Maths Olympiad Prep

Track / Stage 5 / 168 of 400 #1248 of 2444

Problem 1248

AIME late
Combinatorics Difficulty 5.3 Prove it HMMT November · United States · 2012

In a game of rock-paper-scissors with nn people, the following rules are used to determine a champion:

a. In a round, each person who has not been eliminated randomly chooses one of rock, paper, or scissors to play.

b. If at least one person plays rock, at least one person plays paper, and at least one person plays scissors, then the round is declared a tie and no one is eliminated. If everyone makes the same move, then the round is also declared a tie.

c. If exactly two moves are represented, then everyone who made the losing move is eliminated from playing in all further rounds (for example, in a game with 8 people, if 5 people play rock and 3 people play scissors, then the 3 who played scissors are eliminated).

d. The rounds continue until only one person has not been eliminated. That person is declared the champion and the game ends.

If a game begins with 4 people, what is the expected value of the number of rounds required for a champion to be determined?

This one wants a proof. Work it on paper, then read the official solution and mark yourself. Be honest about it: the record is only any use to you if it is.

Next problem →

Official solution

Solution:

Answer: 4514\frac{45}{14}

For each positive integer nn, let EnE_{n} denote the expected number of rounds required to determine a winner among nn people. Clearly, E1=0E_{1}=0. When n=2n=2, on the first move, there is a 13\frac{1}{3} probability that there is a tie, and a 23\frac{2}{3} probability that a winner is determined. In the first case, the expected number of additional rounds needed is exactly E2E_{2}; in the second, it is E1E_{1}. Therefore, we get the relation
E2=13(E2+1)+23(E1+1), E_{2}=\frac{1}{3}\left(E_{2}+1\right)+\frac{2}{3}\left(E_{1}+1\right),
from which it follows that E2=32E_{2}=\frac{3}{2}.

Next, if n=3n=3, with probability 19\frac{1}{9} there is only one distinct play among the three players, and with probability 627=29\frac{6}{27}=\frac{2}{9} all three players make different plays. In both of these cases, no players are eliminated. In all remaining situations, which occur with total probability 23\frac{2}{3}, two players make one play and the third makes a distinct play; with probability 13\frac{1}{3} two players are eliminated and with probability 13\frac{1}{3} one player is eliminated. This gives the relation
E3=13(E3+1)+13(E2+1)+13(E1+1), E_{3}=\frac{1}{3}\left(E_{3}+1\right)+\frac{1}{3}\left(E_{2}+1\right)+\frac{1}{3}\left(E_{1}+1\right),
from which we find that E3=94E_{3}=\frac{9}{4}.

Finally, suppose n=4n=4. With probability 127\frac{1}{27}, all four players make the same play, and with probability 36281=49\frac{3 \cdot 6 \cdot 2}{81}=\frac{4}{9}, two players make one play, and the other two players make the other two plays; in both cases no players are eliminated, with total probability 127+49=1327\frac{1}{27}+\frac{4}{9}=\frac{13}{27} over the two cases. With probability 6481=827\frac{6 \cdot 4}{81}=\frac{8}{27}, three players make one play and the fourth makes another; thus, there is a probability of 427\frac{4}{27} for exactly one player being eliminated and a probability of 427\frac{4}{27} of three players being eliminated.

Then, there is a remaining probability of 6381=29\frac{6 \cdot 3}{81}=\frac{2}{9}, two players make one play and the other two players make another. Similar analysis from before yields
E4=1327(E4+1)+427(E3+1)+29(E2+1)+427(E1+1) E_{4}=\frac{13}{27}\left(E_{4}+1\right)+\frac{4}{27}\left(E_{3}+1\right)+\frac{2}{9}\left(E_{2}+1\right)+\frac{4}{27}\left(E_{1}+1\right)
so it follows that E4=4514E_{4}=\frac{45}{14}.

Source: MathNet, licensed CC-BY-4.0. Statement reproduced verbatim; metadata (topic, difficulty, ordering) added by this project.