On anecdotal evidence: Manyfaces on ""medium" time settings KGS = 2k (accounts manyfaces and manyfaces2) Manyfaces1 playing round 10 sec/move is able maintain 1d rank.
So by reducing oppponents thinking time bot gets relative advantage of 3stones. Also in chess it is uusually considered that humans gain more for longer thinking times. In blitz games chess computer have won worl champion long before in normal thiking times. Some people are more able than others to make move in few seconds, while it is pretty hard for the most. So yes computer do get better with longer thinking times but so do most of the human opponents. Comparing to pro's does not make sense currently as they still have rather superior skill compared to bots. Esspecially their tactical skill are way beyond, allowing them to play with just few seconds per move. And this is quite logical. Easier to see in chess. Game tree grows by factor of 36 on each ply so bot thatis not too smart doubling thinking times will gain a little but narrow searcher like human will gain a lot. I think in correpondence chess humans still hold against computers Petri 2009/10/29 Don Dailey <dailey....@gmail.com> > There is no question that computers play better at longer time controls > even though this has been disputed on this group. Is there any issues with > parallelism at short searches? In the "old days" when I competed in > computer chess with many processors, the program could out-search the > single processor version many times over at long enough time controls, but > the first few ply of search were quite a bit slower, so I would have been > better off using 1 CPU for speed chess games. > > What this meant of course is that at long time controls the CPU advantage > for the computer was exaggerated and it may have even been the case that a > human had a better chance at fast time controls in order to suppress the big > advantage of all those CPU's. I probably could have tuned some of this > effect away but we were not competing at short time controls. > > Is there anything like that going on? > > - Don > > > > 2009/10/29 Olivier Teytaud <olivier.teyt...@lri.fr> > >> Some elements around blitz: >> >> >> - My feeling that blitz games are harder for computers is based on our >> games >> against humans: we always lost games with short time settings. Even in >> 9x9, >> Motoki Noguchi or Pierre Audouard could win plenty of fast games, >> whilst >> playing strange openings for fun. This is for sure on a small sample. >> >> - The newspapers don't take into account or even report the difference >> between >> blitz games and standard games on the 29th of october, and they use the >> not >> very relevant complexity comparisons based on the number of possible >> boards >> or games. But they have nice photos for promoting computer-go :-) >> >> Best regards, >> Olivier >> >> >> Dear all (in particular for your question, Hideki!), please find enclosed >>> some newspapers about the games played on October 29th. Most of them are in >>> chinese. >>> >>> I don't read chinese, if some people can extract some elements... I'll >>> try to have some translations here with our chinese students. >>> >>> Best regards, >>> Olivier >>> >>> >> >> >> -- >> ========================================================= >> Olivier Teytaud (TAO-inria) olivier.teyt...@inria.fr >> Tel (33)169154231 / Fax (33)169156586 >> Equipe TAO (Inria-Futurs), LRI, UMR 8623(CNRS - Universite Paris-Sud), >> bat 490 Universite Paris-Sud 91405 Orsay Cedex France >> (one of the 56.5 % of french who did not vote for Sarkozy in 2007) >> >> >> >> _______________________________________________ >> computer-go mailing list >> computer-go@computer-go.org >> http://www.computer-go.org/mailman/listinfo/computer-go/ >> > > > _______________________________________________ > computer-go mailing list > computer-go@computer-go.org > http://www.computer-go.org/mailman/listinfo/computer-go/ >
_______________________________________________ computer-go mailing list computer-go@computer-go.org http://www.computer-go.org/mailman/listinfo/computer-go/
