Regarding this week’s man vs machine chess match. Tyler writes:
I was surprised to see Kasparov favored. Once he lost to Deep Blue, the last big match (Kramnik vs. Deep Fritz) was a draw. I know it is not as simple as Moore’s Law, but hey, don’t these machines improve their game more rapidly than the human players do?
Indeed, they do. Deep Blue was a very expensive, very fast computer specialized for chess and capable of examining some 200 million positions per second. Fritz is no slouch but it is being run on a more or less ordinary four processor Xeon computer capable of analyzing 3 million moves a second. Fritz is thus about 70 times less powerful than Deep Blue. Yet because of improvements in algorithms, Fritz is almost certainly the better player. So the computer players have improved tremendously over the past 6 years – so much so that the computer side is no longer bothering to field its best against the weak humans!
Addendum: Here’s a graph, from Jeff Sonas, of the top computer player ratings versus the top human player. Although, it is true, as Sonas argues, that the best programs do not beat the very best human players but only draw consistently it may simply be inherent in chess that it is impossible or near impossible to beat someone who is playing at the very highest level (similarly it’s impossible for even a mediocre human player to lose in tic-tac-toe.) We should not conclude from this that the computers aren’t improving. Thanks to Nathan Stocker for the link.