except for the primes
actually contained in the factorization. We do this because it’s much easier to then compare
exponents of corresponding primes of m and n.

It is now much easier to find gcd(m, n) and lcm(m, n) with this notation.
For d = gcd(m.n), for every i,, the exponent of pi in d must not exceed ai or bi . If not, then
by the fundamental theorem of arithmetic, there cannot equal integer k such that dk = m, n.

Moreover, since d is large as possible, then for all i, the exponent of pi must be exactly
min ai , bi . With this, we find that
∞
min{ai ,bi }
Y
d= pi .
i=1

To find q = lcm(m, n), the thinking is very similar. Instead, for all i, the exponent of pi in q
must be at least ai , bi , for m, n|q. Moreover, since q is minimal, then
∞
max{ai ,bi }
Y
q= pi .
i=1

This gives us an important property of GCD and LCM of two integers.

Proposition 17.4. For any positive integers m, n we have

gcd(m, n) · lcm(m, n) = mn.

1
Everaise Academy (2020) Math Competitions I

Proof. This should be clear given the above notations for d and q. If we simply use the previous
notation, as described by the remark, it’ll be much easier to compare exponents and then use
the fundamental theorem of arithmetic.

Remark. To prove two integers are equal, it greatly simplifies the problem if you consider
their prime factorizations and show that the exponent of each prime in their respective
prime factorizations are equal. This is true because of the fundamental theorem of arith-
metic.

∞ ∞ ∞
min{ai ,bi } max{ai ,bi } min{ai ,bi }+max{ai ,bi }
Y Y Y
gcd(m, n) · lcm(m, n) = pi × pi = pi
i=1 i=1 i=1
Now, it suffices to show that
∞ ∞
min{ai ,bi }+max{ai ,bi }
Y Y
pi = piai +bi
i=1 i=1

As mentioned earlier, it’s even simpler to just show each exponent for pi is the same. In other
words, all we have to do is show that

min{ai , bi } + max{ai , bi } = ai + bi .

This should be straightforward. It must be true that either ai ≤ bi or bi ≤ ai . However, since

the equation is symmetric in ai and bi (i.e swapping ai and bi does not change the equation),
then we can say, “Without loss of generality”, and assume that ai ≤ bi .

Then, it is clear that min{ai , bi } = ai and max{ai , bi } = bi , implying that

min{ai , bi } + max{ai , bi } = ai + bi ,

as desired. 

Remark. In summary, we did the following steps in our proof:

• We used the fact that
∞
Y ∞
Y
m= pai i and n= pbi i .
i=1 i=1

• We exploited the fundamental theorem of arithmetic to show that the exponent of

every prime in the prime factorization of both sides equation were the same
• We concluded that because either ai ≤ bi or bi ≤ ai that then min{ai , bi }+max{ai , bi } =
ai + bi .
• By the fundamental theorem of arithmetic, we then conclude that the two numbers
are indeed equal.

Exercise. Prove that gcd(a, b, c) = gcd(gcd(a, b), c).

Example 17.5. Let a and b be positive integers such that

gcd(a2 , b2 ) · lcm(a, b) − 5ab = 6lcm(a, b)

2
Everaise Academy (2020) Math Competitions I

Assume that gcd(a, b) is positive. What is the gcd(a4 , b4 )?

Solution. On the left side of the equation, we have a product of a GCD and LCM, along with
the product ab. This motivates us to consider ab = gcd(a, b) · lcm(a, b), giving the equation is

gcd(a2 , b2 ) · lcm(a, b) − 5 gcd(a, b) · lcm(a, b) = 6lcm(a, b)

We may now simply divide a factor of lcm(a, b) to get

gcd(a2 , b2 ) − 5 gcd(a, b) − 6 = 0.

This almost seems to be a quadratic in terms of gcd(a, b), but unfortunately we have gcd(a2 , b2 )
instead of gcd(a, b)2 . Thus instead, we should try to see if there exists some relationship between
gcd(a2 , b2 ) and gcd(a, b)2 .

How would we do such a thing? Why, using the definition of the GCD, of course! First,
we compute gcd(a, b)2 as
∞ ∞ ∞
min{ai ,bi } min{ai ,bi } 2 min{ai ,bi }
Y Y Y
pi × pi = pi .
i=1 i=1 i=1

Similarly, note that the factorization for a2 and b2 are, respectively,

∞
Y ∞
Y
p2a
i
i
and pi2bi .
i=1 i=1

Hence, this implies that

∞
min{2ai ,2bi }
Y
2 2
gcd(a , b ) = pi .
i=1

Like we did previously, let’s compare the individual exponents of each prime. Like last time,
note that by symmetry, we can without loss of generality let ai ≤ bi . Then,

2 min{ai , bi } = 2ai .

Similarly, since ai ≤ bi , then 2ai ≤ 2bi , so we find

min{2ai , 2bi } = 2ai ,

Meaning that in fact, 2 min{ai , bi } = min{2ai , 2bi }. Then, by the fundamental theorem of
arithmetic, we can conclude that
∞ ∞
min{2ai ,2bi } 2 min{ai ,bi }
Y Y
2 2 2 2
gcd(a , b ) = pi = gcd(a , b ) = pi = gcd(a, b)2 !
i=1 i=1

Now, substituting into our original equation gives the quadratic

gcd(a, b)2 − 5 gcd(a, b) − 6 = 0.

This conveniently factors as

(gcd(a, b) − 6)(gcd(a, b) + 1) = 0.

Clearly, gcd(a, b) > 0, so this gives that gcd(a, b) = 6. Now, with the exact same reasoning as
before, gcd(a4 , b4 ) = gcd(a, b)4 , meaning that our desired answer is gcd(a, b)4 = 64 = 1296 . 

3
Everaise Academy (2020) Math Competitions I

Example 17.6 (2018 AMC). How many ordered pairs (a, b) of positive integers satisfy the
equation
a · b + 63 = 20 · lcm(a, b) + 12 · gcd(a, b),
where gcd(a, b) denotes the greatest common divisor of a and b, and lcm(a, b) denotes their
least common multiple?

Solution. This should look familiar now. We have an equation in terms of lcm(a, b) and
gcd(a, b) as well as the product ab. Thus, in order to get everything in terms of lcm(a, b) and
gcd(a, b), we can substitute ab = gcd(a, b) · lcm(a, b).

gcd(a, b) · lcm(a, b) + 63 = 20 · lcm(a, b) + 12 gcd(a, b).

We may rearranged the equation to move all of the lcm(a, b) and gcd(a, b) on one side and
constants on the other as

gcd(a, b) · lcm(a, b) − 20 · lcm(a, b) − 12 gcd(a, b) = −63.

We now use the colloquial “Simon’s Favorite Factoring Trick (SFFT)”. Where we ultimately
want to incorporate the product (gcd(a, b) − 20)(lcm(a, b) − 12) into the equation.

Remark. In general, Simon’s Favorite Factoring Trick states that

xy + xk + yj + jk = (x + j)(y + k)

Hence, if you ever see an expression of the form xy+xk+yj, it may be helpful to manipulate
the constant term to achieve such a factoring.

Thus, since

(gcd(a, b) − 20)(lcm(a, b) − 12) = gcd(a, b) · lcm(a, b) − 20 · lcm(a, b) − 12 gcd(a, b) + 240,

We should add 240 to both sides of the previous equation to yield

(gcd(a, b) − 20)(lcm(a, b) − 12) = 240 − 63 = 177.

Now, we have a product of integers equalling 177, so we must consider all possible pairs of
integers multiplying to 177. These are:

177 = 177 × 1
177 = 59 × 3
177 = (−59) × (−3)
177 = (−177) × (−1)

However, right away, we can eliminate cases 3 and 4, for they would result in either gcd(a, b) or
lcm(a, b) to be less than 0.

Moreover, note that gcd(a, b) ≤ lcm(a, b), meaning that gcd(a, b) − 20 < lcm(a, b) − 12.

Case 1: lcm(a, b) − 12 = 177 and gcd(a, b) − 20 = 1.

4
Everaise Academy (2020) Math Competitions I

Clearly, then lcm(a, b) = 189 and gcd(a, b) = 21. As a result, by definition, since gcd(a, b) = 21,
then a = 21a0 and b = 21b0 where gcd(a0 , b0 ) = 1. (Note: had it been true that gcd(a0 , b0 ) = k >
1, then gcd(a, b) = 21k.)

Then, lcm(a, b) = 21a0 b0 = 189, meaning that we must find a0 , b0 such that a0 b0 = 9 and a0 b0 are
relatively prime. This has two ordered pairs of (1, 9) and (9, 1), giving 2 ordered pairs.

Case 2: lcm(a, b) − 12 = 59 and gcd(a, b) − 20 = 3.

Clearly, then lcm(a, b) = 71 and gcd(a, b) = 23. As a result, by definition, since gcd(a, b) = 23,
then a = 23a0 and b = 23b0 where gcd(a0 , b0 ) = 1.

Then, lcm(a, b) = 23a0 b0 = 71. However, 23 - 71, resulting in a contradiction. Thus, there
are 0 ordered pairs for this case.

In conclusion, from both cases, there exists only 2 + 0 = 2 ordered pairs satisfying the given
conditions, namely (189, 21) and (21, 189). 

Exercise 17.7. How many ordered pairs of integers (a, b) satisfy

gcd(a, b) = 20 and lcm(a, b) = 180?

Exercise 17.8 (CMC Year 3). How many ordered triples (a, b, c) of not necessarily distinct
positive integer divisors of 216, 000 satisfy
lcm(a, gcd(b, c)) = lcm(b, gcd(a, c)) = lcm(c, gcd(a, b))?

§17.2 Euclidean Algorithm

Before we get into the Euclidean Algorithm, we first need to start with the division algorithm.
Definition 17.9. The division algorithm takes two positive integers and returns the quotient
and remainder when one divides the other. The algorithm starts with the two numbers and
takes the divisor (this is the number that is “doing the dividing”) and finds the largest multiple
of the divisor which is less than or equal to the dividend (the number “being divided”).
Once you find this multiple, the number that you multiplied the divisor by is the quotient,
the difference between the multiple that was found and the dividend is the remainder.

Mathematically, if we have an integer a and are dividing by b, then the division algorithm
finds us an integer q (the quotient) and an integer r (the remainder) such that 0 ≤ r < b and
a = bq + r.
Definition 17.10. The Euclidean Algorithm finds the GCD of two numbers by using the
division algorithm until you either get that one number is the multiple of another or that the
GCD of the numbers is 1. First you start with the two numbers you want, then use the division
algorithm where the smaller number is set as the divisor and the larger number is the dividend,
then get the remainder r as defined above.
If the remainder is 0 then one number is a multiple of the other and the GCD of the two
numbers is the smaller of the two numbers; if the remainder is 1 the GCD of the two numbers
is 1. otherwise replace the larger of the two numbers with the remainder you found.

5
Everaise Academy (2020) Math Competitions I

Mathematically, letting a = bq + r as described in the division algorithm, the Euclidean

algorithm uses the fact that gcd(a, b) = gcd(r, b) repeatedly.

It may not seem evident how this ends up working, but once we do some examples it should
be easier to understand. Before we get into the examples, though, we encourage you to prove
the following statement:
Exercise. Prove that if a and b are integers such that a > b, then gcd(a, b) = gcd(a − b, a).
The reason the Euclidean algorithm works actually boils down to this simple fact; letting
a = bq + r as described in the division algorithm, we have that

gcd(a, b) = gcd(bq + r, b) = gcd(b(q − 1) + r, b) = · · · = gcd(r, b)

which is the fundamental assumption of the Euclidean algorithm.

For finding the GCD of two given integers, using division at each step is the fastest method,
but if we are trying to find the GCD of algebraic expressions its often simpler to just reduce
using the fact that gcd(a, b) = gcd(a − b, a).

Now we’ll get into the examples.

Example 17.11. Find the GCD of 1003 and 68.

Solution. Now we perform the Euclidean Algorithm:

1003 = 68 · 14 + 51
68 = 51 · 1 + 17
51 = 17 · 3 + 0

Note that in each subsequent step, we take the divisor from the previous step and use the
remainder from the previous step as our new divisor. We stop once we reach 0.

The first equation in our Euclidean algorithm signifies that gcd(1003, 68) = gcd(51, 68). The
second signifies that gcd(51, 68) = gcd(51, 17), and the last equation signifies that gcd(51, 17) =
gcd(17, 0), which we know is 17. Thus, gcd(1003, 68) = gcd(51, 68) = gcd(51, 17) = gcd(17, 0) =
17. 

Example 17.12. Find the GCD of 2002 and 78.

Solution.
2002 = 25 · 78 + 52
78 = 1 · 52 + 26
52 = 2 · 26 + 0
This is done in the same way as the previous example. The first equation tells us that
gcd(2002, 78) = gcd(52, 78). The second equation then tells us that gcd(52, 78) = gcd(52, 26).
The last equations tells us that gcd(52, 26) = 26 since 52 is a multiple of 26. From these three
equations we see that gcd(2002, 78) = gcd(52, 78) = gcd(52, 26) = 26. 

6
Everaise Academy (2020) Math Competitions I

21n+4
Example 17.13 (1959 AIME). Prove that the fraction 14n+3 is irreducible for every natural
number n.

Solution. We will prove that gcd(21n + 4, 14n + 3) = 1. Using the Euclidean algorithm gives

gcd(21n + 4, 14n + 3) = gcd(7n + 1, 14n + 3) = gcd(7n + 1, 1) = 1.

Thus, our claim was proven and we are done. 

