Expected Uses of Probability

Evan Chen
August 11, 2014

Email: [email protected]. This is mostly about expected value, both in

its own right and in the context of the probabilistic method.

§1 Definitions and Notation

Nothing tricky here, just setting up notation. I’ll try to not be overly formal.
A random variable is just a quantity that we take to vary randomly. For example,
the outcome of a standard six-sided dice roll, say D6 , is a random variable. We can now
discuss the probability of certain events, which we’ll denote P(•). For instance, we can
write
1
P(D6 = 1) = P(D6 = 2) = · · · = P(D6 = 6) =
6
or P(D6 = 0) = 0 and P(D6 ≥ 4) = 2 . 1

We can also discuss the expected value of a random variable X, which is the “average”
value. The formal definition is
def
X
E[X] = P(X = x) · x.
x

But an example for our dice roll D6 makes this clearer:

1 1 1
E[D6 ] = · 1 + · 2 + · · · + · 6 = 3.5.
6 6 6
In natural language, we just add up all the outcomes weighted by probability they appear.
We’ll assume the reader has some familiarity with basic graph theory terms; see http://
en.wikipedia.org/wiki/Graph_theory#Definitions otherwise. One term we’ll define
here that may not be so known – given a graph G, an independent set is a set of
vertices for which no two are connected by an edge.

§2 Properties of Expected Value

§2.1 A Motivating Example
It is an unspoken law that any introduction to expected value begins with the following
classical example.

Example 2.1
At MOP, there are n people, each of who has a name tag. We shuffle the name tags
and randomly give each person one of the name tags. Let S be the number of people
who receive their own name tag. Prove that the expected value of S is 1.

1
Evan Chen — August 11, 2014 Expected Uses of Probability

This result might seem surprising, as one might intuitively expect E[S] to depend on
the choice of n.
For simplicity, let us call a person a fixed point if they receive their own name tag.1
Thus S is just the number of fixed points, and we wish to show that E[S] = 1. If we’re
interested in the expected value, then according to our definition we should go through
all n! permutations, count up the total number of fixed points, and then divide by n! to
get the average. Since we want E[S] = 1, we expect to see a total of n! fixed points.
Let us begin by illustrating the case n = 4 first, calling the people W , X, Y , Z.
W X Y Z Σ
1 W X Y Z 4
2 W X Z Y 2
3 W Y X Z 2
4 W Y Z X 1
5 W Z X Y 1
6 W Z Y X 2
7 X W Y Z 2
8 X W Z Y 0
9 X Y W Z 1
10 X Y Z W 0
11 X Z W Y 0
12 X Z Y W 1
13 Y W X Z 1
14 Y W Z X 0
15 Y X W Z 2
16 Y X Z W 1
17 Y Z W X 0
18 Y Z X W 0
19 Z W X Y 0
20 Z W Y X 1
21 Z X W Y 1
22 Z X Y W 2
23 Z Y W X 0
24 Z Y X W 0
Σ 6 6 6 6 24
We’ve listed all 4! = 24 permutations, and indeed we see that there are a total of 24
fixed points, which I’ve bolded in red. Unfortunately, if we look at the rightmost column,
there doesn’t seem to be a pattern, and it seems hard to prove that this holds for larger
n.
However, suppose that rather than trying to add by rows, we add by columns. There’s
a very clear pattern if we try to add by the columns: we see a total of 6 fixed points in
each column. Indeed, the six fixed W points correspond to the 3! = 6 permutations of
the remaining letters X, Y , Z. Similarly, the six fixed X points correspond to the 3! = 6
permutations of the remaining letters W , Y , Z.
This generalizes very nicely: if we have n letters, then each letter appears as a fixed
point (n − 1)! times.

1
This is actually a term used to describe points which are unchanged by a permutation. So the usual
phrasing of this question is “what is the expected number of fixed points of a random permutation?”

2
Evan Chen — August 11, 2014 Expected Uses of Probability

Thus the expected value is

 
1  1
E[S] = (n − 1)! + (n − 1)! + · · · + (n − 1)! = · n · (n − 1)! = 1.
n! | {z } n!
n times

Cute, right? Now let’s bring out the artillery.

§2.2 Linearity of Expectation

The crux result of this section is the following theorem.

Theorem 2.2 (Linearity of Expectation)

Given any random variables X1 , X2 , . . . , Xn , we always have

E[X1 + X2 + · · · + Xn ] = E[X1 ] + E[X2 ] + · · · + E[Xn ].

