Arimaa - Comparing Arimaa Challenge To Chess Challenges

Comparing Arimaa Challenge To Chess Challenges

It has been argued that a computer has beaten the world chess champion but not beaten the human in the Arimaa challenge because of six reasons:

1. Arimaa is a new game. Therefore, the number of programmers and amount of time devoted to computer Arimaa is much less than for computer chess. Computer chess had thousands more programmers and 40 more years than computer Arimaa. The later and smaller effort resulted in less and slower progress in computer Arimaa.
2. The rules for the Arimaa challenge required the computer to show a higher playing ability than the rules for the chess matches. In the Arimaa challenge, the computer must beat three human players in three matches. In the chess matches, the computer must win one match against one human player.
3. In the Arimaa challenge, the computer needs to score 2/3 of the total points to win. In chess matches, the computer needs to score more than 1/2 of the total points to win.
4. In the Arimaa challenge, the computer needs to win a qualification match. Then the human studied the computer games to find the computer’s weakness. In chess, there was no qualification match.
5. In the Arimaa challenge, the computer cannot be improved between games. In chess, the computer was improved between games.
6. In the Arimaa challenge, the rules reject powerful or custom made computers priced over $1,000. However, a powerful custom made computer beat the world chess champion.

However, the Arimaa community disputes this argument point by point. To the first point, Arimaa is a new game, so the playing community is still small and even the best players are not professional players and have only been playing the game for a few years. Thus the human players in the Arimaa challenge are much weaker than the human players in the chess challenge. The weakness of human players should make the Arimaa Challenge easier to conquer than chess, which compensates developers for having studied the problem for a shorter time.

The remaining five points compare the Arimaa Challenge only to Kasparov vs. Deep Blue, ignoring all other man vs. machine chess matches in which computers have prevailed. The chess match which can most closely be compared to the Arimaa challenge match is the Man vs Machine World Team Championship. In 2004 and 2005 a team of humans played against a team of computer opponents. In both years the computers won by wide margin. In 2005 all three humans lost, the computers won 2/3 of the total points, the chess engines were commercially available for the humans to study, and the machine hardware used was not a supercomputer, but rather comparable to hardware used in the Arimaa Challenge.

Man-vs.-machine chess matches since 2005 have shown increasing computer dominance. For example, the 2006 Deep Fritz vs. Vladimir Kramnik and 2007 Rybka vs. Jaan Ehlvest matches gave additional advantages to the human player, but the computers (running on commodity hardware) prevailed anyway.