At the beginning of this simple game, the applet below displays a board with two numbers. At any time you can use the edit control to input a positive difference of any two numbers already present on the board. To do that, type in a number and press Enter. The loser is the player unable to make a move. If you wish the computer to move first, check the box "Please start".
Assume the game is over - it's impossible to add a new difference. Let a be the smallest number present. Then the collection of the numbers on the board coincides with the set A of
all multiples of a not exceeding the largest of M and N. The proof is very similar to the way
we established a fundamental property of gcd.
First of all, both a|N and a|M. If, for example, N = na + b, then we could form differences N - a, N - 2a, etc. and eventually get on the board b < a which would contradict the minimality of a. Therefore, for some n and m, N = na and M = ma. So what we have on the board is A = {ia: i = 1,... , max(n, m)}. Now recollect that a|gcd(N, M). On the other hand, as we already remarked,
the difference of any two numbers is divisible by their gcd. Therefore, and number on the board
is divisible by gcd(N, M). In particular, gcd(N, M)|a. Therefore, a = gcd(N, M).
Another way to explain this stems directly from the Euclid's algorithm. Assuming N > M, the first step is to form N = sM + r.
Note that, sooner or later, r would appear on the board since it could be obtained by repeatedly
subtracting M first from N and then from thus obtained numbers. The second step in the
algorithm is to continue with M = tr + u. Having both M and r on the board and proceeding with taking differences we will eventually get u as well. Clearly we can now
continue in this manner until the algorithms stops and gcd(N,M) has been found.
|Contact|
|Front page|
|Contents|
|Algebra|
|Up|
|Store|
Copyright © 1996-2012 Alexander Bogomolny