This theorem is obvious if the X1 , X2 , . . . , Xn are independent of each other – if I roll

100 dice, I expect an average of 350. Duh. The wonderful thing is that this holds even if
the variables are not independent. And the basic idea is just the double-counting we did
in the earlier example: even if the variables depend on each other, if you look only at the
expected value, you can still add just by columns. The proof of the theorem is just a
bunch of sigma signs which say exactly the same thing, so I won’t bother including it.
Anyways, that means we can now nuke our original problem. The trick is to define
indicator variables as follows: for each i = 1, 2, . . . , n let

def 1 if person i gets his own name tag

(
Si =
0 otherwise.

Obviously,
S = S1 + S2 + · · · + Sn .
Moreover, it is easy to see that E[Si ] = P(Si = 1) = n1 for each i: if we look at any
particular person, the probability they get their own name tag is simply n1 . Therefore,

1 1 1
E[S] = E[S1 ] + E[S2 ] + · · · + E[Sn ] = + + · · · + = 1.
n
| n {z n
}
n times

Now that was a lot easier! By working in the context of expected value, we get a
framework where the “double-counting” idea is basically automatic. In other words,
linearity of expectation lets us only focus on small, local components when computing an
expected value, without having to think about why it works.

§2.3 More Examples

Example 2.3 (HMMT 2006)

At a nursery, 2006 babies sit in a circle. Suddenly, each baby randomly pokes either
the baby to its left or to its right. What is the expected value of the number of
unpoked babies?

3
Evan Chen — August 11, 2014 Expected Uses of Probability

Solution. Number the babies 1, 2, . . . , 2006. Define

def 1 if baby i is unpoked

(
Xi =
0 otherwise.

We seek E[X1 + X2 + · · · + X2006 ]. Note that any particular baby has probability 1 2 1


2 = 4
of being unpoked (if both its neighbors miss). Hence E[Xi ] = 14 for each i, and

1 1003
E[X1 + X2 + · · · + X2006 ] = E[X1 ] + E[X2 ] + · · · + E[X2006 ] = 2006 · = .
4 2
Seriously, this should feel like cheating.

§2.4 Practice Problems

The first two problems are somewhat straightforward applications of the methods de-
scribed above.
Problem 2.4 (AHSME 1989). Suppose that 7 boys and 13 girls line up in a row. Let
S be the number of places in the row where a boy and a girl are standing next to each
other. For example, for the row GBBGGGBGBGGGBGBGGBGG we have S = 12.
Find the expected value of S.
Problem 2.5 (AIME 2006 #6). Let S be the set of real numbers that can be represented
as repeating decimals of the form 0.abc where a, b, c are distinct digits. Find the sum of
the elements of S.
The next three problems are harder; in these problems linearity of expectation is not
the main idea of the solution. All problems below were written by Lewis Chen.
Problem 2.6 (NIMO 4.3). One day, a bishop and a knight were on squares in the same
row of an infinite chessboard, when a huge meteor storm occurred, placing a meteor in
each square on the chessboard independently and randomly with probability p. Neither
the bishop nor the knight were hit, but their movement may have been obstructed by
the meteors. For what value of p is the expected number of valid squares that the bishop
can move to (in one move) equal to the expected number of squares that the knight can
move to (in one move)?
Problem 2.7 (NIMO 7.3). Richard has a four infinitely large piles of coins: a pile of
pennies, a pile of nickels, a pile of dimes, and a pile of quarters. He chooses one pile at
random and takes one coin from that pile. Richard then repeats this process until the
sum of the values of the coins he has taken is an integer number of dollars. What is the
expected value of this final sum of money, in cents?
Problem 2.8 (NIMO 5.6). Tom has a scientific calculator. Unfortunately, all keys
are broken except for one row: 1, 2, 3, + and -. Tom presses a sequence of 5 random
keystrokes; at each stroke, each key is equally likely to be pressed. The calculator then
evaluates the entire expression, yielding a result of E. Find the expected value of E.
(Note: Negative numbers are permitted, so 13-22 gives E = −9. Any excess operators
are parsed as signs, so -2-+3 gives E = −5 and -+-31 gives E = 31. Trailing operators
are discarded, so 2++-+ gives E = 2. A string consisting only of operators, such as -++-+,
gives E = 0.)

4
Evan Chen — August 11, 2014 Expected Uses of Probability

§3 Direct Existence Proofs

In its simplest form, we can use expected value to show existence as follows: suppose
we know that the average score of the USAMO 2014 was 12.51. Then there exists a
contestant who got at least 13 points, and a contestant who got at most 12 points. This
is similar in spirit to the pigeonhole principle, but the probabilistic phrasing is far more
robust.

§3.1 A First Example

Let’s look at a very simple example, taken from the midterm of a class at the San Jose
State University.2

Example 3.1 (SJSU M179 Midterm)

Prove that any subgraph of Kn,n with at least n2 −n+1 edges has a perfect matching.

We illustrate the case n = 4 in the figure.

Figure 1: The case n = 4. There are n2 − n + 1 = 13 edges, and the matching is

highlighted in green.