Example 17.14 (1985 AIME). The numbers in the sequence 101, 104, 109, 116,. . . are of
the form an = 100 + n2 , where n = 1, 2, 3, . . . For each n, let dn be the greatest common
divisor of an and an+1 . Find the maximum value of dn as n ranges through the positive
integers.

Solution. From the given information, we have that

dn = gcd(an , an+1 ) = gcd(100 + n2 , 100 + n2 + 2n + 1).

We notice that we can use the Euclidean algorithm here to find the GCD; we just need to figure
out how to cancel out all the terms with n. Applying the Euclidean algorithm once gives

gcd(100 + n2 , 100 + n2 + 2n + 1) = gcd(100 + n2 , 2n + 1).

Now we want to cancel out the n2 term, and we must do so by subtracting a multiple of 2n + 1.
We can subtract n2 (2n + 1), to cancel it, but how do we know if n2 is an integer? It depends on
whether n is even or odd. We could just do casework based on the parity of n, but there’s a
faster solution; we know that 2n + 1 is odd, so

gcd(100 + n2 , 2n + 1) = gcd(200 + 2n2 , 2n + 1)

since changing the even factors of the left integer won’t change the gcd. Then we can subtract
n(2n + 1) which gives

gcd(200 + 2n2 − n(2n + 1), 2n + 1) = gcd(200 − n, 2n + 1)

Finally, we can add 2(200 − n) to the right integer and get dn = gcd(200 − n, 401). We’ve finally
got a constant value. Then the maximum possible value of dn is 401, which is achievable by
taking n = 601, so our answer is 401. 

Exercise 17.15. Find gcd(999, 148).

Exercise 17.16. Find gcd(20, 45, 36) using the Euclidean Algorithm.

§17.3 Homework Problems

Problem 17.1. [3] Let I + 3 = W M for positive integers I, W, M such that W + M = 6. Find
the sum of all distinct values of gcd(I 3 , W M ).

Problem 17.2. [4] Find gcd(999999, 6512) using the Euclidean Algorithm.

7
Everaise Academy (2020) Math Competitions I

Problem 17.3 (2013 AMC 8). [4] What is the ratio of the least common multiple of 180 and
594 to the greatest common factor of 180 and 594?

Problem 17.4 (2016 AMC 8). [5] The least common multiple of a and b is 12, and the least
common multiple of b and c is 15. What is the least possible value of the least common multiple
of a and c?

Problem 17.5 (2020 AMC 10 A). [6] Let n be the least positive integer greater than 1000 for
which
gcd(63, n + 120) = 21 and gcd(n + 63, 120) = 60.
What is the sum of the digits of n?

Problem 17.6 (2019 PUMAC). [6] The least common multiple of two integers a and b is
25 × 35 . How many such ordered pairs (a, b) are there?

Problem 17.7 (2016 AMC 10/12A). [7] How many ordered triples (x, y, z) of positive integers
satisfy lcm(x, y) = 72, lcm(x, z) = 600, and lcm(y, z) = 900?

Problem 17.8 (2017 CMIMC). [7] Determine the sum of all possible values of m + n, where
m and n are positive integers satisfying

lcm(m, n) − gcd(m, n) = 103.

Problem 17.9 (2018 PUMaC). [8] Let n be a positive integer. Let f (n) be the probability that,
if divisors a, b, c of n are selected uniformly at random with replacement, then gcd(a, lcm(b, c)) =
1
lcm(a, gcd(b, c)). Let s(n) be the sum of the distinct prime divisors of n. If f (n) < 2018 , compute
the smallest possible value of s(n).

Problem 17.10 (2020 AMC 12A). [6] How many positive integers n are there such that n is a
multiple of 5, and the least common multiple of 5! and n equals 5 times the greatest common
divisor of 10! and n?
Provide a detailed solution, or explain what progress you made, even if you didn’t
get an answer.

8
18 Modular Arithmetic, Linear
Congruences, and Divisibility Rules

§18.1 Modular Arithmetic

The motivation of modular arithmetic is to simplify problems involving remainders. Consider
the following problem:

Example 18.1. What is the remainder when 62020 is divided by 5?

Solution. Assume that you have no idea what modular arithmetic is. Recall the binomial
theorem, which details the expansion of (a + b)n . Let’s rewrite 6 as 5 + 1, so the problem
changes to: the remainder when (5 + 1)2020 is divided by 5. Then, the binomial theorem tells
us that the expanded version of this is
     
2020 2020 2019 1 2020 2018 2 2020 1 2019
5 + 5 1 + 5 1 + ··· + 5 1 + 12020 .
2019 2018 1

Every term in this expansion except for the last has a factor of 5 in it, so the remainder when
this is divided by 5 will just be equal to 1. 

It doesn’t seem like a coincidence that the remainder of 6 divided by 5 is equal to 1, and the
remainder of 62020 when divided by 5 is equal to 12020 . Let’s try to formalize this notion by
introducing the concept of modular arithmetic.
We say that two integers a, b are congruent modulo n (modulo is sometimes shortened to
mod) if a − b is divisible by n, and we denote this as a ≡ b (mod n). (Read: a is congruent to
b mod n.) (Recall that we say a | b if there is some integer c such that ac = b.) We refer to n
as the modulus. We can heuristically think of a and b as leaving the same remainder when
divided by n.
Going back to the first example, we can write

62020 ≡ 1 (mod 5),

pronounced as “62020 is congruent to 1 modulo 5.” Note that the ≡ sign means “congruent,” so
we refer to these statements as congruences.
Now we’ll introduce some properties of modular arithmetic. You’ll start to see that congru-
ences share many properties with equations, such as addition, subtraction, and multiplication
to both sides preserving it. Try to think about these properties in terms of remainders and the
formal definition.
• a ≡ a (mod n)
• If a ≡ b (mod n), then b ≡ a (mod n)
• If a ≡ b (mod n) and b ≡ c (mod n), then a ≡ c (mod n)
• a is a multiple of n if and only if a ≡ 0 (mod n)
• If a ≡ b (mod n) and c ≡ d (mod n), then a + c ≡ b + d (mod n) and ac ≡ bd (mod n)

1
Everaise Academy (2020) Math Competitions I

• If a ≡ b (mod n), then am ≡ bm (mod n)

• If ma ≡ mb (mod n), then a ≡ b (mod n/ gcd (m, n))
• If a ≡ b (mod n), then ak ≡ bk (mod n)
Exercise. Prove the above properties using our definition of congruence modulo n.

Example 18.2. Prove that if a ≡ b (mod n), then ak ≡ bk (mod n).

Solution. Note that n | a − b by the definition of modular arithmetic, and a − b | ak − bk . Thus

n | ak − bk , as desired. 

Remark. This property applies to our earlier example. Instead of going through the entire
binomial theorem, we can simply say that since 6 − 1 is divisible by 5, then we have 6 ≡ 1
(mod 5) =⇒ 62020 ≡ 12020 ≡ 1 (mod 5).

In the context of this problem and plenty of other MATHCOUNTS remainder problems, it
is helpful to define a least residue.

Definition 18.3. The least residue of a (mod n) is the the value of l with 0 ≤ l ≤ n − 1 such
that a ≡ l (mod n). In common language, l is just the “remainder” when a is divided by n.

Example 18.4. Pi day this year (2020) was on Saturday, March 14. On what day of the
week will Pi day be, in 2023?

Solution. We don’t have to worry about leap years (since February of 2020 passed before Pi
day), so we know that there will be 3 · 365 = 1095 days until Pi day 2023. Since the days of the
week repeat every 7 days, the answer is essentially the least residue of 1095 mod 7, converted
into the day of the week.
We could use long division, but we could also find this least residue by subtracting obvious
multiples of 7 until we get a number small enough. For example, 700 is a multiple of 7, so
we have 1095 ≡ 1095 − 700 ≡ 395 (mod 7). Continuing, 395 ≡ 395 − 350 ≡ 45 ≡ 3 (mod 7).
Therefore, the day of the week will be 3 days after Saturday, or Tuesday . 

Example 18.5 (MATHCOUNTS 2016). What is the remainder when 12 +22 +32 +· · ·+20162
is divided by 17?

Solution. We know that

n
X n(n + 1)(2n + 1)
i2 =
6
i=1
from algebra. Thus, the sum from the problem turns into
2016(2017)(4033)
= 336 · 2017 · 4033.
6
Now, we compute the least residues of each of these three numbers.

336 ≡ 13 (mod 17), 2017 ≡ 2017 − 1700 ≡ 317 ≡ 11 (mod 17),

2
Everaise Academy (2020) Math Competitions I

and
4033 ≡ 2(2017) − 1 ≡ 2(11) − 1 ≡ 21 ≡ 4 (mod 17).
Multiplying these numbers back together, we obtain

13 · 11 · 4 ≡ 143 · 4 ≡ 7 · 4 ≡ 11 (mod 17),

and the remainder is 11 . 

§18.1.1 Negative Residues in Modular Arithmetic

So far, we’ve only worked with positive residues in modular arithmetic. But what about negative
residues? Recall the formal definition: two numbers a, b are congruent modulo n if a − b = kn,
for some integer k. This definition can be thought as: “If the difference between two numbers
is divisible by n, then they are congruent mod n.”
If a or b are negative, then that’s fine, as long as their difference is divisible by n. Negative
numbers also follow all of the earlier properties. A common misconception is that a ≡ −a
(mod n). This is not necessarily true, since this does not necessarily follow the formal definition.
Let’s do an example.

Example 18.6. Find a if a ≡ 30 (mod 13) and a is the greatest negative integer that
satisfies the congruence.

Solution. We have that the difference between 30 and a must be divisible by 13. In other words,
30 − a = 13k. Since we know that a is negative, then on the LHS, we are essentially subtracting
a negative number, or adding. Therefore, the greatest negative value for a is −9 . 

In general, we can easily see that if a 6= 0 is some least residue modulo n, then the least residue
corresponding to −a is simply n − a.

Example 18.7. Determine whether each of the following pairs of integers is congruent
modulo 9:
1. (−1, 899)
2. (−17, −171)
3. (−90, 900)
4. (−55, 755)

Solution. For two numbers to be congruent mod 9, their difference must be divisible by 9. For
the first pair, the difference is 900 or −900, depending on which way you subtract. Either way,
this difference is divisible by 9, so we have −1 ≡ 899 (mod 9).
For the second pair, the difference is 154 = 14 × 11, which is not divisible by 9, so we can
write −17 6≡ −171 (mod 9).
For the third pair, the difference is 990, so we have −90 ≡ 900 (mod 9).
For the fourth pair, the difference is 810, we have −55 ≡ 755 (mod 9). 

Now, let’s see how using residues numbers could be helpful:

3
Everaise Academy (2020) Math Competitions I

Example 18.8. What is the units digit of 177 + 188 ?

Solution. How do we find the units digit of something using modular arithmetic? Note that
the units digit of any number m is equal to the remainder when m is divided by 10. In other
words, it is the least residue of m (mod 10).
Now, we’re ready to tackle this problem. Let’s break it up. First, we find the units digit of
177 .
We have 177 ≡ 77 (mod 10). Notice that 72 ≡ −1 (mod 10) =⇒ 77 ≡ (76 )7 ≡ (−1)3 · 7 ≡ 3
(mod 10).
Now, we focus on the 188 . We have 188 ≡ 88 ≡ (−2)8 ≡ (−2)4 (−2)4 ≡ 16 · 16 ≡ 6 · 6 ≡ 6
(mod 10). Therefore, 177 + 188 ≡ 3 + 6 ≡ 9 (mod 10). 

Notice how we used negative numbers to significantly reduce the numbers that we were
working with, in order to simplify calculations and to arrive at an answer quicker.

§18.1.2 Modular inverses

You might notice that we’ve stayed away from division thus far. This is because division is
quite weird in modular arithmetic.

Definition 18.9. A multiplicative inverse of a modulo n is denoted as a−1 (you might

also sometimes see a), and satisfies a · a−1 ≡ 1 (mod n).

This is an important concept because 1 is a very nice number to work with, especially when
dealing with multiplication of big numbers or exponents, and it will also help us understand
linear congruences better. More on that later.
It is also important because in situations where we might divide both sides when we’re solving
an equation, we might not be able to do that in a congruence without tweaking the modulus
or making other changes. Thus, we introduce the inverse as a way to remedy this; instead of
dividing both sides, we multiply both sides by an inverse.
However, modular inverses do not always exist. We’ll go into when they do later in the
section.
Finally, note that if b is the inverse of a modulo n, then any other value in the same residue
class as b would also be an inverse of a. In terms of congruences, if

ab ≡ 1 (mod n) and b ≡ c (mod n) =⇒ ac ≡ 1 (mod n) =⇒ b ≡ c ≡ a−1 (mod n).

You have to be careful when stating a−1 ≡ a1 . Fractions and rational numbers are not well-
defined with remainders and modular arithmetic, so it’s always best to prove statements such as
1 1 a+b
a + b ≡ ab (mod n). Grading issues have occurred in olympiads because of this, so we present
an example that properly uses addition of inverses.
Feel free to ignore this example if you do not know Fermat’s Little Theorem. This example
is absolutely not essential.

Example 18.10 (2005 IMO). Determine all positive integers relatively prime to all the terms
of the infinite sequence
an = 2n + 3n + 6n − 1, n ≥ 1.

4
Everaise Academy (2020) Math Competitions I

Solution. We claim the answer is just 1. It is easy to show that it works.

