Author: Uri Blass
Date: 22:40:51 07/22/03
Go up one level in this thread
On July 22, 2003 at 17:38:29, Matthew Hull wrote: >On July 22, 2003 at 17:24:11, Uri Blass wrote: > >>On July 22, 2003 at 10:02:28, Matthew Hull wrote: >> >>>On July 22, 2003 at 00:08:51, Dann Corbit wrote: >>> >>>>On July 22, 2003 at 00:01:07, Matthew Hull wrote: >>>> >>>>>On July 21, 2003 at 23:29:11, Dann Corbit wrote: >>>>> >>>>>>On July 21, 2003 at 23:14:52, Derek Paquette wrote: >>>>>> >>>>>>>Ok here is a hypothetical situation for you all. >>>>>>>I love debating chess, and so here is something to debate. >>>>>>> >>>>>>>There was talk a few years ago of a program actually being able to play for the >>>>>>>world championship. While this isn't happening, let us pretend for the sake of >>>>>>>this debate that it is true. >>>>>>> >>>>>>>How much money would it take to build a machine and the salary of programmers to >>>>>>>win a world championship match outright, >>>>>>>so a point where it is embarasing for the Grandmasters >>>>>>> >>>>>>>no draws, all wins, no loses >>>>>>>Is this possible right now? How much money would it cost >>>>>> >>>>>>Way, way more than the reward in monetary terms. >>>>>> >>>>>>>The saying is, "money can't buy everything" >>>>>>>only most things, is this possible? >>>>>>> >>>>>>>In my own opinion yes. >>>>>>>No investment by any one or two people could possibly afford this, >>>>>>>However if a corporation were to invest millions, they could topple the best in >>>>>>>the world, thoroughly, >>>>>>> >>>>>>>my own opinion of course >>>>>> >>>>>>Possible? Maybe. Hsu/Campbell could shrink and improve the chips by a couple >>>>>>orders of magnitude. They could use 1 million of them instead of 480. They >>>>>>could use a cluster of top of the line RS/6000 machines and improve/debug the >>>>>>programs and hardware. >>>>>> >>>>>>Probably a cost of 100 million dollars. >>>>>> >>>>>>There is absolutely no way that's going to happen. >>>>>> >>>>>>Of course, 20 years from now your desktop PC will be able to do the same thing. >>>>>>So why not just wait a bit. >>>>> >>>>> >>>>>Don't need to spend all that money! Not even one cent more. >>>>> >>>>>I'd bet any of the top programs could win a championship on current hardware, >>>>>simply because of the human fatigue factor. The programs we have now would >>>>>wear-down any of the top players in a 12 game match, no problem. >>>> >>>>Without a single draw for the best player in the world? >>>> >>>>I don't think you read the OP's question. >>> >>> >>>You're right. They could not do it without draws or losses. But I'd bet even >>>crafty on current hardware could win a 12+ game match against any human, just by >>>fatigue factor alone. >>> >>>Matt >> >>I believe that Smirin is going to beat Crafty in a 12 game match. >>The fatigue factor is only a problem for kasparov and kramnik and the conclusion >>should be that sponsors should not pay them money for matches against computers. >> >>I do not believe that humans who play a match against another player become >>weaker in the last games if they have to play 12 games when they have 48 hours >>between every 2 games. >> >>There were a lot of matches between humans of more than 12 games. >>Karpov was leading 5-0 against kasparov and could not win the match because of >>the fatique factor but it did not happen after less than 12 games but after more >>than 24 games. > >However, there seems to be a psychological fatigue when playing strong computers >for several hours at a stretch...on just one game. There is definitely a >physical fatigue for the human in that situation, and zero for the computer. In >a match between humans, this is not an issue since both players tire at about >the same rate. > >But grinding out a match with a strong computer, even every two days is probably >psychologically wearing over time. That combined with the physical wear of a >five hour game becomes quite noticable, since one player has "infinite" stamina >and the other does not. It's just not an issue in human/human matches. > >Do you not think so? > >Matt I do not believe in it and I think that humans can play at a constant level for 12 games. Uri
This page took 0 seconds to execute
Last modified: Thu, 15 Apr 21 08:11:13 -0700
Current Computer Chess Club Forums at Talkchess. This site by Sean Mintz.