This problem doesn’t “feel” like it should be very hard. After all, there’s only a total
of n2 possible edges, so having n2 − n + 1 edges means we have practically all edges
present.3
So let’s be really careless and just randomly pair off one set of points with the other,
regardless of whether there is actually an edge present. We call the score of such a pairing
the number of pairs which are actually connected by an edge. We wish to show that
some pairing has score n, as this will be the desired perfect matching.
So what’s the expected value of a random pairing? Let v1 , . . . , vn be the n vertices on
the left. For each i, let4 ,

def 1 if the pair with vi has an edge

(
Xi =
0 otherwise.

deg vi
Then the score of the configuration is X = X1 +X2 +· · ·+Xn . Now we have E[Xi ] = n ,

2
For a phrasing of the problem without graph theory: given n red points and n blue points, suppose we
connect at least n2 − n + 1 pairs of opposite colors. Prove that we can select n segments, no two of
which share an endpoint.
3
On the other hand, n2 − n + 1 is actually the best bound possible. Can you construct a counterexample
with n2 − n?
4
Thanks to D. Grozev for a correction here

5
Evan Chen — August 11, 2014 Expected Uses of Probability

so
E[X] = E[X1 ] + · · · + E[Xn ]
deg v1 deg v2 deg vn
= + + ··· +
n n n
n2 − n + 1 1
= =n−1+ .
n n
Since X takes only integer values, there must be some configuration which achieves
X = n. Thus, we’re done.

§3.2 Ramsey Numbers

Let’s do another simple example. Before we begin, I will quickly introduce a silly algebraic
lemma, taken from [5, page 30].

Lemma 3.2
For any positive integers n and k,
 
n 1  en k
< .
k e k

Here e ≈ 2.718 . . . is Euler’s constant.

Proof. Do n nk
and then use calculus to prove that k! ≥ e(k/e)k . Specifically,


k < k!
Z k
ln 1 + ln 2 + · · · + ln k ≥ ln x dx = k ln k − k + 1
x=1
whence exponentiating works.

Algebra isn’t much fun, but at least it’s easy. Let’s get back to the combinatorics.

Example 3.3 (Ramsey Numbers)

Let n and k be integers with n ≤ 2k/2 and k ≥ 3. Then it is possible to color
the edges of the complete graph on n vertices each either red
or blue with the
following property: one cannot find k vertices for which the 2 edges among them
k

are monochromatic.

Remark. In the language of Ramsey numbers, prove that R(k, k) > 2k/2 .

Solution. Again we just randomly color the
edges and hope for the best. We use a coin
flip to determine the color of each of the n2 edges. Let’s call a collection of k vertices
bad if all k2 edges are the same color. The probability that any collection is bad is

 (k)−1
1 2
.
2
The number of collections in n
, so the expected number of bad collections is


k
n


E[number of bad collections] = k
k
.
2(2)−1

6
Evan Chen — August 11, 2014 Expected Uses of Probability

We just want to show this is less than 1. You can check this fairly easily using Lemma 3.2;
in fact, we have a lot of room to spare.

§3.3 A Tricky Application

To cap off this section, we give a tricky proof (communicated to me via [3]) of the
following result.

Theorem 3.4 (Ajtai-Komlós-Szemerédi)

Given a triangle-free graph G with average degree d and N vertices, we can find an
independent set with size at least 0.01 Nd log d.

Here, triangle-free just means there are no three vertices which are all adjacent to each
other.5 Another phrase for this is locally sparse.
Our first move is to try and replace the “average degree” d with “maximum degree”
∆. Here’s the trick: notice that at most half of the vertices have degree greater than 2d.
So if we throw away these vertices, we still have half the vertices and left, and now the
maximum degree is ∆ ≤ 2d. If we let n = N /2, then we just need an independent set of
size 0.04 ∆
n
log ∆ in our new graph.
So now we have n vertices with maximum degree ∆. Here’s the trick: consider all
possible independent sets, and pick one set S uniformly at random (!). For this set S, we
define a score X as follows:

• For each vertex u in S, we write a +∆ at that u.

• For each vertex v adjacent to something in S, we write +1 at that vertex. A vertex

can receive +1 multiple times. However, note that since S is independent, this
means v ∈/ S.

• Define the score X to be the sum of all numbers written.

+1

+1

+∆
+∆ +2

+1

Figure 2: Assigning scores. The elements of S are the large red vertices.

Obviously, X ≤ 2∆ |S|, since each vertex in S bestows ∆ to itself and at most ∆

among its neighbors. Now, we will place a bound on E[X], which will give us the result.
Consider any vertex v, and consider its set of neighbors. Note that by the triangle-free
condition, no neighbors are adjacent to each other. Let Xv denote the sum of the scores

5
If you’re familiar with the notation R(m, n), here’s some food for thought: what’s the connection between
this and R(3, t)?