Note that a2 = 12, showing 2 and 3 do not work.
Now we just show that 2p−1 + 3p−1 + 6p−1 − 1 ≡ 0 (mod p). This heuristically should work
since
1 1 1
2p−1 + 3p−1 + 6p−1 − 1 ≡ + + − 1 ≡ 0 (mod p).
2 3 6
1 1 1
Now we have to show that 2 + 3 + 6 ≡ 1 (mod p) is valid.
We claim that in general, a−1 + b−1 ≡ (a + b)(a−1 b−1 ) (mod p). Note that by the distributive
property, (a + b)(a−1 b−1 ) ≡ aa−1 b−1 + bb−1 a−1 ≡ b−1 + a−1 (mod p), so adding fractions is
valid. 

When working with congruences, it’s important to keep in mind that we’re always working
in integers. Fractions are just a convenient shorthand that happen to work - you must keep in
mind why and when they do.

Example 18.11. Find the inverse of 2, 12, and 5 modulo 55.

Solution. Let’s start with finding 2−1 . We want to find a number x such that 2x ≡ 1 (mod 55).
You might observe that x = 28 works because 2 · 28 = 56 ≡ 1 (mod 55), so we know that any
number congruent to 28 modulo 55 would be an inverse of 2.
There is a more rigorous approach for finding modular inverses using the Euclidean algorithm.
Recall that modular arithmetic is fundamentally just another way to represent remainders.
When working with modular arithmetic, don’t forget that in the end, we’re ultimately dealing
with a division problem.
Now let’s find 12−1 . We want to find a solution for x in the congruence 12x ≡ 1 (mod 55). In
other words, we want to find integer solutions for the equation 12x = 55q+1 =⇒ 12x−55q = 1.
Now, we’ll illustrate how to use the Euclidean algorithm:

55 = 12 · 4 + 7 gcd (55, 12)

12 = 7 · 1 + 5 = gcd (7, 12)
7=5·1+2 = gcd (7, 5)
5=2·2+1 = gcd (2, 5).

Note that we ended with 1; this is no coincidence, but we’ll clear this up later. Because we have
a 1, we can backtrack to express it as a linear combination of 12 and 55:

1=5−2·2
= 5 − 2(7 − 5 · 1)
=5·3−2·7
= (12 − 7 · 1) · 3 − 2 · 7
= 12 · 3 − 5 · 7
= 12 · 3 − 5(55 − 12 · 4)
= 12 · 23 − 55 · 5.

Thus, x = 23 and q = 5, and the inverse of 12 modulo 55 is any number congruent to 23

modulo 55. Notice how we used the equations that we found using the Euclidean Algorithm

5
Everaise Academy (2020) Math Competitions I

and substitution to arrive at an answer. Besides finding the gcd of two very big numbers, this
is one of the many applications of the Euclidean Algorithm.
Now, to find 5−1 . We can use the same technique, and write the equation 5x = 55q + 1 =⇒
5x − 55q = 1 =⇒ 5(x − 11q) = 1 =⇒ x − 11q = 15 . Hold on. There’s no way we can
find integers x and q such that x − 11q is a fraction! What does this mean? This means that
5−1 simply does not exist. There is no number that we can multiply to 5 in order to obtain a
number that is congruent to 1 mod 55. 

This raises the question: When does a−1 modulo n exist?

Proposition 18.12. a−1 exists modulo n. iff gcd (a, n) = 1.

Proof. First we’ll prove the forwards direction (i.e. that the first statement implies the second).
Suppose to the contrary that g = gcd (a, n) > 1. We want to show that there are no solutions
for x in the congruence
ax ≡ 1 (mod n).
We can use the same strategy as before and write out the equation: ax = 1+qn =⇒ ax−qn = 1.
Don’t forget that a, x, q, and n are all integers! We factor out a g on the LHS, so our new equation
becomes
 
a n a n 1
g x − q = 1 =⇒ x − q = .
g g g g g
However, by definition, g is a factor of both a and n, so we know that ag and ng are both integers.
Therefore, the LHS must be an integer. The RHS will not be an integer because g > 1, and
thus we have reached a contradiction.

For the reverse direction, note that the Euclidean algorithm with a and n must terminate in
1 (the gcd of the two numbers), so backtracking will give us a linear combination ax + ny = 1,
where x will be our desired a−1 . 

Exercise. Make the argument for the reverse direction more rigorous using the Euclidean algo-
rithm and strong induction.

Remark. Note that this means that any nonzero residue mod a prime p will have an inverse;
this is handy to remember since we’ll often be working with a prime modulus in problems.

Proposition 18.13. If the multiplicative inverse of a exists modulo n, then exactly one
value between 0 and n − 1 is an inverse of a. In other words, if a has an inverse, then it
is unique modulo n.

Proof. Suppose to the contrary that the inverse of a exists and there exist two values i, j, such
that 0 ≤ i < j ≤ n − 1, and aj ≡ ai ≡ 1 (mod n). Combining, we have
aj ≡ ai (mod n) =⇒ a(j − i) ≡ 0 (mod n).
Since a has an inverse, then gcd (a, n) = 1. However, we must have a(j − i) as a multiple of n,
so we conclude that j − i is a multiple of n. By construction, j − i ≤ n − 1, so we have arrived
at a contradiction. 

6
Everaise Academy (2020) Math Competitions I

Example 18.14. Given that p is a prime, for which integer values of a between 0 and p − 1,
inclusive, satisfy the property that a is its own inverse?

Solution. Recalling the definition, we want to find all values of a for which a·a ≡ 1 (mod p) =⇒
a2 ≡ 1 (mod p) =⇒ a2 − 1 ≡ 0 (mod p) =⇒ (a + 1)(a − 1) ≡ 0 (mod p). From this, we see
that either a + 1 or a − 1 must be a multiple of p. Since we are restricted by 0 ≤ a ≤ p − 1, our
only possible values for a would be a = p − 1 and a = 1. 

Remark. It should be noted that the integers modulo n are divided into n distinct residue
classes. Each residue class is a set of integers that are all congruent to each other modulo
n.

§18.1.3 Exercises for this section

Exercise 18.15. What time does a clock read 100 hours after it reads 2 o’ clock? (The clock is
an analog, 12-hour clock)
Exercise 18.16. Find the remainder when 212 is divided by 13. Hint: Build up! 212 ≡ 28 · 24 , and
28 ≡ (24 )2 .

Exercise 18.17. Find the remainder when 10! is divided by 11. Hint: Use some negative numbers
to simplify calculations.

Exercise 18.18. Determine which integers a where 1 ≤ a ≤ 24 have an inverse modulo 24.
Exercise 18.19. For which positive integers n is it true that
n−1
X
i2 ≡ 0 (mod n)?
i=1

What about if n satisfies

n−1
X
i3 ≡ 0 (mod n)
i=1

instead?
Exercise 18.20. Prove that if p is a prime number, and x and y are positive integers such that
x ≡ y (mod p), then xp ≡ y p (mod p2 ). Hint: Move everything to one side. Now what can we do?

§18.2 Linear Congruences

A congruence in the form
ax ≡ b (mod n)
is called a linear congruence, for positive integers a, b, n, because the degree of x is 1. In this
congruence, we want to solve for x, which is also an integer.
Similar to inverses, if x0 is a solution to the congruence ax ≡ b (mod n) for x, then any other
number in the same residue class as x0 is also a solution.

7
Everaise Academy (2020) Math Competitions I

Now, when does this not have a solution for x? Let’s consider the congruence

2x ≡ 1 (mod 6).

Doing a quick analysis, we see that 2x is always an even number, and that it must be 1 more
than a multiple of 6. Well, any number that’s congruent to 1 modulo 6 must be odd, so there
aren’t any solutions to this linear congruence. This raises the question: When does a linear
congruence have a solution for x? Let’s do some examples to get more comfy with the idea of
linear congruences.

Example 18.21. Solve for x if 2x ≡ 3 (mod 5).

Solution. We’ve solved linear congruences before when we dealt with inverses, you just didn’t
know that they were called linear congruences until now. Recall how we introduced inverses as
an alternative to division in modular arithmetic. Normally, to solve the equation 2x = 3, you
would just divide by sides by 2, to get x = 32 . However, modular arithmetic and fractions don’t
go together! Let’s use inverses here. Then, we have

2x ≡ 3 (mod 5) =⇒ 2x · 2−1 ≡ 3 · 2−1 (mod 5) =⇒ x ≡ 4 (mod 5),

because 2 · 2−1 ≡ 1 (mod 5), and 2−1 ≡ 3 (mod 5). Thus any number congruent to 4 (mod 5)
is a solution to the congruence. 

Hmmm, so this congruence had a solution, but our earlier one (2x ≡ 1 (mod 6)) did not. You
might suspect that it’s because we needed to multiply both sides by 2−1 in order to isolate x,
and that gcd (2, 6) 6= 1, so the inverse of 2 doesn’t exist. Let’s see how we might need to modify
this argument...

Example 18.22. Solve for x in the congruence 9x ≡ 12 (mod 15).

Solution. Here, gcd (9, 15) 6= 1, so by our previous argument, this congruence should not have
a solution. However, recall that we can divide both sides by 3, so we get
9 12
9x ≡ 12 (mod 15) =⇒ x≡ (mod 15/ gcd (3, 15)) =⇒ 3x ≡ 4 (mod 5).
3 3
Now, this congruence has a solution: x ≡ 3 (mod 5). Let’s list some actual concrete numbers
that x could be: x = 3, 8, 13, 18, 23, ... Looking back at the original problem, we see that the
original modulus was 15, and 3 6≡ 8 6≡ 13 (mod 15), which means that x has three different
solutions modulo 15! 

This example brings us to the following proposition:

Proposition 18.23. Consider the linear congruence

ax ≡ b (mod n)

and let gcd (a, n) = g. If g - b, then no solutions exist. Otherwise, the congruence has g
distinct solutions modulo n.

8
Everaise Academy (2020) Math Competitions I

§18.2.1 Exercises for this section

Exercise 18.24. Verify that Proposition 5.21 holds for the three examples in this section.
Exercise 18.25. Solve for x in the congruence 5x + 7 ≡ 9 (mod 11).

§18.3 Divisibility Rules

The last part of this handout deals with a direct application of modular arithmetic: divisibility
rules. Let’s recap with some basic information. We use a base 10 number system, which means
that the number 478 = 4 · 100 + 7 · 10 + 8 · 1 = 4 · 102 + 7 · 101 + 8 · 100 , where the 8 is called
the units or ones digit, the 7 is called the tens digit, and the 4 is called the hundreds digit. For
the rest of this section, we will denote the units digit as a0 , the tens digit as a1 , and so on.
Essentially, if we have an n-digit number, then it can be written as
n−1
X
10i ai ,
i=0

where 0 ≤ ai ≤ 9, for all i.

Recall that if a number k is divisible by n, then we can write k ≡ 0 (mod n). We will prove
most of our divisibility rules this way.

Example 18.26 (Divisibility rule for 3). Prove that if the sum of the digits of a number is
divisible by 3, then the original number is also divisible by 3.

Solution. You probably learned this fact in elementary or middle school, but now, it’s time to
prove it. Let this number have n digits, so we can write it as

n = 10n−1 an−1 + 10n−2 an−2 + · · · + 101 a1 + 100 a0 .

Now, notice that 10 ≡ 1 (mod 3) =⇒ 10k ≡ 1 (mod 3). Taking n modulo 3 gives us

10n−1 an−1 + 10n−2 an−2 + · · · + 101 a1 + 100 a0 ≡ an−1 + an−2 + · · · + a1 + a0 (mod 3).

The RHS of the congruence is just the sum of the digits! In the problem statement, we’re
given that the sum of the digits is divisible by 3, so we know an−1 + an−2 + · · · + a1 + a0 ≡ 0
(mod 3) =⇒ 10n−1 an−1 + 10n−2 an−2 + · · · + 101 a1 + 100 a0 ≡ n ≡ 0 (mod 3). 

The divisibility rule for 9 follows a similar argument.

Exercise 18.27. Prove the divisibility rule for 9: If the sum of the digits of a number is divisible
by 9, then the original number is also divisible by 9.

Example 18.28 (Divisibility rule for powers of 2). Prove that a number is divisible by 2k
if the number formed by its k rightmost digits are divisible by 2k , where k is a positive
integer.

Solution. You might know the divisibility rule for 4: If the number formed by the last two
digits of a number is divisible by 4, then the original number is divisible by 4. The divisibility
rule for 8 is similar: If the number formed by the last three digits of a number is divisible by 8,

9
Everaise Academy (2020) Math Competitions I

then the original number is divisible by 8. Let’s prove this generalized rule. Suppose we have
an n > k-digit number. We can write out the base 10 expansion as in the previous example,
while noting that 10 = 5 · 2 :

10n−1 an−1 + 10n−2 an−2 + · · · + 101 a1 + 100 a0

≡ 2n−1 5n−1 an−1 + 2n−2 5n−2 an−2 + · · · + 21 51 a1 + 20 50 a0 (mod 2k ).

Since we’re taking this modulo 2k , we see that for any value of l ≥ k, it follows that 2l ≡ 0
(mod 2k ). Therefore, we can chop off a bunch of digits, until we’re left with the k rightmost
digits, because those are the only ones that are not congruent to 0 modulo 2k . Continuing, we
get

2n−1 5n−1 an−1 + 2n−2 5n−2 an−2 + · · · + 21 51 a1 + 20 50 a0

≡ 2k−1 5k−1 ak−1 + 2k−2 5k−2 ak−2 + · · · + 21 51 a1 + 20 50 a0
≡ 10k−1 ak−1 + 10k−2 ak−2 + · · · 101 a1 + 100 a0 (mod 2k ).

Thus, n is congruent to the number formed by the last k digits of n, modulo 2k . 

Exercise 18.29. Prove the divisibility rule for 11: Starting from the leftmost digit, add and
subtract digits in an alternating manner, until you get to the end. If that sum is divisible by
11, then so is the original number.

