The only way to win is not to play the game…
Here’s a new match. Deep Fritz against the current world champion, Vladimir Kramnik.
The situation is quite different in the world of computer technology, where time never stands still. Today even an off-the-shelf Core Duo processor can match the speed of the four-processor system on which Fritz was running in 2002. In the current match Kramnik will face a Dual Intel Core 2 Duo 5160 system which allows Deep Fritz – also improved from version seven to version ten – to calculate around eight million positions per second.
Note that 8 million positions/sec is a far cry from the 200 milion that Deep Blue achieved. Of course, Fritz is remarkable for different reasons. It’s a piece of commercial chess software on commodity hardware. The average american consumer could easily afford to buy both the software and the hardware. And it’s playing the current world chess champion!
In the continuing quest to see if humans can outpace their electronic creations, the humans have lost another, perhaps decisive, round.
A six-game chess match between Vladimir Kramnik of Russia, the world champion, and Deep Fritz, a souped-up version of commercially available chess software made by Chessbase, ended today in victory for the computer, which won the final game and clinched the match, 4 games to 2.
Mr. Kramnik fell behind in the match when he lost Game 2 by walking into a checkmate in one move with hardly any pieces remaining on the board, a mistake that ranks as one of the biggest in championship-level chess history. Needing a win today to tie the match, Mr. Kramnik took some chances, eventually lost a pawn, and was then outmaneuvered by the computer.
http://www.nytimes.com/2006/12/05/crosswords/chess/05cnd-chess.html
hi,
i think the new processor is the grapic processor for the chessengine’s with cuda or mainstream,
thats the new talk.
greetings Jan Commandeur
I still cannot accept that Deep Blue beat Gary Kasparov. This is why. The terms of the match were that two GMs would “escort” Deep Blue through the first 16 mjoves. Second, the Deep Blue programmers had access to Kasparov’s chess history which they used to prepare Deep Blue to play Kasparov. My point is that Deep Blue had been prepared to play Kasparov more so than chess. I wonder what the outcome would have been had Kasparov got snowed in at some airport and some other top class GM had sat in for him. The match would probably have been called off, or so I believe. Kasparov said that when he asked to see sample games from Deep Blue he was flatly refused. These are terms that no one would accept in World Championship play.
As a novice in this field, I have a question about the amount of freedom of choice in the most complex chess-playing programs. If there is a given configuration of pieces on the board, will such a program invariably make the same move? Or might it make different moves, as a human player might, depending on matters such as tendencies toward falling into traps or other characteristics of the opponent that have been observed in prior moves?
I always get somewhat irked when people claim that computers can’t do something the human mind can do, because the human mind is itself a computing device. Its based on a very different chemical system, but the brain is still composed of what can be broken down into logic circuits. All of the calculations the brain makes can be made by circuit boards, we just don’t know how yet.
Here is a link to a very interesting online chess game in Java. You play against the computer. But, this isn’t a typical, boring online game – this one shows all of the moves that the computer is thinking about while it is doing its thinking. You’d have to see it to know why I think it is so interesting…
http://archive.turbulence.org/spotlight/thinking/gallery.html
1 Like
At some point, it becomes a question of whether or not chess is ‘solvable’. It may well be that keeping up means simply not losing. Kind of like tic-tac-toe, a “solved” game. No matter how good you are at it, if you play someone else reasonably good, you’ll never beat them.
Nope. Combinatorial explosion.
Douglas Hofstadter’s comments in Scientific American following the Kasparov-Deep Blue matches (reprinted in his book Metamagical Themas) on the philosophical and cognitive-science significance of computer chess are still the last word on the subject.
The verdict that computers are the equal of human beings in chess could hardly be more official, which makes the caviling all the more pathetic. The excuses sometimes take this form: “Yes, but machines don’t play chess the way human beings play chess!” Or sometimes this: “What the machines do isn’t really playing chess at all.” Well, then, what would be really playing chess?