7
Evan Chen — August 11, 2014 Expected Uses of Probability

given to vertex v (so X = v Xv ). We are going to show that E[Xv ] ≥ 0.08 log ∆. This
P
is enough, because then E[X] ≥ 0.08n log ∆, and for a good choice of X, we then have
|S| ≥ 0.04n log∆∆ .

Neighbors

v ...

Figure 3: Ignoring things.

Suppose we’re selecting an independent set, and we’re done selecting everything aside
from v and its neighbors. We’ll prove that regardless of how the stuff outside is chosen,
E[Xv ] ≥ 0.08 log ∆ still holds. Assume that, not including v, there are m other vertices in
the neighborhood which we can still pick (i.e. they are not adjacent to anything outside
that has been selected).
There are a few ways we can pick the remaining set:
• We can pick v, but then we can no longer pick any of its neighbors.
• We can pick any nonempty subset of the m remaining vertices, but then we can no
longer pick v.
• We can pick no vertices.
There are a total of 1 + (2m − 1) + 1 possibilities. In the first scenario, the Xv = +∆. In
the second and third scenario, Xv = E[# neighbors chosen] = 12 m. So,

1 · ∆ + 2m · 12 m
 
∆ 1 m 1 ∆
E[Xv ] = = + · > max , m .
2m + 1 2m + 1 1 + 2−m 2 4 2m
It remains to prove this is at least 0.08 log ∆. You can check this, because if m ≥ 1
log2 ∆,
√ 2
then 14 m is enough; otherwise, 2∆m ≥ ∆ which is certainly sufficient.

§3.4 Practice Problems

The first two problems are from [2]; the last one is from [4].
Problem 3.5. Show that one can construct a (round-robin) tournament with more than
1000 people such that in any set of 1000 people, some contestant beats all of them.
Problem 3.6 (BAMO 2004). Consider n real numbers, not all zero, with sum zero.
Prove that one can label the numbers as a1 , a2 , . . . , an such that
a1 a2 + a2 a3 + · · · + an a1 < 0.
Problem 3.7 (Russia 1996). In the Duma there are 1600 delegates, who have formed
16000 committees of 80 people each. Prove that one can find two committees having no
fewer than four common members.

8
Evan Chen — August 11, 2014 Expected Uses of Probability

§4 Heavy Machinery
Here are some really nice ideas used in modern theory. Unfortunately I couldn’t find
many olympiad problems that used them. If you know of any, please let me know!

§4.1 Alteration
In previous arguments we often proved a result by showing E[bad] < 1. A second method
is to select some things, find the expected value of the number of “bad” situations, and
subtract that off. An example will make this clear.

Example 4.1 (Weak Turán)

A graph G has n vertices and average degree d. Prove that it is possible to select an
independent set of size at least 2d
n
.

Proof. Rather than selecting 2d n

vertices randomly and hoping the number of edges is 1,
we’ll instead select each vertex with probability p. (We will pick a good choice of p later.)
That means the expected number of vertices we will take is np. Now there are 12 nd
edges, so the expected number of “bad” situations (i.e. an edge in which both vertices
are taken) is 12 nd · p2 .
Now we can just get rid of all the bad situations. For each bad edge, delete one of its
endpoints arbitrarily (possibly with overlap). This costs us at most 12 nd · p2 vertices, so
the expected value of the number of vertices left is
 
1 2 1
np − ndp = np 1 − dp .
2 2
It seems like a good choice of p is d,
1
which now gives us an expected value of 2d ,
n
as
desired.
A stronger result is Problem 6.5.

§4.2 Union Bounds and Markov’s Inequality

A second way to establish existence is to establish a nonzero probability. One way to do
this is using a union bound.

Proposition 4.2 (Union Bound)

Consider several events A1 , A2 , . . . , Ak . If

P(A1 ) + P(A2 ) + · · · + P(Ak ) < 1

then there is a nonzero probability that none of the events occur.

The following assertion is sometimes useful for this purpose.

Theorem 4.3 (Markov’s Inequality)

Let X be a random variable taking only nonnegative values. Suppose E[X] = c.
Then
1
P(X ≥ rc) ≤ .
r

9
Evan Chen — August 11, 2014 Expected Uses of Probability

This is intuitively obvious: if the average score on the USAMO was 7, then at most 16 of
the contestants got a perfect score. The inequality is also sometimes called Chebyshev’s
inequality or the first Chebyshev inequality.

§4.3 Lovász Local Lemma

The Lovász Local Lemma (abbreviated LLL) is in some sense a refinement of the union
bound idea – if the events in question are “mostly” independent, then the probability no
events occur is still nonzero.
We present below the “symmetric” version of the Local Lemma. An asymmetric version
also exists (see Wikipedia).