§18.4 Challenging examples

Example 18.30 (AoPS). Given that p ≥ 7 is a prime number, evaluate

1−1 · 2−1 + 2−1 · 3−1 + 3−1 · 4−1 + · · · + (p − 2)−1 · (p − 1)−1 (mod p).

Solution. First off, we can be sure that all integers between 1 and p − 1, inclusive, have an
inverse, because p is prime. Because modular arithmetic and solving regular equations are so
similar, we might look to regular equations for help. If we were just solving a regular sum, then
we would have

1−1 · 2−1 + 2−1 · 3−1 + 3−1 · 4−1 + · · · + (p − 2)−1 · (p − 1)−1

1 1 1 1 1 1 1 1
= · + · + · + ··· + · .
1 2 2 3 3 4 p−2 p−1

Recall that n1 · n+1

1
= n1 − n+1
1
, so we can get a telescoping sum. Let’s see if we can achieve the
same thing in modular arithmetic. In other words, can we prove that

n−1 · (n + 1)−1 ≡ n−1 − (n + 1)−1 (mod p)?

Multiplying both sides by n(n + 1) lets us cancel out some things, by the definition of inverse.
We do that, to get

n−1 (n + 1)−1 n(n + 1) ≡ (n−1 − (n + 1)−1 )(n(n + 1)) (mod p)

=⇒ 1 ≡ n + 1 − n (mod p) =⇒ 1 ≡ 1 (mod p).

10
Everaise Academy (2020) Math Competitions I

We’ve proven that this fact also holds modulo p. This gives us a nice telescoping sum to work
with! Continuing, we can see that

1−1 · 2−1 ≡ 1−1 − 2−1 ,

2−1 · 3−1 ≡ 2−1 − 3−1 ,
3−1 · 4−1 ≡ 3−1 − 4−1 ,

and so on. Our sum becomes 1−1 − 2−1 + 2−1 − 3−1 + 3−1 − 4−1 + · · · + (p − 2)−1 − (p − 1)−1 ≡
1−1 − (p − 1)−1 ≡ 1 − (−1) ≡ 2 (mod p). 

Example 18.31 (2014 AIME I). The positive integers N and N 2 both end in the same
sequence of four digits abcd when written in base 10, where digit a is not zero. Find the
three-digit number abc.

Solution. If N and N 2 have the same last 4 digits, then we can say that

N2 ≡ N (mod 104 ) =⇒ N (N − 1) ≡ 0 (mod 104 ). (18.1)

We also know that a 6= 0, so N is at least a 4-digit number. Finally, N can’t be something

along the lines of 10000, because that also clearly violates our a 6= 0 condition.
From (5.1), we can see that N (N − 1) must be a multiple of 104 = 24 · 54 . Furthermore, one of
N, N − 1 must be a multiple of 54 but not divisible by 2, and the other one must be a multiple
of 24 but not divisible by 5 (do a proof by contradiction to see why; what happens when one of
N, N − 1 is a multiple of both 2 and 5?).
Let’s use the fact that one of N, N − 1 is an odd multiple of 54 = 625. We can write this
number as 625 + 2 · 625 · k = 625 + 1250k, for some positive integer k.
If N = 625 + 1250k, then N − 1 must be divisible by 16. We can write

N −1≡0 (mod 16) =⇒ 625 + 1250k ≡ 1 (mod 16).

If N − 1 = 625 + 1250k, then N must be divisible by 16, and we can write

N ≡0 (mod 16) =⇒ N − 1 ≡ −1 (mod 16) =⇒ 625 + 1250k ≡ −1 (mod 16).

Combining, we see that we must find a k such that

625 + 1250k ≡ ±1 (mod 16).

Continuing, we have 625 ≡ 1 (mod 16) =⇒ 1250 ≡ 2 · 625 ≡ 2 (mod 16) =⇒ 625 + 1250k ≡
1 + 2k (mod 16). An obvious value of k that would make the congruence equivalent to ±1
modulo 16 would be k = 0. However, we would get N = 625 and N − 1 = 624, which, although
N and N 2 do end in the same 4 digits, does not satisfy our given restriction of a 6= 0.
The next smallest possible value of k would be

k = 7 =⇒ 1 + 2k ≡ −1 (mod 16) =⇒ N − 1 = 9375, N = 9376.

This satisfies our condition on a not being 0. Therefore, our answer is 937 . Indeed, N = 9376
and N 2 = 87909376. 

11
Everaise Academy (2020) Math Competitions I

Example 18.32 (2011 AIME I). Find the number of positive integers m for which there
exist nonnegative integers x0 , x1 , . . . , x2011 such that
2011
X
x0
m = mxk .
k=1

Solution. This problem just spits out a bunch of symbols, so let’s try to make sense of this.
It wants us to find how many positive integer values of m there are such that m to some power
can be written as the sum of 2011 not necessarily distinct powers of m.
Let’s begin by taking both sides modulo m − 1. Note that m ≡ 1 (mod m − 1), so the LHS
reduces nicely to mx0 ≡ 1x0 ≡ 1 (mod m − 1), and the RHS reduces to
2011
X 2011
X
xk
m ≡ 1xk ≡ 2011 (mod m − 1).
k=1 k=1

Setting the LHS congruent to the RHS, we get

1 ≡ 2011 (mod m − 1) =⇒ 2010 ≡ 0 (mod m − 1).

This tells us that m − 1 must be a factor of 2010. Now, we have to make sure that all values of
m, for which m − 1 is a factor of 2010, are achievable. In other words, we’re still not sure if this
condition is our tightest restriction, but let’s try to prove that it is. The smallest possible value
for m is 2, because 2 − 1 = 1 is a factor of 2010. Let’s try to find x0 , x1 , . . . , x2011 such that
2011
X
2x0 = 2xk .
k=1

Note that if we have x0 = 2010, then on the RHS, we can have the sum 20 +20 +21 +22 +· · ·+22009 ,
and it satisfies the equation.
We can do a similar thing for m = 3. On the RHS, we can have the sum 30 + 30 + 30 + 31 +
31 + 32 + 32 + · · · + 31004 + 31004 (check if this sum has exactly 2011 terms), which sums up to
31005 .
In general, we can construct a power of m using 2011 powers of m by having

m0 + m0 + · · · + m0 + m1 + m1 + · · · + m1 + · · · + mk + mk + · · · + mk = mk+1 ,
| {z } | {z } | {z }
m terms m−1 terms m−1 terms

where k is some positive integer (in the case of m = 2, k = 2009, and in the case of m = 3, k =
1004. What matters is that k changes as m changes). Lastly, we need to verify that this sum
has exactly 2011 terms whenever m − 1 is a factor of 2010. Sure enough, we can count the
number of terms. Looking back at our sum, we have

m0 + m0 + · · · + m0 + m1 + m1 + · · · + m1 + · · · + mk + mk + · · · + mk = mk+1 .
| {z } | {z } | {z }
m terms m−1 terms m−1 terms
| {z }
k of these

Therefore, the number of terms is

m + (m − 1) + (m − 1) + · · · + (m − 1) = m + k(m − 1) = m − 1 + 1 + k(m − 1)
| {z }
k of these

12
Everaise Academy (2020) Math Competitions I

= (k + 1)(m − 1) + 1 = 2011 =⇒ (k + 1)(m − 1) = 2010.

Hence, m − 1 must be a factor of 2010, since we have defined k to be an integer. We have

shown that this way of constructing a power of m works for all m with m − 1|2010, so all
that’s left to do is find how many values of m satisfy m − 1|2010. This is simply equal to
τ (2010) = τ (2 · 3 · 5 · 67) = 016 . 

§18.5 Homework Problems

Problem 18.1. [3] How many positive integers n ≤ 2020 are there such that 2n + 1 is divisible
by 3?

Problem 18.2. [3] Find the remainder when 2020 is divided by 13.

Problem 18.3 (Dennis Chen). [4] How many integers x with 1 ≤ x ≤ 100 are there such that
x2 + 8x + 15 is divisible by 10?

Problem 18.4 (AoPS). [4] What is the smallest positive integer a such that a−1 is undefined
(mod 55) and a−1 is also undefined (mod 66)?

Problem 18.5. [4] What is the remainder when

2020
X
n · 20n
n=1

is divided by 19?

Problem 18.6 (2017 AMC 12). [5] Claire adds the degree measures of the interior angles of a
convex polygon and arrives at a sum of 2017. She then discovers that she forgot to include one
angle. What is the degree measure of the forgotten angle?

Problem 18.7 (2010 AIME I). [5] Find the remainder when 9 × 99 × 999 × · · · × |99 {z
· · · 9} is
999 9’s
divided by 1000.

Problem 18.8 (2018 AMC 10B). [5] Let a1 , a2 , . . . , a2018 be a strictly increasing sequence of
positive integers such that
a1 + a2 + · · · + a2018 = 20182018 .
What is the remainder when a31 + a32 + · · · + a32018 is divided by 6?

Problem 18.9 (2017 AIME II). [9] Find the number of positive integers n less than 2017 such
that
n2 n3 n4 n5 n6
1+n+ + + + +
2! 3! 4! 5! 6!
is an integer.

Problem 18.10 (2003 AMC 10). [7] Let n be a 5-digit number, and let q and r be the quotient
and the remainder, respectively, when n is divided by 100. For how many values of n is q + r
divisible by 11?
Provide a detailed solution, or explain what progress you made, even if you didn’t
get an answer.

13
19 Chinese Remainder Theorem, Fermat’s
Little Theorem, and Euler’s Theorem
In this handout, we’ll go over some more advanced techniques relating to modular arithmetic.

§19.1 The Chinese Remainder Theorem

You might remember being introduced to the methods of substitution and elimination to solve
a system of equations. Analogously, the Chinese Remainder Theorem is used to solve a system
of linear congruences.

Example 19.1. Find the smallest positive integer solution for x given that it satisfies
2x ≡ 7 (mod 11) and 5x ≡ 1 (mod 3).

Solution. We should begin by solving each congruence for x. The first congruence requires us
to find the inverse of 2 modulo 11. Since 2 · 6 ≡ 12 ≡ 1 (mod 11), we know that 2−1 ≡ 6
(mod 11). It follows that

2x ≡ 7 (mod 11) =⇒ x ≡ 7 · 2−1 ≡ 7 · 6 ≡ 42 ≡ 9 (mod 11).

For the second congruence, we can already reduce it down:

5x ≡ 1 (mod 3) =⇒ 2x ≡ 1 (mod 3) =⇒ x ≡ 2−1 (mod 3).

We know that 2 ≡ −1 (mod 3). Since (−1)(−1) ≡ 1 (mod 3), then we can replace the −1s with
2s, and get that 2 is its own inverse modulo 3. Thus,

x ≡ 2−1 ≡ 2 (mod 3).

Then, we list out the first few positive integers that are congruent to 9 modulo 11: 9, 20, 31, ...
We see that 20 is also congruent to 2 modulo 3, so our answer is 20 . 

That was pretty unsatisfying. What would we do if we had bigger numbers? We just kind of
guessed and checked our way through the problem, and it worked, simply because the numbers
were manageable.
To generalize the idea of constructing a solution to a system of modular equations, we intro-
duce the Chinese Remainder Theorem (CRT):

Theorem 19.2 (Chinese Remainder Theorem). If b1 , b2 , . . . , br are pairwise relatively prime

positive integers, then the system of congruences

x ≡ a1 (mod b1 )
x ≡ a2 (mod b2 )
..
.
x ≡ ar (mod br )

1
Everaise Academy (2020) Math Competitions I

has a unique solution for x modulo b1 b2 · · · br .

In particular, the solution for x is constructed as follows: Let B = b1 b2 · · · br and let
Bi = B/bi , for all i, with 1 ≤ i ≤ r. Let ci ≡ Bi−1 (mod bi ). In other words, ci is the
inverse of Bi , mod bi . Then,
r
X
x≡ ai Bi ci ≡ a1 B1 c1 + a2 B2 c2 + · · · + ar Br cr (mod B).
i=1

Proof. We prove that the solution for x is unique. We first prove the following general statement:
If

x≡y (mod b1 )
x≡y (mod b2 )
..
.
x≡y (mod br ),

then x ≡ y (mod lcm (b1 , b2 , . . . , br )).

We can rearrange the congruences to get

x−y ≡0 (mod b1 )
x−y ≡0 (mod b2 )
..
.
x−y ≡0 (mod br ).

It follows that b1 |(x − y), b2 |(x − y), . . . , br |(x − y). Recall that if for positive integers a, b, and
c, we have a|c and b|c, then lcm (a, b)|c.
Therefore,

lcm (b1 , b2 , . . . , br )|(x − y) =⇒ x ≡ y (mod lcm (b1 , b2 , . . . , br )).

An immediate corollary to this is when b1 , b2 , . . . , br are pairwise relatively prime. This means
that lcm (b1 , b2 , . . . , br ) = b1 b2 · · · br =⇒ x ≡ y (mod b1 b2 · · · br ).
Now, we turn our eyes back to the CRT. Suppose that x1 and x2 are both solutions to the
following system of congruences for x :

x ≡ a1 (mod b1 )
x ≡ a2 (mod b2 )
..
.
x ≡ ar (mod br ),

where b1 , b2 , . . . , br are pairwise relatively prime. It follows that x1 ≡ x2 ≡ ai (mod bi ) for all i
between 1 and r, inclusive. From what we proved earlier, we have

x1 ≡ x2 (mod b1 b2 · · · br ),

so the solution of the system is unique modulo b1 b2 · · · br . 

Existence of the solution is already given by our construction in the theorem statement. We’ll
verify it now.

