Art of Problem Solving

WOOT 2012–13
Combinatorics A

Solutions to Exercises
1. How many 10-digit binary numbers start with 2 ones or end with 2 ones (or both)?

Solution. First, we count the number of 10-digit binary numbers that start with two ones. The first
two digits are determined, and then each of the remaining 8 digits can be 0 or 1, so there are 28 such
binary numbers.
Next, we count the number of 10-digit binary numbers that end with two ones. The last two digits are
determined. The first digit must be a 1, otherwise the number would not have 10 digits. Each of the
remaining 7 digits can be 0 or 1, so there are 27 such binary numbers.
Finally, we count the number of 10-digit binary numbers that both start with two ones and end with
two ones. The first two digits and the last two digits are determined, and then each of the remaining
6 digits can be 0 or 1, so there are 26 such binary numbers.
Therefore, by PIE, there are 28 + 27 − 26 = 320 10-digit binary numbers that either start with two
ones or end with two ones.
2. Suppose that 80% of U.S. households own a DVD player and that 70% of U.S. households own a
computer. What is the range of possible percentages of U.S. households that own both?

Solution. Let x% be the percentage of U.S. households that own both a DVD player and a computer.
Then the percentage of U.S. households that own a DVD player but not a computer is (80 − x)%,
and the percentage of U.S. households that own a computer but not a DVD player is (70 − x)%.
Then by PIE, the percentage of U.S. households that own either a DVD player or a computer is
(80 − x)% + (70 − x)% + x% = (150 − x)%.
Each of these percentages must be between 0% and 100%, giving us the ranges:

0 ≤ x ≤ 100 ⇒ 0 ≤ x ≤ 100
0 ≤ 80 − x ≤ 100 ⇒ −20 ≤ x ≤ 80
0 ≤ 70 − x ≤ 100 ⇒ −30 ≤ x ≤ 70
0 ≤ 150 − x ≤ 100 ⇒ 50 ≤ x ≤ 150

So 50 ≤ x ≤ 70. Therefore, the range of possible percentages of U.S. households that own both a DVD
player and a computer is between 50% and 70% .
Intuitively, we can see that this answer makes sense. Clearly not more than 70% of households can
own both, since only 70% own a computer. If everyone who owns a computer also owns a DVD player,
then the 70% of households who own a computer will own both. On the other hand, if the 20% who do
not own a DVD player and the 30% who do not own a computer are disjoint sets, then the remaining
50% of people will own both. There is no way fewer people can own both, because there is no way
that more people cannot own one or the other.
3. How many 3-letter words have at least one A? (Solve using PIE.)

Solution. Let W1 , W2 , and W3 be the sets of three-letter words, such that the first letter, second
letter, and third letter is an A, respectively. Then by PIE, the number of three-letter words that have

1
 Art of Problem Solving
WOOT 2012–13
Combinatorics A

at least one A is

#(W1 ∪W2 ∪W3 ) = #(W1 )+#(W2 )+#(W3 )−#(W1 ∩W2 )−#(W1 ∩W3 )−#(W2 ∩W3 )+#(W1 ∩W2 ∩W3 ).

First we count #(W1 ), the number of three-letter words where the first letter is an A. The second and
third letter can be any letters, so #(W1 ) = 262 . Similarly, #(W2 ) = #(W3 ) = 262 .
Next we count #(W1 ∩ W2 ), the number of three-letter words where the first and second letters are A’s.
The third letter can be any letter, so #(W1 ∩ W2 ) = 26. Similarly, #(W1 ∩ W3 ) = #(W2 ∩ W3 ) = 26.
Finally, a word is in W1 ∩ W2 ∩ W3 if and only if all three letters are A’s, so #(W1 ∩ W2 ∩ W3 ) = 1.
Therefore, the number of three-letter words that have at least one A is

#(W1 ∪ W2 ∪ W3 ) = 3 · 262 − 3 · 26 + 1 = 1951 .

Alternatively, we could have found the answer by using complementary counting. If a word does not
have at least one A, then it does not have any As. The number of three-letter words that do not have
any A’s is 253 , so the number of three-letter words that have at least one A is 263 − 253 = 1951.