Theorem 4.4 (Lovász Local Lemma)

Consider several events, each occurring with probability at most p, and such that
each event is independent of all the others except at most d of them. Then if

epd ≤ 1

the probability that no events occur is positive.

Note that we don’t use the number of events, only the number of dependencies.
As the name implies, the local lemma is useful in situations where in a random
algorithm, it appears that things do not depend much on each other. The following
Russian problem is such an example.

Example 4.5 (Russia 2006)

At a tourist camp, each person has at least 50 and at most 100 friends among the
other persons at the camp. Show that one can hand out a T-shirt to every person
such that the T-shirts have (at most) 1331 different colors, and any person has 20
friends whose T-shirts all have pairwise different colors.

The constant C = 1331 is extremely weak. We’ll reduce it to C = 48 below.

Solution. Give each person a random T-shirt. For each person P , we consider the event
E(P ) meaning “P ’s neighbors have at most 19 colors of shirts”. We wish to use the Local
Lemma to prove that there is a nonzero probability that no events occur.
If we have two people A and B, and they are neither friends nor have a mutual
friend (in graph theoretic language, the distance between them is at least two), then the
events E(A) and E(B) do not depend on each other at all. So any given E(P ) depends
only on friends, and friends of friends. Because any P has at most 100 friends, and
each of these friends has at most 99 friends other than P , E(P ) depends on at most
100 + 100 · 99 = 1002 other events. Hence in the lemma we can set d = 1002 .
For a given person, look at their 50 ≤ k ≤ 100 neighbors. The probability that there
are at most 19 colors among the neighbors is clearly at most
   k
C 19
· .
19 C

To estimate the binomial coefficient, we can again use our silly Lemma 3.2 to get that

10
Evan Chen — August 11, 2014 Expected Uses of Probability

this is at most
 19  k  k−19  31
1 eC 19 18 19 18 19
· =e · ≤e .
e 19 C C C

Thus, we can put p = e18 19 31

. Thus the Lemma implies we are done as long as


C
 31
19 19
e · 1002 ≤ 1.
C

It turns out that C = 48 is the best possible outcome here. Needless to say, establishing
the equality when C = 1331 is trivial.

§5 Grand Finalé – IMO 2014, Problem 6

This article was motivated by the following problem, given at the 55th International
Mathematical Olympiad, and the talk by Po-Shen Loh [3] given on it.

Example 5.1 (IMO 2014/6)

A set of lines in the plane is in general position if no two are parallel and no three
pass through the same point. A set of lines in general position cuts the plane into
regions, some of which have finite area; we call these its finite regions. Prove that for
all sufficiently large n, in any set of n lines in general position it is possible to colour
√
at least n lines blue in such a way that none of its finite regions has a completely
blue boundary.
√ √
Note: Results with n replaced by c n will be awarded points depending on the
value of the constant c.

We’ll present two partial solutions (c < 1), one using Local Lovász, and one using
alteration. For completeness we also present the official solution obtaining
√ c = 1, even
though it is not probabilistic. Then, we will establish the bound O( n log n) using some
modern tools (this was [3]).

§5.1 Partial Solution Using LLL

√
We’ll show the bound c n where c = (6e)√− 2 .
1

√
Split the n lines into c n groups of size cn each, arbitrarily. We are going to select one
line from each of the groups at random to be blue. Let the regions be R1 , R2 , . . . , Rm .
For each region Rk we consider an event Ak meaning “the three chosen lines bounding
Rk are blue”; We will show there is a nonzero probability that no events occur.

3
The probability of Ak is at most cn−1/2 . (It is equal to this if the three of the
chosen lines are from different groups, and is zero if any two are in the same
√
group.)
For each Rk , we have three groups to consider. Each group consists of c lines. Each
n
√
line is part of at most 2n − 2 regions. Hence Ak depends on at most 3 · c
n
· (2n − 2)
events.
Thus,
  √
c 3
 
n
e √ 3· · (2n − 2) < 6ec2 = 1
n c
and we are done by LLL.

11
Evan Chen — August 11, 2014 Expected Uses of Probability

§5.2 Partial Solution Using Alteration

√
We’ll show the bound c n for any c < 23 .
First, we need to bound the number of triangles.

Claim — There are at most 13 n2 triangles.

Proof. Consider each of the n2 intersection of two lines. One can check it is the vertex

of at most
two triangles. Since each triangle has three vertices, this implies there are at
most 23 n2 < 13 n2 triangles.

It is also not hard to show there are at most 12 n2 finite regions6 .