2
Everaise Academy (2020) Math Competitions I

Example 19.3. Verify that the given solution for x in the theorem satisfies all of the
congruences.

Solution. To get an idea of what we’re working with, let’s take the sum modulo b1 . Since
B2 = b1 b3 b4 · · · br , then we know that B2 ≡ 0 (mod b1 ). Also, B3 = b1 b2 b4 b5 · · · br , then we
know that B3 ≡ 0 (mod b1 ) as well. In fact, Bj ≡ 0 (mod b1 ) for all j 6= 1, because b1 will
always be a factor. Thus,

a1 B1 c1 + a2 B2 c2 + · · · + ar Br cr ≡ a1 B1 c1 (mod b1 ).

By the definition of ci , we know that c1 is the inverse of B1 modulo b1 . In other words, we can
write c1 B1 ≡ 1 (mod b1 ). Therefore,

a1 B1 c1 ≡ a1 · 1 ≡ a1 (mod b1 ),

so x satisfies the first congruence. Now, we prove that x ≡ ak (mod bk ), for all k between 2
and r, inclusive.
Notice that bk |Bj for all j 6= k. Hence,

a1 B1 c1 + a2 B2 c2 + · · · + ar Br cr ≡ ak Bk ck (mod bk ).

Recall that Bk ck ≡ 1 (mod bk ) for all k, so

ak Bk ck ≡ ak (mod bk ),

as desired. 

Remark. Don’t memorize this long statement! It’s more important to know how the formula
works. The most important thing for you is to see how we used inverses and divisibility to
force a sum to satisfy many different congruences at once.

Example 19.4. Solve the system of congruences:

x≡1 (mod 3)
x≡3 (mod 5)
x≡5 (mod 7)

Solution. We use the Chinese Remainder Theorem to solve this. Afterwards, we will show
another way to solve systems of congruences, called back-substitution.
Be careful! In order to use CRT, then we have to make sure that all of our moduli are pairwise
congruent. For this problem, we’re good, since we’re using only primes. Now, we compute

B = 3 · 5 · 7 = 105 =⇒ B1 = 35, B2 = 21, B3 = 15.

We work term by term. We must make sure that the first term in the sum is congruent to 1
mod 3. Therefore,

B1 c1 ≡ 35 · c1 ≡ 1 (mod 3) =⇒ c1 = 2.

3
Everaise Academy (2020) Math Competitions I

Similarly,

B2 c2 ≡ 21 · c2 ≡ 1 (mod 5) =⇒ c2 = 1,
B3 c3 ≡ 15 · c3 ≡ 1 (mod 7) =⇒ c3 = 1.

Thus, x ≡ 1 · 35 · 2 + 3 · 21 · 1 + 5 · 15 · 1 ≡ 208 ≡ 103 (mod 105) . 

Solution. For this solution, we build up using “back-substitution”. Since x ≡ 1 (mod 3),
then recall that we can write x = 3q + 1, for some integer q. Substituting this into our second
congruence, we get

x≡3 (mod 5) =⇒ 3q + 1 ≡ 3 (mod 5) =⇒ q ≡ 4 (mod 5).

Thus, we write q = 5r + 4, for some integer r. Back-substituting, we have x = 3q + 1 =

3(5r + 4) + 1 = 15r + 13. Now, we use our last congruence:

x≡5 (mod 7) =⇒ 15r + 13 ≡ 5 (mod 7) =⇒ r ≡ 6 (mod 7) =⇒ r = 7s + 6,

for some integer s. Back-substituting again, we arrive at

x = 15r + 13 = 15(7s + 6) + 13 = 105s + 103 =⇒ x ≡ 103 (mod 105) .

Solution. There is a slicker solution, that actually doesn’t involve any advanced method. Ob-
serve that we can rewrite the congruences as

x ≡ −2 (mod 3)
x ≡ −2 (mod 5)
x ≡ −2 (mod 7).

Using the ideas discussed in the proof of the uniqueness of the solution in CRT, we have that
x + 2 is a multiple of 3, 5, and 7. It follows that

x+2≡0 (mod 3 · 5 · 7) =⇒ x ≡ 103 (mod 105) .

Example 19.5. The students in a classroom are arranged in a rectangular formation. When
they are arranged in 9 rows, there are 3 positions unoccupied. When they are arranged in
11 rows, there is 1 position unoccupied. How many students are in the class if the class
size is between 100 and 200?

Solution. Let n be the number of students in the class. We can write a system of modular
congruences to model this problem.

n ≡ −3 (mod 9)
n ≡ −1 (mod 11)

This is equivalent to the following:

n≡6 (mod 9)
n ≡ 10 (mod 11)

4
Everaise Academy (2020) Math Competitions I

Using the Euclidean Algorithm, we find that the inverse of 9 mod 11 is 5 and the inverse
of 11 mod 9 is 5. Therefore, using our earlier summation formula, we have n = 6(11)(5) +
10(9)(5) = 780 ≡ 87 (mod 99). The problem states that 100 ≤ n ≤ 200, so the answer is
87 + 99 = 186 . 

Example 19.6. Show that there exists a multiple of 2013 whose last four digits are 2014.

Solution. At first glance, this number, which we can call X, seems really big. However, we
don’t actually need to find this number, we simply need to show that it exists.
Recall that the core idea of CRT is that a system of congruences with pairwise coprime moduli
always has a solution. Once we construct a system of congruences that satisfies the problem’s
constraints, we’re done, by CRT. From the first part of the problem statement, we have

X≡0 (mod 2013).

We’re also given that the last four digits of X are 2014, so we can write

X ≡ 2014 (mod 10000).

Since gcd (2013, 10000) = 1, CRT applies, and it tells us that there exists a solution. 

Example 19.7. Find the last three digits of 1242020 + 2020 without a calculator.

Solution. For the sake of brevity, let’s say that N = 1242020 + 2020. Finding the last three
digits of N is equivalent to finding the remainder when N is divided by 1000. In other words,
we want to reduce N down modulo 1000.
It would be really hard to do this without a calculator, so we look to a new strategy. We
want to split up the problem into smaller moduli so that we can use CRT. This means that we
need to find b1 and b2 such that gcd (b1 , b2 ) = 1 and b1 b2 = 1000. The only pair that works is
b1 = 8, b2 = 125.
Our new idea is this: We want to solve the following system of congruences for N :

N ≡ l1 (mod 8)
N ≡ l2 (mod 125),

where l1 is the least residue of N modulo 8 and l2 is the least residue of N modulo 125. To
proceed, we need to find l1 and l2 . Since 1242020 = 42020 312020 is clearly divisible by 8, we have

1242020 ≡ 0 (mod 8) =⇒ 1242020 + 2020 ≡ 2020 ≡ 4 (mod 8) =⇒ l1 = 4.

For modulo 125, we use the fact that 124 ≡ −1 (mod 125), to get

1242020 ≡ (−1)2020 ≡ 1 =⇒ N ≡ 1 + 2020 ≡ 21 (mod 125) =⇒ l2 = 21.

Thus, our system of congruences that we want to solve is

N ≡4 (mod 8)
N ≡ 21 (mod 125).

5
Everaise Academy (2020) Math Competitions I

Since we only have two congruences, we back-substitute instead of breaking out the CRT sum
formula. From the first congruence, we can write x = 8q + 4, for some integer q. Then, our
second congruence becomes 8q + 4 ≡ 21 (mod 125) =⇒ 8q ≡ 17 (mod 125). Ooh, this is a bit
tough to solve, since our modulus is so big. Let’s go the other way around and start with the
second congruence.
We have N = 125q + 21, for some integer q. Then, we substitute this in for N in our first
congruence, to get

125q + 21 ≡ 4 (mod 8) =⇒ 5q + 5 ≡ 4 (mod 8) =⇒ q ≡ 3 (mod 8).

We can write q = 8r + 3, for some integer r, which gives us

N = 125q + 21 = 125(8r + 3) + 21 = 1000r + 375 + 21 = 1000r + 396.

Therefore, the last three digits of N are 396 . 

Remark. No one likes big numbers. We like working with small moduli, so we reduced the
problem down until the size of the moduli was manageable.

If the number didn’t involve 124 to some power and instead involved, say, 97 to some power,
then we would have had a much harder time reducing it modulo 125. Do not worry! FLT and
Euler’s Theorem are both very helpful in this regard.

§19.1.1 Exercises for this section

Exercise 19.8. Solve the following system of congruences for x :

x ≡ 19 (mod 21)
x ≡ 15 (mod 20)
x ≡ 11 (mod 19).

Exercise 19.9. Show that there exists a multiple of 2019 that is 1 more than a multiple of 2020.
Exercise 19.10. Find the smallest positive integer solution for x if

2x ≡ 4 (mod 6)
3x ≡ 6 (mod 25)
3x ≡ 8 (mod 21)
5x ≡ 10 (mod 11).

Hint: What needs to be true in order for CRT to work?

§19.2 Fermat’s Little Theorem

The remaining sections are very important, because they contain methods to reduce large
numbers. These theorems are worth memorizing. We start off this section with another theorem:
Wilson’s Theorem. This will help us in our proof of FLT.

6
Everaise Academy (2020) Math Competitions I

Example 19.11. Find the remainder when 10! is divided by 11.

Solution. Don’t bash it out! Let’s use inverses to solve this. Note that
2−1 ≡ 6 (mod 11)
3−1 ≡ 4 (mod 11)
−1
5 ≡9 (mod 11)
−1
7 ≡8 (mod 11).
Thus,
10! ≡ 1 · (2 · 6)(3 · 4)(5 · 9)(7 · 8) · 10 ≡ 10 (mod 11),

and the remainder is 10 . 

Theorem 19.12 (Wilson’s Theorem). If p is a prime number, then

(p − 1)! ≡ −1 (mod p).

Proof. Similar to the example above, we use inverses. In Example 5.13 from the previous
handout, we proved that the only time when a is its own inverse modulo p, where p is a prime,
is when a ≡ −1 or 1 (mod p). For this proof, we assume that p ≥ 5 (it is easy to see that p = 2
and p = 3 work).
Remark that (p − 1)! = 1 · 2 · 3 · · · (p − 3)(p − 2)(p − 1) and consider the product 2 · 3 · · · (p −
3)(p − 2). Recall that the inverse of a is unique modulo p, which implies that we can group the
set of numbers {2, 3, . . . , p − 3, p − 2} into pairs of inverses modulo p.
For instance, in the previous example, the pairs would be (2, 6), (3, 4), (5, 9), and (7, 8). We
can be sure that all our pairs are disjoint, meaning no two pairs share a common number,
because of the uniqueness property.
Why did we pair up numbers like this? So that we can use inverses! The product of the two
numbers in every pair is congruent to 1 modulo p, so we have
2 · 3 · · · (p − 3)(p − 2) ≡ 1 (mod p) =⇒ (p − 1)! ≡ p − 1 ≡ −1 (mod p).


Now, we can move on to FLT. We first illustrate how Fermat’s Little Theorem is useful with
an example.

Example 19.13 (MATHCOUNTS). Find the remainder when 1112 is divided by 13.

Solution. We want to find the least residue of 1112 modulo 13. Note that 11 ≡ −2 (mod 13),
so we have
1112 ≡ (−2)12 ≡ 26 · 26 ≡ 64 · 64 ≡ (−1)(−1) ≡ 1 (mod 13),

so our answer is 1 . 

That took a bit of work, and without prior knowledge of modular arithmetic, would’ve taken
longer. However, FLT completely trivializes this problem.

7
Everaise Academy (2020) Math Competitions I

Theorem 19.14 (Fermat’s Little Theorem). If p is a prime number, and a is a positive

integer with p - a, then

ap−1 ≡ 1 (mod p).

We present two proofs of FLT: one using Wilson’s theorem, and one using induction, in hopes
that both will be beneficial.

Proof. Suppose that p is a prime number a is some positive integer such that p - a. Then, the
elements in the following set:

S = {a, 2a, 3a, . . . , (p − 2)a, (p − 1)a}

are pairwise incongruent and hit all possible residues except for 0 modulo p.
We first prove that they are pairwise in congruent. Suppose to the contrary that there exist
i, j such that 1 ≤ i 6= j ≤ p − 1 and ia ≡ ja (mod p). Rearranging the congruence gives us

ia − ja ≡ 0 (mod p) =⇒ (i − j)a ≡ 0 (mod p).

It follows that i−j must be a multiple of p, but i−j ≤ p−1, so we’ve arrived at a contradiction.
Note that the possible residues modulo p are {0, 1, 2, . . . , p−2, p−1}. S contains p−1 elements,
which are pairwise incongruent to each other, and none of the elements are congruent to 0
(mod p). This means that the elements in S taken modulo p are a permutation of {1, 2, . . . , p −
2, p − 1}. Thus, if P is the product of all the elements in S, then

P ≡ (p − 1)! (mod p)
=⇒ a · 2a · 3a · · · (p − 1)a ≡ −1 (mod p)
p−1
=⇒ (p − 1)!a ≡ −1 (mod p)
p−1
=⇒ − a ≡ −1 (mod p) =⇒ ap−1 ≡ 1 (mod p).

Proof. Our second proof uses induction. Note that

ap−1 ≡ 1 (mod p) ⇐⇒ ap ≡ a (mod p).

Thus, we want to prove that ap ≡ a (mod p) for all a 6≡ 0 (mod p). It suffices to prove the
congruence holds for all a in the range 1 ≤ a ≤ p − 1.

For our base case, we have a = 1. Indeed, 1p ≡ 1 (mod p).