Computers have mastered us humans in Chess… lets move on to Wei-Chi (aka Go by the Japanese and Baduk by Koreans)!!!
http://en.wikipedia.org/wiki/Computer_chess
http://csdl2.computer.org/comp/mags/ex/2006/04/x4002.pdf
IEEE Spectrum: Cracking GO
http://www.spectrum.ieee.org/oct07/5552
Enjoy!
Sighris
Rybka 3 is the newest killer chess engine. It is vastly stronger than any human at any tournament time control (anything but postal). The strongest human was Kasparov at 2851 ELO. Rybka is at 3200 on a quad 2.4 Ghz machine. An Intel i7 overclocked to 4000 mhz is twice as fast and probably 50 elo stronger. A 400 Elo differential is utterly unsurmountable. The engine is so strong that some of the people who test engines have just been testing opening books rather than different engines because it is 100 elo stronger than any other chess engine. Coupled with the best book it gains yet another 80 Elo over the default book. It is doubtful Kasparov in his prime could win a single game in a 8 game match whole the machine would win 3/4 or so.
The Grandmasters do not even presume to play it in a fair public match. They have played several odds matches. It has won most of those.
1 Like
to the post about Rybka 3… computer ratings are not in the same rating board as human FIDE ratings… just to let you know rybka 3 has not won any odds matches with any of the top 100 players… alot of the info stated in your bulletin is false or made up… I do Think eventually computer wll be the better chess players only due to how eventually they will process insane amounnts of nodes and what not but humans have something special computer dont have at this point in time… MEMORY! in the movie man vrs machine by kasperov he explains his beliefs about computers never being better then humans due to humans memorization of games… But nothing is impossible with the minds of all the great scientists including giving computers memory… well now i hear they have programs that play with the same streangth as SHREDDER, Rybka 3, fruit, ect… but memorizes games played so that the programmers and techs can then after the game make changes… well i think eventually computers will keep improving and probably alot faster then the greatest human minds in chess… in the past 10 years household computers came from 200 megahertz processors with 2 gigabyte hardrives and 64mgb ram to Some alian ware computers that have a 1000 GB hardrives or a TB, 4 sticks of 2gb of ram, and 4 processors of 4GB each…
1 Like
In 1989 a lucked into beating a chess program. It did something pretty smart - it reversed the board and told me that it had won. I couldn’t convince any of my friends that I’d beat the game; it sure didn’t look like I had. If you can’t win, cheat!
Rybka and Stockfish and others are IMHO winning the debate of whether a “Type A” program is best.
There is a discussion in the www.bookup.com forum called “MyDBSM” which outlines a hybrid approach of solving chess openings. Using brute force to pick the first layer of candidates, analysis by these killer selective program is then used to pick the next layer of candidates.
The topic that shows how Rybka is used to analyze chess opening is called MSbyDM. The “DM” in MSbyDM stands for data mining.
No, I didn’t make this up. 
1 Like
I agree computers are stronger than most humans, especially in Game 30 like the Kasparov Vs Deep Blue match. In normal controls, such as 45/2, then 25/1, and even longer for championship matches, I think Kasparov’s positional ability might have overcome. Moreover, I wonder if the computer would have won if the operators would have accepted obvious draws, instead of making G.K. Sweat it out for many more tedious moves. A human would have agreed to the draw and rested for the next game. Moreover, the programmers/operators operated Deep Blue from a sealed room, so we really don’t know what went on in there. Humans can help a computer search out obviously positive lines to greater depth. But back to the draw thing, it’s been known since the beginning that computers don’t get scared, they don’t get tired, and they don’t get hangovers. Companies like Novag and Fidelity would enter them in the big, long chess tournaments and through that attrition, their rating was always inflated. My Super V.I.P. was rated 2000, but I doubt it would give a 1750 player much problem in 45/2 if he were fresh. Ultimately, I think type A will be the answer, brute force. In all honesty, strategy and positional play is merely a substitute for perfect tactics, something G.K. refused to see for the longest time, stating “Computers will never have a human’s positional ability”…they don’t have to.
1 Like