Now color each line blue with probability p. The expected value of the number of lines
chosen is
E[lines] = np.
The expected number of completely blue triangles is less than
1
E[bad triangles] < n2 · p3 .
3
For the other finite regions, of which there are at most 12 n2 , the probability they are
completely blue is at most p4 . So the expected number of completely blue regions here is
at most
1
E[bad polygons with 4+ sides] < n2 · p4 .
2
Note that the expected number of quadrilaterals (and higher) is really small compared
to any of the preceding quantities; we didn’t even bother subtracting off the triangles
that we already counted earlier. It’s just here for completeness, but we expect that it’s
going to die out pretty soon.
Now we do our alteration – for each bad, completely blue region, we un-blue one line.
Hence the expected number of lines which are blue afterwards is
 2   2
np2 np3
  
n 3 n 4
np − ·p − · p = np 1 − − .
3 2 3 2
√
Ignore the rightmost np2 for now, since it’s really small. We want p = k/ n for some k;
3

the value is roughly k · (1 − k 2 /3) at this point, so an optimal value of p is p = n−1/2

(that is, k = 1); this gives

√ 2√
 
2 27 1 81
n· − √ = n− .
3 16 n 3 32
√
For n sufficiently large, this exceeds c n, as desired.

§5.3 Interlude – Sketch of Official Solution Obtaining c = 1

This is not probabilistic, but we include it for completeness anyways. It is in fact just a
greedy algorithm.
Suppose we have colored k of the lines blue, and that it is not possible to color any
additional lines. That means any of the n − k non-blue lines is the side of some finite
region with an otherwise entirely blue perimeter. For each such line `, select one such

12
Evan Chen — August 11, 2014 Expected Uses of Probability

Figure 4: Here ` is the eyelid of v.

region, and take the next counterclockwise vertex; this is the intersection of two blue
lines v. We’ll say ` is the eyelid of v.
You can prove without too much difficulty that every intersection of two blue lines has
at most two eyelids. Since there are n2 such intersections, we see that

 
k
n−k ≤2 = k2 − k
2

so n ≤ k 2 , as required.

√
Figure 5: The greedy algorithm cannot do better than n.

It’s interesting to note that the greedy algorithm cannot be extended to achieve a
√
result better than n. To show this, note that if
n = m2 , we can consider m arbitrary
blue lines in general position, and then add 2 m 2 lines, two on either side of a given
intersection point. (Po-Shen Loh called these “tubes” in his talk.) Thus each of the new
lines is the edge of a triangle with two blue sides, and so the greedy algorithm must stop
here.

§5.4 Overkill Solution

This
√ solution is due to Po-Shen Loh [3]. We are now going to establish the bound
cn log n. The heart is the following theorem.

Theorem 5.2 (Duke-Lefmann-Rödl)

Given a hypergraph G with N vertices and with edges all of size 3, suppose that for
any two vertices at most one √ 3-edge joins them. Then we can find an independent
set with size at least c · √Nd log d.

Here a hypegraph is a graph in which an “edge” is any subset of vertices, as opposed to

just two vertices. In the above theorem, all edges have three endpoints, and we require
that any two vertices are joined by at most one edge.
In the context of the IMO problem, suppose we consider each of the n lines as a vertex
and each finite region as a hyper-edge. Like in the previous solution, we treat pentagons,
hexagons, . . .as just quadrilaterals; hence we can assume all edges have size either 3 or 4.
6
Say, use V − E + F = 2 on the graph whose vertices are the n
intersection points and whose edges are


2
the n(n − 2) line segments.

13
Evan Chen — August 11, 2014 Expected Uses of Probability

Once again we use a coin flip weighted with probability p to pick whether a vertex is
chosen. Define the following random variables:

• Let W be the number of vertices remaining. Then E[W ] = pn.

• Let Y be the number of 4-edges. There are at most n2 such edges, so E[Y ] ≤ p4 n2 .

• Let Z be the number of pairs (u, v) with two 3-edges containing both

(in the context
of geometry, there are at most two such edges). Then E[Z] ≤ n2 p4 < p4 n2 .

If we eliminate the situations in Y and Z then we reach a situation in which the theorem
can be applied.
Finally, let X be the number of edges altogether remaining. Since each edge has ≥ 3
vertices and there are ≤ n2 edges, E[X] ≤ n2 p3 .
Using Markov’s Inequality,
1
P(Y > 4p4 n2 ) < .
4
Similarly,
1 1
P(Z > 4p4 n2 ) < and P(X > 4n2 p3 ) < .
4 4
Meanwhile, W is a binomial distribution, so one can actually show that,

P(W < 0.99pn) → 0 as n → ∞.

Consequently, the union bound implies there is a nonzero chance that all these inequalities
fail, meaning Y ≤ 4p4 n2 , Z ≤ 4p4 n2 , and X ≤ 4n2 p3 , and W ≥ 0.99pn.
Now using alteration again, we delete the “bad” situations in Y and Z. Then the
number of vertices, N , is at least

N ≥ W − Y − Z ≥ 0.99pn − 8p4 n2 ∼ pn(1 − 8p3 n)

Let’s pick p = 0.01n−1/3 . Now N ∼ pn.