4. Four coins are flipped one at a time. What’s the probability of getting two consecutive tails?

Solution. Let A, B, and C be the outcomes in which the first two coins, middle two coins, and last
two coins (respectively) come up tails. Note that #(A) = #(B) = #(C) = 4, since for each of these
outcomes, we have 2 possibilities for the other two flips.
Outcomes in A ∩ B have the first three flips come up tails, so there are 2 such outcomes (depending
on whether the 4th flip is heads or tails). Similarly, outcomes in B ∩ C have the last three flips come
up tails, so there are 2 such outcomes. However, outcomes in A ∩ C have all four flips come up tails,
so there is only 1 such outcome. Lastly, outcomes in A ∩ B ∩ C also have all four flips tails, so there
is only 1 such outcome.
Therefore, by PIE, we have that the number of outcomes in A ∪ B ∪ C (that is, the number of outcomes
in which at least one pair of coins comes up tails) is 4 + 4 + 4 − 2 − 2 − 1 + 1 = 8. Since there are
24 = 16 total possible outcomes for the four flips, the probability of getting two consecutive tails is
8 1
16 = 2 .

Note that we can use PIE and symmetry to check our answer. We computed 8 outcomes with two
consecutive tails. By symmetry, there should be 8 outcomes with two consecutive heads. We also
know that there are 2 outcomes with both consecutive heads and consecutive tails (namely, HHTT
and TTHH). So, by PIE, there are 8 + 8 − 2 = 14 outcomes with consecutive heads or consecutive
tails. And, indeed, there are only 16 − 14 = 2 outcomes with no consecutive heads nor consecutive
tails, namely HTHT and THTH.
5. How many positive integers less than 211 are relatively prime to 126?

Solution. First, we factor 126 = 2 · 32 · 7. Thus, a number is relatively prime to 126 if and only if it
is not divisible by 2, 3, or 7.

2
 Art of Problem Solving
WOOT 2012–13
Combinatorics A

We can now follow the solution to Problem 3.8. There are 210 positive integers less than 211. Then
by PIE, the number of positive integers less than or equal to 210 that are divisible by at least one of
2, 3, or 7 is
210 210 210 210 210 210 210
+ + − − − +
2 3 7 2·3 2·7 3·7 2·3·7
= 105 + 70 + 30 − 35 − 15 − 10 + 5
= 150,

which means that there are 210 − 150 = 60 positive integers less than or equal to 210 that are
relatively prime to 126.
6. 3 fans each from Austin High School, Butler High School, and Central High School are seated in a row
of 9 seats. In how many ways can we seat the fans if no three fans from the same school are seated
consecutively?

Solution. Let the letters A, B, and C denote a fan from Austin High School, Butler High School, and
Central High School, respectively. Then we can think of a seating of the nine fans as an arrangement
of the nine letters A, A, A, B, B, B, C, C, and C. We want to count the number of arrangements where
no three letters appear together.
Let SA , SB , and SC be the sets of arrangements where the three letters A, B, and C appear together,
respectively. Then by PIE, the number of arrangements where any of the three letters A, B, or C
appear consecutively is

#(SA ∪SB ∪SC ) = #(SA )+#(SB )+#(SC )−#(SA ∩SB )−#(SA ∩SC )−#(SB ∩SC )+#(SA ∩SB ∩SC ).

First we count #(SA ), the number of arrangements where the As are consecutive. We consider the
three letters A as a single block. Then the number of arrangements of the single block AAA, three
letter Bs, and three letter Cs is
7!
= 140.
1!3!3!
Once the positions of the letters have been determined, the fans are distinguishable, so there are 3!
ways to seat the fans from each school. Therefore, #(SA ) = 140 · 3! · 3! · 3! = 30,240. Similarly,
#(SB ) = #(SC ) = 30,240.
Next we count #(SA ∩ SB ), the number of arrangements where the A’s are consecutive and the B’s
are consecutive. We consider both the three letters A and the three letters B as a single block. Then
the number of arrangements of the single block AAA, the single block BBB, and three letter Cs is
5!
= 20.
1!1!3!
As before, there are 3! ways to seat the fans from each school, so #(SA ∩ SB ) = 20 · 3! · 3! · 3! = 4,320.
Similarly, #(SA ∩ SC ) = #(SB ∩ SC ) = 4,320.