For the inductive step, we assume that ap ≡ a (mod p) for some a between 1 and p − 2.
We seek to prove that (a + 1)p ≡ a + 1 (mod p) as well. Indeed, by the binomial theorem,
     
p p p p−1 p p−2 p
(a + 1) ≡ a + a + a + ··· + a1 + 1 ≡ ap + 1 ≡ a + 1 (mod p).
1 2 p−1

Thus, ap ≡ a (mod p) for all a between 1 and p − 1, inclusive. If instead we have a > p, then
a will still be congruent to some integer between 1 and p − 1, because p - a, so our conclusion
holds for all positive a that is not divisible by p. 

8
Everaise Academy (2020) Math Competitions I

Example 19.15 (2008 AMC 12). Let k = 20082 + 22008 . What is the units digit of k 2 + 2k ?

Solution. If we want the units digit, we think of the problem modulo 10. However, reducing
the 22008 looks nasty, and FLT requires the modulus to be prime, so we split up the problem
into mod 2 and mod 5, then use CRT. We find k 2 + 2k modulo 2 first. Note that k is even, so

k 2 + 2k ≡ 0 + 0 ≡ 0 (mod 2).

For modulo 5, we use FLT to reduce k. In particular, it tells us that 25−1 ≡ 1 (mod 5). Therefore,

502
20082 + 22008 ≡ 32 + 24 ≡9+1≡0 (mod 5) =⇒ k 2 ≡ 0 (mod 5).

To find 2k modulo 5, we need to find k modulo 4. To see why, simply let k = 4q + r, where r is
the remainder when k is divided by 4. Then, we have

2k ≡ 24q+r ≡ 2| 4 24 2{z
4
· · · 2}4 ·2r ≡ 2r (mod 5).
q of these

Note that 4|20082 and 4|22008 , so k ≡ 0 (mod 4) =⇒ r = 0 =⇒ 2k ≡ 2r ≡ 1 (mod 5) =⇒

k 2 + 2k ≡ 1 (mod 5). Thus, we have the following system of congruences:

k 2 + 2k ≡ 0 (mod 2)
2 k
k +2 ≡1 (mod 5).

It follows that k 2 + 2k ≡ 6 (mod 10), so the units digit is 6 . 

Remark. If a and n are relatively prime positive integers such that ax ≡ 1 (mod n) and
y > x, then ay ≡ ar (mod n), where r is the remainder when y is divided by x. We will
use this idea a lot in the next section.

Exercises for this section:

Exercise 19.16 (AoPS). Find 220 + 330 + 440 + 550 + 660 (mod 7).
Exercise 19.17. Find all n such that 5n − 4n is divisible by 61.
Exercise 19.18. The order of a modulo n is the smallest positive integer r such that ar ≡ 1
(mod n). If n is a prime and a is a positive integer that does not divide n, then is the order of
a always n − 1? Find two values of a and n for which the order is n − 1 and two values of when
it is not.
Exercise 19.19. Prove the converse of Wilson’s Theorem: If (p − 1)! ≡ −1 (mod p), then p is a
prime number. Hint: Do a proof by contradiction: assume that p is composite.

§19.3 Euler’s Theorem

What if the modulus isn’t prime? FLT only deals with prime moduli, after all. Euler’s theorem
generalizes this to all moduli.

9
Everaise Academy (2020) Math Competitions I

Theorem 19.20 (Euler’s Theorem). If gcd (a, n) = 1, then

aφ(n) ≡ 1 (mod n),

where φ(n) is the number of positive integers less than or equal to n that are relatively
prime to n.

Proof. This proof will be very similar to the proof for FLT. We define a reduced residue system
(RRS) modulo n to be a set of φ(n) numbers such that every integer relatively prime to n is
congruent to some element in the set. For example, the following sets would be RRS’s modulo
10:

{1, 3, 7, 9}, {11, 13, 17, 19}, {−19, 1113, 917, 9}

and these sets would not be RRS’s modulo 10:

{2, 4, 6, 9}, {1, 2, 3, 4, . . . , 9, 10}, {1, 3, 5, 7}.

Recall that in the proof of FLT, we showed that {a, 2a, 3a, . . . , (p − 2)a, (p − 1)a} was a permu-
tation of {1, 2, 3, . . . , p − 2, p − 1}, modulo p. It turns out that the same thing holds for a RRS!
Let

S = {r1 , r2 , . . . , rφ(n)−1 , rφ(n) }

be the set of positive integers less than n that are relatively prime to n. Note that it is a RRS,
modulo n. Then, for any a such that gcd(a, n) = 1, the set

S 0 = {ar1 , ar2 , . . . , arφ(n)−1 , arφ(n) }

is also a RRS. To show this, we want to show that all of the elements in S 0 are relatively prime
to n, and that the elements are pairwise relatively prime. We will use proof by contradiction to
prove both of these facts.

To prove that all of the elements are in S 0 are relatively prime to n, we suppose to the contrary
that there exists i between 1 and φ(n), inclusive, such that gcd(ari , n) > 1. This tells us that
either gcd(a, n) > 1 or gcd(ri , n) > 1. We defined a to be relatively prime to n, and ri was also
defined to be part of a reduced residue system, so we know that gcd(ri , n) = 1. Thus, we have
arrived at a contradiction.
To prove that all the elements in S 0 are pairwise relatively prime, we suppose to the contrary
that there exist i, j, where 1 ≤ i 6= j ≤ φ(n), such that

ari ≡ arj (mod n) =⇒ a(ri − rj ) ≡ 0 (mod n).

Hence, one of a, ri − rj must be a multiple of n. However, gcd(a, n) = 1, so we must have ri − rj

be a multiple of n. However, we defined S to only contain elements that are positive and less
than n, so ri − rj < n. Thus, we have arrived at another contradiction, meaning that we’ve
proved that

{ar1 , ar2 , . . . , arφ(n)−1 , arφ(n) }

10
Everaise Academy (2020) Math Competitions I

is a RRS. To finish, we take the product of all elements in S and all elements in S 0 , and set
them to be congruent to each other:

r1 r2 · · · rφ(n)−1 rφ(n) ≡ ar1 ar2 · · · arφ(n)−1 arφ(n) (mod n)

φ(n)
=⇒ r1 r2 · · · rφ(n)−1 rφ(n) ≡ r1 r2 · · · rφ(n)−1 rφ(n) a (mod n).

We can divide both sides by r1 r2 · · · rφ(n)−1 rφ(n) , since we know that gcd(r1 r2 · · · rφ(n)−1 rφ(n) , n) =
1. Doing so gives us

aφ(n) ≡ 1 (mod n).

Anyways, recall that if n = p1 k1 p2 k2 . . . pm km , then

    
1 1 1
φ(n) = n 1 − 1− ··· 1 − .
p1 p2 pm

Note that if n is a prime number, then φ(n) = n − 1. By Euler’s Theorem,

aφ(n) ≡ an−1 ≡ 1 (mod n),

which is just FLT! Hence, we say that Euler’s Theorem is a generalization of FLT. Euler’s The-
orem also gives us a reliable way to find inverses, besides using the Euclidean Algorithm:

Proposition 19.21. If gcd (a, n) = 1, then

a−1 ≡ aφ(n)−1 (mod n).

Proof. By Euler’s, we know that aφ(n) ≡ 1 (mod n). Recall that a · a−1 ≡ 1 (mod n), so it
follows that a−1 ≡ aφ(n)−1 (mod n). 

Now let’s do some examples to better understand how to use Euler’s Theorem.

2021
Example 19.22. Find the remainder when 20192020 is divided by 1000.

Solution. As usual, we split the problem up into mod 8 and mod 125 to downsize our numbers.
2021
Let N = 20192020 , for shortness. First, we want to find N modulo 8. By Euler’s, 2019φ(8) ≡ 1
1
(mod 8). Computing, we have φ(8) = 8 1 − 2 = 4 =⇒ 20194 ≡ 1 (mod 8). It follows that

2021
20192020 ≡1 (mod 8),

because 20202021 is divisible by 4.

1


Now for modulo 125. We use Euler’s Theorem again, and note that φ(125) = 125 1 − 5 =
100, to get

2019φ(125) ≡ 1 (mod 125) =⇒ 2019100 ≡ 1 (mod 125).

Thus, our next step is to find 20202021 modulo 100. We don’t have to do Euler’s again, because
it’s easy to see that

20202021 = 2022021 · 102021 ,

11
Everaise Academy (2020) Math Competitions I
2021
so 20202021 ≡ 0 (mod 100) =⇒ 20192020 ≡ 1 (mod 125). Combining, we have

N ≡1 (mod 8)
N ≡1 (mod 125),

implying that N ≡ 1 (mod 1000), so the remainder is 1 . 

Example 19.23. Find the inverse of 5 modulo 36 using Euler’s Theorem.

Solution. We use the fact that a−1 ≡ aφ(n)−1 (mod 36) when gcd (a, n) = 1. Here, we have
gcd (5, 36) = 1, so we’re good to go. Plugging in the numbers, we see that

5−1 ≡ 5φ(36)−1 (mod 36),

where φ(36) = 36 1 − 21 1 − 13 = 12, so we just need to find 5φ(36)−1 = 511 mod 36. This

isn’t too bad as long as we keep the number reasonably small:

52 ≡ 25 (mod 36)
3
5 ≡ 125 ≡ 17 (mod 36)
6 3 3
5 ≡ 5 5 ≡ 17 · 17 ≡ 289 ≡ 1 (mod 36)
11 6 3 2
5 ≡ 5 5 5 ≡ 1 · 17 · 25 ≡ 425 ≡ 29 (mod 36).

Hence, 5−1 ≡ 29 (mod 36) . 

Let’s do a slightly harder example to finish off this section.

Example 19.24 (2011 AMC 10). What is the hundreds digit of 20112011 ?

Solution. We essentially want to find 20112011 mod 1000. Before we split up the problem
into mod 8 and mod 125, notice that we can already reduce this mod 100, since 2011 ≡ 11
(mod 1000). Now, we split up. We consider 112011 modulo 8 first. We have 11 ≡ 3 (mod 8), so

112011 ≡ (3)2011 (mod 8).

By Euler’s, we know that φ(8) = 8 1 − 12 = 4, so 34 ≡ 1 (mod 8). We know that 2011 ≡ 3

(mod 4). Hence,

32011 ≡ 33 ≡ 27 ≡ 3 (mod 8).

For modulo 125, we use Euler’s again to get that 11φ(125) ≡ 1 (mod 125), where φ(125) =
125 1 − 51 = 100. It follows that

112011 ≡ 1111 (mod 25).

Then, we bash it out:

112 ≡ 121 ≡ −4 (mod 125)

4 2 2
11 ≡ 11 11 ≡ (−4)(−4) ≡ 16 (mod 125)
8 4 4
11 ≡ 11 11 ≡ 16 · 16 ≡ 256 ≡ 6 (mod 125)
11 8 2
11 ≡ 11 11 · 11 ≡ 6 · (−4) · 11 ≡ −264 ≡ −14 ≡ 111 (mod 125).

12
Everaise Academy (2020) Math Competitions I

Thus, our system of congruences is

112011 ≡ 3 (mod 8)
2011
11 ≡ 111 (mod 125).

We let 112011 = 125q + 111 for some integer q, then substitute it into the first congruence:

112011 ≡ 125q + 111 ≡ 3 (mod 8) =⇒ 5q ≡ 4 (mod 8) =⇒ q ≡ 4 (mod 8).

Thus, q = 8r + 4 for some integer r, which implies that

112011 = 125q + 111 = 125(8r + 4) + 111 = 1000r + 500 + 111 =⇒ 112011 ≡ 611 (mod 1000).

Therefore, hundreds digit is 6 . 

Remark. Don’t forget your roots! We started off our modular arithmetic portion of this
week by using the binomial theorem to solve the following problem:

What is the remainder when 62020 is divided by 5?

Can we do something similar for the previous example?

The point is to not get sucked too deep in the rabbit hole that is modular arithmetic. It’s
easy to get caught up in all of the calculations and numbers that are involved. If you are
doing a contest and modular arithmetic seems too bashy/time consuming, think of other
ways to approach the problem!

§19.3.1 Exercises for this section

2122
Exercise 19.25. Find the last two digits of 1920 .
Exercise 19.26. Find the remainder when 1234567890 is divided by 999.
Exercise 19.27. Find the smallest positive value of x such that 2x ≡ 2123 (mod 65).
Exercise 19.28. Show that aφ(b) + bφ(a) ≡ 1 (mod ab) if a and b are relatively prime positive
integers. Verify that this congruence holds when a = 15 and b = 14.

§19.4 Challenging examples

Example 19.29. Find the last two non-zero digits of 2020!.

Solution. First, we must find the number of trailing zeroes in 2020!. The number of trailing
zeroes is equal to the greatest power of 10 that divides 2020!. Since there are more powers of
2 than powers of 5 among the integers from 1 to 2020, the highest power of 5 dividing 2020!
should equal the highest power of 10 that divides 2020!.
Here, we can use Legendre’s formula for p = 5.
∞ j
X nk
f (n!) = .
5k
k=1

This will give us the number of trailing zeroes in the expression n!.

13
Everaise Academy (2020) Math Competitions I

Remark. In general, for a prime p and n, Legendre’s formula gives that the number of
factors of p in n! is
∞  
X n
.
pk
k=1
j k
You can think of each term pnk as counting the number of integers less than n divisible
by pk . It’s easy to see that the number of factors of p in n! is simply the number of integers
less than n divisible by p, plus the number of integers less than n divisible by p2 , and
so on; this is because if an integer between 1 and n inclusive has highest power k of the
prime factor p, then we count it once each time we are counting numbers divisible by pi
for 1 ≤ i ≤ k, meaning we count all k factors in total. This is the reasoning for Legendre’s
formula.