The average degree is at most

3X ∼ n2 p 3
d= ≤ ∼ np2 .
N ∼ np
The theorem then gives us a bound of
q p
N pn
√ log d ∼ √ log pn2 ∼ n log n
p
d p n

as desired.

§6 Practice Problems
These problems are mostly taken from [2, 4].
Problem 6.1 (IMC 2002). An olympiad has six problems and 200 contestants. The
contestants are very skilled, so each problem is solved by at least 120 of the contestants.
Prove that there exist two contestants such that each problem is solved by at least one
of them.

14
Evan Chen — August 11, 2014 Expected Uses of Probability

Problem 6.2 (Romania 2004). Prove that for any complex numbers z1 , z2 , . . . , zn ,
satisfying |z1 |2 + |z2 |2 + · · · + |zn |2 = 1, one can select ε1 , ε2 , . . . , εn ∈ {−1, 1} such that
 
Xn 
εk zk  ≤ 1.
 

 
k=1

Problem 6.3 (Shortlist 1999 C4). Let A be a set of N residues (mod N 2 ). Prove that
there exists a set B of of N residues (mod N 2 ) such that A + B = {a + b|a ∈ A, b ∈ B}
contains at least half of all the residues (mod N 2 ).
Problem 6.4 (Iran TST 2008/6). Suppose 799 teams participate in a round-robin
tournament. Prove that one can find two disjoint groups A and B of seven teams each
such that all teams in A defeated all teams in B.
Problem 6.5 (Caro-Wei Theorem). Consider a graph G with vertex set V . Prove that
one can find an independent set with size at least
X 1
.
deg v + 1
v∈V

Remark. Note that, by applying Jensen’s inequality, our independent set has size at least
d+1 , where d is the average degree. This result is called Turán’s Theorem (or the
n

complement thereof).

Problem 6.6 (USAMO 2012/6). For integer n ≥ 2, let x1 , x2 , . . . , xn be real numbers

satisfying x1 +x2 +. . .+xn = 0 and x21 +x22 +. . .+x2n = 1. For each subset A ⊆ {1, 2, . . . , n},
define X
SA = xi .
i∈A

(If A is the empty set, then SA = 0.) Prove that for any positive number λ, the number
of sets A satisfying SA ≥ λ is at most 2n−3 /λ2 . For which choices of x1 , x2 , . . . , xn , λ
does equality hold?
Problem 6.7 (Online Math Open, Ray Li). Kevin has 2n − 1 cookies, each labeled with
a unique nonempty subset of {1, 2, . . . , n}. Each day, he chooses one cookie uniformly at
random out of the cookies not yet eaten. Then, he eats that cookie, and all remaining
cookies that are labeled with a subset of that cookie. Compute the expected value of the
number of days that Kevin eats a cookie before all cookies are gone.
Problem 6.8. Let n be a positive integer. Let ak denote the number of permutations of
n elements with k fixed points. Compute

a1 + 4a2 + 9a3 + · · · + n2 an .

Problem 6.9 (Russia 1999). In a certain school, every boy likes at least one girl. Prove
that we can find a set S of at least half the students in the school such that each boy in
S likes an odd number of girls in S.
Problem 6.10 (Sperner). Consider N distinct subsets S1 , S2 , . . . , SN of {1, 2, . . . , n}
such that no Si is a subset of any Sj . Prove that
 
n
N ≤ 1  .
2n

15
Evan Chen — August 11, 2014 Expected Uses of Probability

Problem 6.11. Let n be a positive integer. Suppose 11n points are arranged in a circle,
colored with one of n colors, so that each color appears exactly 11 times. Prove that one
can select a point of every color such that no two are adjacent.
Problem 6.12 (Sweden 2010, adapted). In a town with n people, any two people either
know each other,√or they both know someone in common. Prove that one can find a
group of at most n log n + 1 people, such that anyone else knows at least one person in
the group.

Remark. In graph theoretic language

√ – given a graph with diameter 2, prove that a
dominating set of size at most n log n + 1 exists.

Problem 6.13 (Erdös). Prove that in any set S of n distinct positive integers we can
always find a subset T with 13 n or more elements with the property that a + b 6= c for
any a, b, c ∈ T (not necessarily distinct).

Remark. Such sets are called sum-free.

Problem 6.14 (Korea 2016). Let U be a set of m triangles. Prove that there exists a
subset W ⊆ U with at least 0.45m0.8 triangles, with the following property: there are no
points A, B, C, D, E, F for which ABC, BCD, CDE, DEF , EF A, F AB are all in W .