3
 Art of Problem Solving
WOOT 2012–13
Combinatorics A

Finally, to count #(SA ∩ SB ∩ SC ), we consider the three letters A, the three letters B, and the three
letters C each as a single block. Then the number of arrangements is 3!, so #(SA ∩ SB ∩ SC ) =
3! · 3! · 3! · 3! = 1,296.
Therefore,
#(SA ∪ SB ∪ SC ) = 3 · 30,240 − 3 · 4,320 + 1,296 = 79,056.
There are a total of 9! = 362,880 ways of seating the nine fans, which means that there are 362,880 −
79,056 = 283,824 ways of seating the nine fans so that no three fans of any high school are seated
consecutively.
7. Yeechi has a deck of cards consisting of the 2 through 5 of hearts and the 2 through 5 of spades. She
deals two cards (at random) to each of four players. What is the probability that no player receives a
pair?

Solution. Let A1 , A2 , A3 , and A4 be the sets of deals where the first, second, third, and fourth player
receives a pair, respectively. Then by PIE, the number of hands where at least one player receives a
pair is

#(A1 ∪ A2 ∪ A3 ∪ A4 ) = #(A1 ) + #(A2 ) + #(A3 ) + #(A4 )

− #(A1 ∩ A2 ) − #(A1 ∩ A3 ) − #(A1 ∩ A4 ) − #(A2 ∩ A3 ) − #(A2 ∩ A4 ) − #(A3 ∩ A4 )
+ #(A1 ∩ A2 ∩ A3 ) + #(A1 ∩ A2 ∩ A4 ) + #(A1 ∩ A3 ∩ A4 ) + #(A2 ∩ A3 ∩ A4 )
− #(A1 ∩ A2 ∩ A3 ∩ A4 ).

First we count #(A1 ), the number of ways the first player can receive a pair. There are four possible
pairs (one for each denomination from 2 through 5),
and

the number of ways the remaining six cards
can be dealt to the remaining three players is 62 42 = 90, because we have 62 choices for cards

for the first remaining player, then 42 choices for cards for the next remaining player. Therefore,

#(A1 ) = 4 · 90 = 360. Similarly, #(A2 ) = #(A3 ) = #(A4 ) = 360.

Next we count #(A1 ∩ A2 ), the number of ways both the first player and second player can receive
a pair. There are four possible pairs for the first player, then three possible pairs for the second
player. Then the number of ways the remaining four cards can be dealt to the remaining two players is
4


2 = 6. Therefore, #(A1 ∩ A2 ) = 4 · 3 · 6 = 72. Similarly, #(A1 ∩ A3 ) = #(A1 ∩ A4 ) = #(A2 ∩ A3 ) =
#(A2 ∩ A4 ) = #(A3 ∩ A4 ) = 72.
Finally, if three players receive a pair, then the fourth remaining player must also receive a pair. This
can occur in 4! = 24 ways. Therefore, #(A1 ∩ A2 ∩ A3 ) = #(A1 ∩ A2 ∩ A4 ) = #(A1 ∩ A3 ∩ A4 ) =
#(A2 ∩ A3 ∩ A4 ) = #(A1 ∩ A2 ∩ A3 ∩ A4 ) = 24.
So the number of hands where at least one player receives a pair is

#(A1 ∪ A2 ∪ A3 ∪ A4 ) = 4 · 360 − 6 · 72 + 4 · 24 − 24 = 1080.

The total number of possible hands is

8!
= 2520.
2!2!2!2!

4
 Art of Problem Solving
WOOT 2012–13
Combinatorics A

Hence, the probability that no player receives a pair is

1080 3 4
1− =1− = .
2520 7 7

Seeing a simple answer might make us wonder if there’s a more direct solution, and there is. We will
present just the highlights of the solution, and let you fill in the details.
We see that player #1 does not get a pair with probability 67 (since for any first card that player #1
receives, 6 of the 7 remaining cards do not match it). Suppose without loss of generality that player
#1 receives a 2 and a 3, and that a 4 is dealt to player #2. (Think about why we can make these
assumptions.) The other 4 gets dealt to a different player with probability 45 , since there are 5 slots
remaining for the 5 cards left in the deck. Now the deck is left with a 2, and 3, and the pair of 5’s, and
the probability that the pair doesn’t get dealt to the remaining person is 56 . Therefore, the probability
of no pair is 67 · 45 · 56 = 47 .