Maths Olympiad Prep

Library / Taiwan IMO Selection Camp /243 of 359

Taiwan IMO Selection Camp

Number theory Difficulty 6.3 National Olympiad Prove it Taiwan

Let pp be a prime. In the following game, Ed and Alfie take turns to play, each taking his own turn. On a person's turn, he first chooses an ii from {0,1,…,p−1}\{0, 1, \dots, p-1\} that has not yet been chosen by either side, and then chooses an element aia_i from {0,1,2,…,9}\{0, 1, 2, \dots, 9\}. Ed goes first. Once all of {0,1,…,p−1}\{0, 1, \dots, p-1\} have been chosen, the game ends and the following number is computed
M=a0+10×a1+⋯+10p−1×ap−1=∑j=0p−1aj×10j. M = a_0 + 10 \times a_1 + \cdots + 10^{p-1} \times a_{p-1} = \sum_{j=0}^{p-1} a_j \times 10^j.
If MM is divisible by pp, Ed wins; otherwise, Alfie wins.
Prove that Ed has a winning strategy.

Solution

We say that a player makes the move (i,ai)(i, a_i) if he chooses the index ii and then the element aia_i of the {0,1,…,9}\{0, 1, \dots, 9\} in this move.

(1) If p=2p=2 or p=5p=5, then the first player simply chooses (0,0)(0, 0), which forces 10∣M10|M, which guarantees he wins.

(2) So assume p∉{2,5}p \notin \{2, 5\}. Let the first player choose (p−1,0)(p-1, 0) in the first move. By Fermat's Little Theorem, (10(p−1)/2)2=10p−1≡1(modp)(10^{(p-1)/2})^2 = 10^{p-1} \equiv 1 \pmod p, so p∣(10(p−1)/2)2−1=(10(p−1)/2−1)(10(p−1)/2+1)p|(10^{(p-1)/2})^2 - 1 = (10^{(p-1)/2} - 1)(10^{(p-1)/2} + 1). Since pp is prime, we are left with two cases:

(i) p∣10(p−1)/2−1p|10^{(p-1)/2} - 1

In this case, for each move (i,ai)(i, a_i) made by the second player, the first player immediately makes a move
(j,aj)={(i+p−12,ai),0≤i≤p−32;(i−p−12,ai),p−12≤i≤p−2. (j, a_j) = \begin{cases} (i + \frac{p-1}{2}, a_i), & 0 \le i \le \frac{p-3}{2}; \\ (i - \frac{p-1}{2}, a_i), & \frac{p-1}{2} \le i \le p-2. \end{cases}

This makes the sum be divisible by pp, thus wins the game. One can easily check that this move is always possible.

(ii) p∣10(p−1)2+1p|10^{\frac{(p-1)}{2}} + 1

In this case, for each move (i,ai)(i, a_i) made by the second player, the first player immediately makes a move
(j,aj)={(i+p−12,9−ai),0≤i≤p−32(i−p−12,9−ai),p−12≤i≤p−2. (j, a_j) = \begin{cases} (i + \frac{p-1}{2}, 9 - a_i), & 0 \le i \le \frac{p-3}{2} \\ (i - \frac{p-1}{2}, 9 - a_i), & \frac{p-1}{2} \le i \le p-2. \end{cases}

In any case, we will have 10j≡10i(modp)10^j \equiv 10^i \pmod p, and hence, aj10j+ai10i≡(aj+ai)10i=9×10i(modp)a_j 10^j + a_i 10^i \equiv (a_j + a_i) 10^i = 9 \times 10^i \pmod p. Therefore, at the end of the game, MM would be congruent to
∑i=0(p−3)/29×10i=10(p−1)2−1≡0(modp). \sum_{i=0}^{(p-3)/2} 9 \times 10^i = 10^{\frac{(p-1)}{2}} - 1 \equiv 0 \pmod{p}.
This ensures that the first player wins. Again, one can easily check that this move is always possible.

Want a route through all this instead of an archive? The track puts 2,604 problems in a working order, from Junior Challenge level to the IMO shortlist.

Source: MathNet, licensed CC-BY-4.0. Statement translated into English from zh; metadata (topic, difficulty) added by this project.