§7 Solution Sketches
2.4 Answer: 9.1. Make an indicator variable for each adjacent pair.

2.5 Answer: 360. Pick a, b, c randomly and compute E[0.abc]. Then multiply by |S|.

2.6 8(1 − p) = 4 · (1 − p) + (1 − p)2 + (1 − p)3 + . . . .

2.7 Let xn be the EV at a state with n (mod 100). Then x0 = 0 and

1
xn = ((xn+1 + 1) + (xn+5 + 5) + (xn+10 + 10) + (xn+25 + 25)) .
4
Do algebra.

2.8 Answer: 1866. Show that one can replace + or - buttons with STOP. Show that one
can replace 1 and 3 buttons with 2. Let p = 35 . Compute 2(p + 10p2 + · · · + 104 p5 ).

3.5 Suppose there are n people, and decide each edge with a coin flip. Compute the
expected number of 1000-subsets for which there is no one better than all. Check that
this is less than 1 for very large n.

3.6 Show that a random permutations has expected value at most 0. Why are the
inequalities strict?

3.7 Let ni be the number of committees which the ith delegate is in. Pick two committees
randomly and find the expected value
P ofnithe

number of common members. Use Jensen’s
inequality to get a lower bound on 2 .

16
Evan Chen — August 11, 2014 Expected Uses of Probability

6.1 Pick the contestants randomly. Find the expected number of problems both miss.

6.2 Select each of the εi randomly with a coin flip. Square the left-hand side and use
the fact that |z|2 = zz for any z.

6.3 Randomly selecting B
works; you can even permit repeated elements in B. You may
need the inequality 1 − n1 ≤ 1e .
n

6.4 Let dk be the number of teams which defeat the kth team (here 1 ≤ k ≤ 799). Select
A randomly and compute the expected number of teams dominated by everyone in A.
You need Jensen on the function x7 .

6.5 Use the following greedy algorithm – pick a random vertex, then delete it and all its
neighbors. Repeat until everything is gone.

6.6 Compute E[SA

2 ] for a random choice of A. Markov Inequality.

6.7 The number of days equals the number of times a cookie is chosen. Compute the
probability any particular cookie is chosen; i.e. the expected value of the number of times
the cookie is chosen. Sum up.

6.8 For a random permutation let X be the number of fixed points. We already know
E[X] = 1. Compute E[ X2 ]. Use this to obtain E[X 2 ].

6.9 Use a coin flip to decide whether to select each girl, then take as many boys as
possible. Show that any person, girl or boy, has exactly a 50% chance of being chosen.

6.10 First prove that

N
X 1
n

≤ 1.
k=1 |Sk |

To do this, consider a random maximal chain of subsets

∅ = T0 ⊂ T1 ⊂ T2 ⊂ · · · ⊂ Tn = {1, 2, . . . , n}.

Compute the expected number of intersections of this chain with {S1 , S2 , . . . , SN }.

6.11 LLL. Here p = 11−2 and d = 42.

6.12 If any vertex has small degree, √ then its neighbors are already the desired set. So
assume all degrees are greater than n log n. Pick each person with probability p for
some well-chosen p; then we expect to pick np people. Show that the probability someone
fails is less than n1 and use a union bound. The inequality 1 − p ≤ e−p is helpful.

6.13 Work modulo a huge prime p = 3k + 2. Find a nice sum-free (mod p) set U of size
k + 1 first, and then consider Un = {nx | x ∈ U } for a random choice of n. Compute
E[|S ∩ Un |].

6.14 Fix U and use alteration. Add a triangle to W with probability p, then for every
bad 6-tuple contained in W , delete one of the triangles from W .

17
Evan Chen — August 11, 2014 Expected Uses of Probability

References
[1] pythag011 at http://www.aops.com/Forum/viewtopic.php?f=133&t=481300

[2] Ravi B’s collection of problems, available at:

http://www.aops.com/Forum/viewtopic.php?p=1943887#p1943887.

[3] Problem 6 talk (c > 1) by Po-Shen Loh, USA leader, at the IMO 2014.

[4] Also MOP lecture notes: http://math.cmu.edu/~ploh/olympiad.shtml.

[5] Lecture notes by Holden Lee from an MIT course:

http://web.mit.edu/~holden1/www/coursework/math/18997/notes.pdf

Thanks to all the sources above. Other nice reads that I went through while preparing
this, but eventually did not use:

1. Alon and Spencer’s The Probabilistic Method. The first four chapters are here:
http://cs.nyu.edu/cs/faculty/spencer/nogabook/.
2. A MathCamp lecture that gets the girth-chromatic number result:
http://math.ucsb.edu/~padraic/mathcamp_2010/class_graph_theory_probabilistic/
lecture2_girth_chromatic.pdf

18