Hence, we find that the number of trailing zeroes in 2020! is

       
2020 2020 2020 2020
+ + + = 404 + 80 + 16 + 3 = 503.
5 25 125 625
2020!
Therefore, the expression 10 503 represents 2020! with no trailing zeroes. In order to find the last
2020!
two digits of 10503 , we take the expression mod 100. However, we note that it is much easier
2020!
to find 10503 in mod 4 and mod 25 and then combine the equations using Chinese Remainder

Theorem.
2020! 2020! 2020! z
We will first consider 10 503 (mod 25). Let z = 5503 . Note that 10503 is equivalent to 2503 .

Additionally, note that z is the same as saying all the factors of 2020! that are multiples of 5
have been divided through by their highest power of 5. I.e.
z = (1 · 2 · 3 · 4 · 1) · (6 · 7 · 8 · 9 · 2) . . . (2016 · 2017 · 2018 · 2019 · 404)

We will split these factors into sets based on the highest power of 5 they were divisible by. For
example, since 75 = 3 · 52 , 3 belongs in the 52 set
50 : {1, 2, 3, 4, 6, 7, 8, 9, . . . , 2016, 2017, 2018, 2019}
51 : {1, 2, 3, 4, 6, 7, 8, 9, . . . 401, 402, 403, 403}
52 : {1, 2, 3, 4, 6, 7, 8, 9, . . . 76, 77, 78, 79}
53 : {1, 2, 3, 4, 6, 7, 8, 9, . . . 16}
54 : {1, 2, 3}

Interestingly, 1·2·3·4 ≡ −1 (mod 25) and 6·7·8·9 ≡ −1 (mod 25). In fact, every 4 consecutive
numbers starting from 1 has a product of -1 mod 25. We can prove this by representing the
four consecutive numbers as 5z + 1, 5z + 2, 5z + 3, and 5z + 4 and expanding and taking mod
25. Therefore, let us group them into 4 consecutive numbers.
50 : {(1, 2, 3, 4), (6, 7, 8, 9), . . . , (2016, 2017, 2018, 2019)}
51 : {(1, 2, 3, 4), (6, 7, 8, 9), . . . (401, 402, 403, 404)}
52 : {(1, 2, 3, 4), (6, 7, 8, 9), . . . (76, 77, 78, 79)}
53 : {(1, 2, 3, 4), (6, 7, 8, 9), . . . 16}
54 : {1, 2, 3}

14
Everaise Academy (2020) Math Competitions I

Since we only care about the product mod 25, we can substitute each group with −1. The 50
set has 404 groups, so its product is (−1)404 = 1. Similarly, the product of the sets denoted by
51 , 52 , 53 , and 54 are −1, 1, −16, and 6 respectively. Taking the product of each of these sub-
products, we find that z ≡ 96 ≡ 21 (mod 25). We also have that 2−503 ≡ 2−3 ≡ 22 (mod 25)
2020!
by Euler’s theorem. Therefore, 10 503 ≡ 21 · 22 ≡ 12 (mod 25).

2020!
Thankfully the process to find 10503 mod 4 is much quicker. Simply note that because there are

far more powers of 2 than powers of 5 in the prime factorization of 2020!, 2505 |2020!. Therefore,
2020! 2020!
4| 10503 , so 10503 ≡ 0 (mod 4).

Finally, we have arrived at two congruences

2020!
≡ 12 (mod 25)
10503
2020!
≡ 0 (mod 4),
10503
2020!
which we can combine using the Chinese Remainder Theorem to get 10 503 ≡ 12 (mod 100)

This example shows how the Chinese Remainder Theorem can be used in order to break down
difficult problems in simpler, more managaeble ones. 

9
Example 19.30. Find the last three digits of 78 .

Solution. By this point, you should automatically think of modulo 1000 when it asks about
the last three digits. To reduce the size of the numbers, we split up the problem into modulo 8
and modulo 125. Nothing new so far. Let’s do modulo 8 first. We have 7 ≡ −1 (mod 8), which
implies that
9 9
78 ≡ (−1)8 ≡ 1 (mod 8).
1
Now for modulo 125. By Euler’s, we have 7φ(125) ≡ 1 (mod 125), where φ(125) = 125 1 −


5 =
100.
Now, we need to find the remainder when 89 is divided by 100. We can’t use Euler’s right
away, because gcd (8, 100) 6= 1, so we split it up into modulo 4 and modulo 25. For modulo 4,
we have

89 ≡ 0 (mod 4).

By Euler’s, we have 8φ(25) ≡ 1 (mod 25), where φ(25) = 25 1 − 51 = 20. This does not help us

at all, because we
want to find 89 (mod 25), but Euler’s tells us about 820 instead. To proceed,
9
we write 89 = 23 = 227 . By changing this exponent to have a smaller base, we can get Euler’s
to tell us something useful now. Using the theorem again, we have

2φ(25) ≡ 220 ≡ 1 (mod 25) =⇒ 227 ≡ 27 ≡ 3 (mod 25).

Hence,

89 ≡ 0 (mod 4)
9
8 ≡3 (mod 25),

so 89 ≡ 28 (mod 100). It follows that

9
78 ≡ 728 (mod 125).

15
Everaise Academy (2020) Math Competitions I

Now is the hard part. There’s really no nice way to find this, except for to bash. However,
there are different methods of bashing, the most efficient of which is binary exponentiation.
Suppose you want to find ax (mod n), for some large x. We write x in its base two representation,
and repeatedly square a, then reduce it mod n. For this example, we have 28 = 16 + 8 + 4 and

71 ≡ 7 (mod 125)
2
7 ≡ 49 (mod 125)
2
74 ≡ 72 ≡ 492 ≡ 2401 ≡ 26 (mod 125)



2
78 ≡ 74 ≡ 262 ≡ 676 ≡ 51 (mod 125)

2
716 ≡ 78 ≡ 512 ≡ 2601 ≡ 101 (mod 125).

Therefore, 728 ≡ 716 78 74 ≡ 101 · 51 · 26 ≡ 5151 · 26 ≡ 26 · 26 ≡ 51 (mod 125). We’re left with
one final system of congruences:
9
78 ≡ 1 (mod 8)
89
7 ≡ 51 (mod 125).
9
We can write 78 = 125q + 51 for some integer q and plug it into our first congruence to get
9
78 ≡ 125q + 51 ≡ 5q + 3 ≡ 1 (mod 8) =⇒ q ≡ 6 (mod 8) =⇒ q = 8r + 6,

for some integer r. Finishing, we have

9
78 = 125q + 51 = 125(8r + 6) + 51 = 1000r + 801,

so the last three digits are 801 . 

§19.5 Homework Problems

Problem 19.1 (MATHCOUNTS). [4] Find the least positive integer that leaves a remainder
of 3 when divided by 5, a remainder of 7 when divided by 9, and a remainder of 7 when divided
13.

Problem 19.2 (MATHCOUNTS). [4] For each positive integer n, an = 9n + 2 and bn = 7n + 3.

If the values common to both sequences are written as a sequence, the nth term of that sequence
can be expressed as pn + q, for integers p and q. What is the value of p − q?
1 1 1 1 a
Problem 19.3 (ARML 2002). [6] Let a be an integer such that 1 + 2 + 3 + ··· + 23 = 23! .
Find the remainder when a is divided by 13.

Problem 19.4 (2002 AIME II). [7]It is known that, for all positive integers k,
12 + 22 + 32 + . . . + k 2 = k(k+1)(2k+1)
6 . Find the smallest positive integer k such that 12 + 22 +
32 + . . . + k 2 is a multiple of 200.

Problem 19.5 (Dennis Chen). [7] Let p > 3 be a prime number. Find the remainder when
(13 )(13 + 23 )(13 + 23 + 33 ) · · · (13 + 23 + 33 · · · + (p − 2)3 ) is divided by p.

Problem 19.6 (1989 AIME). [8] One of Euler’s conjectures was disproved in the 1960s by three
American mathematicians when they showed there was a positive integer such that 1335 +1105 +
845 + 275 = n5 . Find the value of n.

16
Everaise Academy (2020) Math Competitions I

Problem 19.7 (2019 AIME I). [8] Find the least odd prime factor of 20198 + 1.

Problem 19.8 (2012 AIME II). [8] For a positive integer p, define the positive integer n to
be p-safe if n differs in absolute value by more than 2 from all multiples of p. For example,
the set of 10-safe numbers is {3, 4, 5, 6, 7, 13, 14, 15, 16, 17, 23, . . . }. Find the number of positive
integers less than or equal to 10, 000 which are simultaneously 7-safe, 11-safe and 13-safe.
Provide a detailed solution, or explain what progress you made, even if you didn’t
get an answer.

17
20 Diophantine Equations, Bases

§20.1 Diophantine Equations

A Diophantine equation is an equation where we only seek integer solutions; like normal
equations, diophantine equations will often come in systems. We often use techniques in both
algebra and number theory to solve these equations.

Example 20.1. Find all integers (x, y) such that 3x + 2y = 5.

Solution. Since we are looking for integer solutions, this is a Diophantine equation. At first
glance, this equation has infinitely many solutions across the integers, and an obvious solution
is (1, 1).
In order to find the other solutions, we can build off of the first solution we found, (1, 1).
Since we are dealing with the expression 3x + 2y, it becomes clear that x increases when y
decreases, and vice versa. Now, we just need to find the specific balance between x and y to
solve for the remaining solutions.
Observe that the GCD of 2 and 3 is 6. Applying this, we have that 3(x+2)+2(y−3) = 3x+2y.
We can take our original solution (1, 1) and repeatedly add 2 to x and subtract 3 from y (and
vice versa) to obtain our infinite solutions.
Thus, our solutions are in the form of (1 + 2a, 1 − 3a) for all integers a. Plugging this in to
our original equation, we have that 3(1 + 2a) + 2(1 − 3a) = 3 + 2 + 6a − 6a = 5, so it does
work. 

Exercise 20.2. Does the equation 2x + 6y = 11 have any integer solutions?

Example 20.3. Find all integers (x, y) such that x2 + 3x + 1 = 3y 2 .

Solution. Notice that there is a coefficient of 3 on both sides of the equation. This motivates
us to take the equation mod 3. Doing so gives us, x2 + 1 ≡ 0 mod 3 which becomes x2 ≡ 2
mod 3.However, a perfect square is never equivalent to 2 mod 3!
Let’s prove this. Consider a number that is equivalent to 0 mod 3. Then squaring it will
result in another number that is also 0 mod 3. Now consider a number 1 mod 3. Squaring
it results in another number 1 mod 3. Finally, consider a number that is 2 mod 3. When we
square it, it becomes a number 4 mod 3 → 1 mod 3. Thus, we have for all perfect squares x2 ,
x2 ≡ 0, 1 mod 3.
This contradicts with our solution that x2 ≡ 2 mod 3, so our original equation has no integral
solutions. 

In the next example, we will introduce Simon’s Favorite Factoring Trick which is a powerful
way to factor certain equations.

Example 20.4. Find all integer solutions (x, y) such that xy = x + y

1
Everaise Academy (2020) Math Competitions I

Solution. After doing a bit of guess and check, we find that (2, 2) amd (0, 0) are valid solutions,
and there don’t seem to be any other solutions.
However, this is a bit unsatisfying because we don’t know for sure whether these were the
only solutions. Let’s actually solve it.
Moving the x and y to the LHS, we have that
xy − x − y = 0
It’s not clear how to proceed from here, but we might think to factor out an x which gives us
x(y − 1) − y = 0

We would be able to factor this expression if we had x(y − 1) − (y − 1) = 0, but there’s no

constant term for us to do so. However, notice that we can add one to both sides to achieve
this factorization! We now have
x(y − 1) − y = 0
x(y − 1) − y + 1 = 1
x(y − 1) − (y − 1) = 1
(x − 1)(y − 1) = 1

Since we seek integer solutions, (x − 1) and (y − 1) must both be factors of 1. We either have
that x − 1 = y − 1 = 1 or that x − 1 = y − 1 = −1. Thus, our two solutions are (2, 2) and
(0, 0). 

Now let’s look at an example where our coefficients are not 1.

Example 20.5. Find all integer solutions (x, y) such that 2xy − 5y − 8x = −13

Solution. Factoring out the y, we have that y(2x − 5) − 8x = −13. Remembering what we did
in the previous problem, we seek to add some constant c to both sides such that we are able to
factor out a (2x − 5) from (−8x + c). Notice that −4(2x − 5) = −8x + 20. Thus, we wish to
add 20 to both sides to be able to completely factor the LHS. We have that
2xy − 5y − 8x = −13
y(2x − 5) − 8x = −13
y(2x − 5) − 8x + 20 = 7
y(2x − 5) − 4(2x − 5) = 7
(y − 4)(2x − 5) = 7
Now, we just need to find the factors of 7 to solve for the values of x and y.
7 = 7 · 1 = (−7) · (−1). However, these two factorizations each give us two cases to consider
because we flip the order of y − 4 and 2x − 5. Thus, we have four cases to consider.
2x − 5 = 7 2x − 5 = 1
y−4=1 y−4=7

2x − 5 = −7 2x − 5 = −1
y − 4 = −1 y − 4 = −7
Solving for (x, y), we have our desired solutions (x, y) = (6, 5), (3, 11), (−1, 3), (2, −3). 

2
Everaise Academy (2020) Math Competitions I

Remark. The strategy we used in the previous two examples is known as Simon’s Favorite
Factoring Trick, or SFFT for short. Whenever we see some expression in the form of
xy + ax + by = c, we add the constant ab to both sides so that we can factor the LHS
expression into (x + a)(x + b). The coefficient in front of the xy term will not always be 1,
but the same idea can be applied.

