It seems like these are happening more and more often. Kasparov is playing one of the world's best computers at a chess tournament in NYC. I've been to at least one game of the last three of these tournaments I think. And watched most of the rest online. Since my schedule is pretty free now, I just signed up for all four games of this tournament. Tickets are free, first come first serve.
Kasparov vs. X3D Fritz
Now, I actually suck at chess, and don't understand most of what happens in these games. But watching them and the commentary surrounding them is fun and educational. But I must admit, I am mainly there to root for the computer. I think it is only a matter of time before the puny humans are routinely crushed, and hope to see the day that happens, and Kasp can't even eek out one win. We probably are not there yet, but soon perhaps. Soon...
I do have a couple beefs with how the computer teams have been run int he past though. Here they are:
#1) While the computer chooses its own moves, the TEAM of human programmers and handlers has always decided when to offer resignations and draws, and has decided if they shuld accept draw offers. This is just wrong. The computer should have to make these decisions itself. This is a part of the game, and if the computer is truely to be the one playing, humans should not be involved in ANY decisions.
#2) Even if the computer makes the decision, it should remember that it is a computer. It will see any mistakes immediately, and not make any obvious ones. The human can. Even the best human in the world. As a strategy to WIN the computer should NEVER resign, offer a draw, or accept a draw. It should force each game to conclude via the rules only. Checkmate, stalemate, draw by repeated poisition, 50 move rule or lack of time. (Or I suppose the human can resign.) First, this would give the human many more opportunities to make a mistake. A mistake the computer could take advantage of. Two, such a strategy would fluster the hell out of the human and anger him (especially Kasparov with his temper). That would make the human MORE likely to make a mistake. Third, in an extended draw sequence, the human is far more likely to get into time pressure, again increasing the chances of an error, or even running out of time. If the computer truly wants to win, it needs to fully take advantage of its strengths... which includes giving the human every possible opportunity to make a mistake. Finally, there are a lot of chess newcomers watching these things. And seeing draws accepted early in the game all the time does not really help to encourage the game in the mind of those people. These games are going to be televised on ESPN! Give the people a good show darn it!
Anyway, both of those tick me off. There is no reason for a computer player to EVER agree to end a game before it has to end. Give the humans chances to make stupid mistakes! Yes, this shows some lack of respect for the opponant. But screw it. Go for the jugular. Feed Kasp the rope and let him hang himself!
Sam, I agree with #1 - the humans should have no involvement in those decisions.
I agree with most of #2... The computer should never resign. That is a gentlemanly thing to do, and the computer is a machine, not a gentleman.
But. If the computer is offered a draw, and believes it is at a disadvantage then it should accept. Sure humans can make mistakes, but humans at that level do it rarely enough that the computer is better off accepting a draw.
I don't know how you could logically set the computer to offer draws, because it seems that the computer should never offer a draw if it believes it has an advantage, and so then if it did offer one the human would know that the computer thinks it is at a disadvantage and would never accept the offer. Whatever.