Example 20.6 (AIME). Find 3x2 y 2 if x and y are integers such that y 2 +3x2 y 2 = 30x2 +517.

Solution. Moving all variables to the LHS, we get the equation 3x2 y 2 − 30x2 + y 2 = 517.
Factoring out 3x2 from the first two terms, we have that 3x2 (y 2 − 10) + y 2 = 517. Remembering
SFFT, we can subtract 10 from both sides so that the LHS factors. Our equation now is

(3x2 + 1)(y 2 − 10) = 507.

Since we have that x and y are both integers, we know that 3x2 + 1 and y 2 − 10 are both factors
of 507. We have that 507 = 3 · 132 . Notice that 3x2 + 1 ≡ 1 mod 3. Thus, we can narrow our
potential solutions to 3x2 + 1 = 13 or 3x2 + 1 = 169. Testing 3x2 + 1 = 169 does not yield an
integer solution for x. However, we find that 3x2 + 1 = 13 and y 2 − 10 = 39 gives the valid
solution of (x, y) = (2, 7). Thus, 3x2 y 2 = 588. 

Exercise 20.7 (AHSME). How many ordered pairs (m, n) of positive integers are solutions to
4 2
+ = 1?
m n
Exercise 20.8. Let A, M , and C be digits with

(100A + 10M + C)(A + M + C) = 2005

What is A?
Now we’ll work through a few more examples of Diophantine equations.

Example 20.9 (AHMSE). The number of triples (a, b, c) of positive integers which satisfy
the simultaneous equations

ab + bc = 44
ac + bc = 23

is

Solution. As we have done in the previous examples, our first step here will be to factor both
equations. Doing so gives us

b(a + c) = 44
c(a + b) = 23

Since we know that our variables are positive integers, we have that b and a + c are both factors
of 44 and that c and a + b are both factors of 23.

3
Everaise Academy (2020) Math Competitions I

Since 23 is prime, we have that either c = 23 or that c = 1. However, we can’t have c = 23

because that implies that a+b = 1. If a+b = 1, a and b can’t both be positive which contradicts
with the “positive integers” condition of our problem.
Thus, we know that c = 1 and a + b = 23. Plugging in c = 1 to our first equation, we have
that ab + b = 44. We also have that a + b = 23. Since we have ab, a, b terms between the two
equations, we are motivated to add them and use SFFT.
Adding the two equations, we have that ab + a + 2b = 67. Factoring the a from the first two
terms, we have a(b + 1) + 2b = 69. Adding 2 to both sides allows us to factorize the equation
as (a + 2)(b + 1) = 69. Since we want positive a, b and remembering that a + b = 23, we have
the two equations
a + 2 = 23 a+2=3
b+1=3 b + 1 = 23
From this, we find our desired solutions (a, b, c) = (21, 2, 1), (1, 22, 1). Thus, there are two
ordered triples of positive integers that satisfy our original two equations. 

Example 20.10 (Math League HS). What are all pairs of positive integers (a, b) for which
a2 + b exceeds a + b2 by 36?

Solution. We are given that a2 + b = a + b2 + 36. Moving all variables to the LHS, we have
that a2 − b2 + b − a = 36. Notice that we can rewrite the LHS as
a2 − b2 + b − a = 36
(a + b)(a − b) − (a − b) = 36
(a + b − 1)(a − b) = 36
Thus, both (a + b − 1) and (a − b) are factors of 36. Notice that a + b − 1 > a − b because when
we rearrange this inequality, we get that 2b > 1 which is always true since we are dealing with
positive integers. Thus, we have four cases
a + b − 1 = 36 a + b − 1 = 18
a−b=1 a−b=2

a + b − 1 = 12 a+b−1=9
a−b=3 a−b=4
Solving these equations, we actually find 3 solutions because the set of equations a + b −
1 = 18, a − b = 2 does not give us integer solutions. Our desired solutions are (x, y) =
(19, 18), (8, 5), (7, 3). 

The main strategies behind solving Diophantine equations are either finding ways to factorize
the original equation(s), or taking the equation mod n to simplify it. Additionally, if we have
fractions or square roots, find ways to get rid of them perhaps by multiplying or squaring both
sides.
Exercise 20.11 (AMC 10). In multiplying two positive integers a and b, Ron reversed the digits
of the two-digit number a. His erroneous product was 161. What is the correct value of the
product of a and b?
Exercise 20.12 (PUMaC). If p, q, and r are primes with pqr = 7(p + q + r), find p + q + r.

4
Everaise Academy (2020) Math Competitions I

§20.2 Bases
Definition 20.13. A base is a system of writing numbers. Our familiar number system uses
base-10. Take the number 237, for example. What do the digits 2, 3, and 7 represent in base-10?
They signify that
237 = 2 · 102 + 3 · 10 + 7
In general, a base-b number an an−1 an−2 . . . a0 , where the digits ai are nonnegative integers less
than b (with an > 0, has the value
n
X
ai bi = an · bn + an−1 · bn−1 + · · · + a0
i=0

This notation may look slightly intimidating, but bases will make a lot more sense after
looking at some examples. Notation-wise, if a number is not written in base-10, we write the
base as a subscript. For example, 3245 denotes 324 in base-5. For base-10, the subscript may
also be used for clarification.

Example. Convert 2315 , 11112 and 3219 to their base-10 equivalent.

Solution.

2315 = 2 · 52 + 1 · 52 + 1 = 5610
11112 = 1 · 23 + 1 · 22 + 1 · 21 + 1 = 1510
3219 = 3 · 92 + 2 · 9 + 1 = 26210

Example. Convert 59 to base-16.

Solution. We can do this by looking for the leading digit, then working our way down. 59 <
256, so the first digit occurs in the “16” column. 59 = 3·16+11. Thus, we have 5910 = 3B16 . 

Something to note is that the digits used in base-b are the digits 0→ b-1. For example,
in base-10, we used the digits 0 → 9. For bases higher than 10, digits greater than 9 are
usually denoted using letters of the alphabet starting from A=10. For example, in hexadecimal
(base-16), we use A, B, C, D, E, and F to represent 10, 11, 12, 13, 14, and 15, respectively.
Operating on numbers in a different base system follows the same general idea as operating
on numbers in base 10, except that “carrying” happens with respect to the base and not 10.
Let’s look at some examples:

Example. What is the sum of 4768 and 3218 ?

4768
Solution. +3218
10178
This addition was done similar to how it is normally done, except that we “carried” over
with respect to 8. 6 +1 = 7 in the last column is done as normal. In the second-to-last

5
Everaise Academy (2020) Math Competitions I

column, 7 + 2 = 910 = 118 , so a 1 is placed down and a 1 is carried over. Then, we have
4 + 3 + 1 = 810 = 108 . Thus, the total is 10178 . 

Example. What is the product of 268 and 218 ?

268
×218
Solution. 268
+5408
5668
Similarly, we can use normal multiplication procedures but just carry over at different values.
Multiply 1 × 26 = 268 as normal, but 268 × 2 = 548 because we carry over at 8 instead. Then,
addition proceeds as normal, where we get 5668 as a result.


Another way to operate on numbers in a different base system is to convert everything to base-
10, operate normally, then convert them back. The method is a matter of personal preference,
although competition problems rarely involve more than a few operations.
Bases frequently show up in both AMC and AIME problems, and base problems are often
simply diophantine equations in disguise.

Example 20.14 (AHSME). Let the product (12)(15)(16), each factor written in base b,
equals 3146 in base b. Let s = 12 + 15 + 16, each term expressed in base b. Then s, in base
b, is (A) 43 (B) 44 (C) 45 (D) 46 (E) 47

Solution. We can convert 12b , 15b , 16b , 3146b to base-10.

12b = b + 2
15b = b + 5
16b = b + 6
3146b = 3b3 + b2 + 4b + 6

Plugging these conversions back into the original equation, we get

(b + 2)(b + 5)(b + 6) = 3b3 + b2 + 4b + 6

b3 + 13b2 + 5b + 60 = 3b3 + b2 + 4b + 6
2(b3 − 6b2 − 24b − 27) = 0

By the Rational Root Theorem, we know that b must divide 27. Furthermore, we must have
b ≥ 7 because the digit 6 is valid in base b. The only factors of 27 that satisfy this criteria are 9
and 27. 27 appears much too large to be a root of the cubic, so b = 9 (this can also be verified
through computation). Thus, s = 129 + 159 + 169 = B)449 

Example 20.15 (2018 AIME I). The number n can be written in base 14 as a b c, can be

6
Everaise Academy (2020) Math Competitions I

written in base 15 as a c b, and can be written in base 6 as a c a c , where a > 0. Find the
base-10 representation of n.

Solution. Let’s first convert the given numbers to base 10. Since n is a b c14 , we have that

n = a · 142 + b · 14 + c = 196a + 14b + c

Converting a c b15 and a c a c6 similarly, we get the two equations

n = 225a + 15c + b
n = 222a + 37c

Setting the above two equations equal to each other,

225a + 15c + b = 222a + 37c

22c = 3a + b

Since c is nonzero, we know that 3a + b = 22c ≥ 22. Also, a is a digit in base 6 and b is a
digit in base 14, so a ≤ 5 and b ≤ 13. Therefore, 22 ≤ 3a + b ≤ 28. So the only possibility is
c = 1. Then, we have

3a + b = 22

Subtracting the first and second equation and plugging in c = 1, we get

196a + 14b + c = 225a + 15c + b

13b = 29a + 14c
13b = 29a + 14

Then, we can easily solve for a and b through guess and check because a, b are integers with a
limited range. We get a = 4 and b = 10. Plugging a, b, and c into one of the original expressions
gives 196a + 14b + c = 925. 

Remark. It is a common technique to convert numbers written in other bases to base-10

for ease of comparison.

Example 20.16 (2020 AIME I). A positive integer N has base-eleven representation abc
and base-eight representation 1 bca, where a, b, and c represent (not necessarily distinct)
digits. Find the least such N expressed in base ten.

Solution. First, convert both numbers to base-10 in terms of the variables so that they can be
compared easily. We have

abc11 = 121a + 11b + c

1 bca8 = 512 + 64b + 8c + a

7
Everaise Academy (2020) Math Competitions I

Equating the two,

121a + 11b + c = 512 + 64b + 8c + a

120a = 512 + 53b + 7c

Notice N = 121a + 11b + c, so we can start by minimizing the variable a. Minimizing a

minimizes N because a is the leading digit when expressing N in base-11. Variables a, b, c must
all be positive, so the RHS of

120a = 512 + 53b + 7c

is at least 512. Therefore, the minimum value of a is 5. Now, we can test if b, c can have
appropriate values if a = 5. Plugging in a = 5, we get

600 = 512 + 53b + 7c

88 = 53b + 7c

Then, we continue by trying to minimize b because b is the next digit in the base-11 repre-
sentation of N . Letting b = 0 clearly won’t work, so we try b = 1. If b=1,

88 = 53(1) + 7c
35 = 7c
5=c

Variable c works out to be an integer, which is exactly what we hoped for. Thus, the solution
is

N = 121a + 11b + c = 5 · 121 + 11 · 1 · 5

= 605 + 11 + 5 = 621

§20.3 Homework Problems

Problem 20.1 (Purple Comet HS 2004). [4] Find n such that n − 76 and n + 76 are both cubes
of positive integers.

Problem 20.2 (2017 AIME I). [5] A rational number written in base eight is ab.cd, where all
digits are nonzero. The same number in base twelve is bb.ba. Find the base-ten number abc.

Problem 20.3 (2015 AMC 10). [5] Hexadecimal (base-16) numbers are written using numeric
digits 0 through 9 as well as the letters A through F to represent 10 through 15. Among the
first 1000 positive integers, there are n whose hexadecimal representation contains only numeric
digits. What is the sum of the digits of n?

Problem 20.4 (1987 AIME). [6] Find 3x2 y 2 if x and y are integers such that y 2 + 3x2 y 2 =
30x2 + 517.

8
Everaise Academy (2020) Math Competitions I

Problem 20.5 (2014 AMC 12). [6] There are exactly N distinct rational numbers k such that
|k| < 200 and
5x2 + kx + 12 = 0
has at least one integer solution for x. What is N ?

Problem 20.6 (2008 AIME I). [6] There exist unique positive integers x and y that satisfy the
equation x2 + 84x + 2008 = y 2 . Find x + y.

Problem 20.7 (2013 AMC 10A). [6] In base 10, the number 2013 ends in the digit 3. In base
9, on the other hand, the same number is written as (2676)9 and ends in the digit 6. For how
many positive integers b does the base-b-representation of 2013 end in the digit 3?

Problem 20.8 (2019 AMC 10). [6] For some positive integer k, the repeating base-k represen-
7
tation of the (base-ten) fraction 51 is 0.23k = 0.232323...k . What is k?

Problem 20.9 (2013 AMC 10). [7] Bernardo chooses a three-digit positive integer N and writes
both its base-5 and base-6 representations on a blackboard. Later LeRoy sees the two numbers
Bernardo has written. Treating the two numbers as base-10 integers, he adds them to obtain
an integer S. For example, if N = 749, Bernardo writes the numbers 10,444 and 3,245, and
LeRoy obtains the sum S = 13,689. For how many choices of N are the two rightmost digits of
S, in order, the same as those of 2N ?

Problem 20.10 (2003 AIME I). [8] Let N be the number of positive integers that are less than
or equal to 2003 and whose base-2 representation has more 1’s than 0’s. Find the remainder
when N is divided by 1000.
Provide a detailed solution, or explain what progress you made, even if you didn’t
get an answer.