What's new , what's hot , and what's cutting-edge in chess ?
Create topic | To list
Hi Guys! As far as we all know, todays engines are better (much-much better) in strenght, than Deep Blue in the 1997's. However Kramnik and Kasparov in their recent matches with Fritz and Deep Junior respectivly, didn't lost! So, how do you explain Kasparov loss with Deep Blue? Was this match a fake? Opinions please! Let us attempt to resolve the mystery!
As far as we all know, todays engines are better (much-much better) in strenght, than Deep Blue in the 1997's. However Kramnik and Kasparov in their recent matches with Fritz and Deep Junior respectivly, didn't lost! So, how do you explain Kasparov loss with Deep Blue? Was this match a fake? Opinions please! Let us attempt to resolve the mystery!
For one thing Deep blue was not an engine in the normal sense but was , like Hydra is , an inextricable fusion of software and hardware . Its hardware just can not be compared with the hardware of a normal dual core . If the chess playing part of its algorithm were to be translated to uci or chessbase protocol and run on a quad it would't give Kasparov much trouble . The second thing is that it was after all a short match and also Kasparov did not adopt the best possible approach to the match . Azaad
The conditions were unfavourable for Kasparov as he had to play every day, and his mental balance was not normal after the second game. But I do not think these are the real reasons for the result. In my judgement the evidence points strongly to the possibility of Kasparov playing not Deep Blue but a Centaur. Kasparov won the first game easily in the manner he had won his match against DB the previous year. DB looked like a very ordinary program, rather like the one an early Fritz had defeated in the 1995 WCCC. Then came the infamous second game. Kasparov had an inferior position but played a move he knew from experience would cause the computer to play a move leading to a draw because the later position Kasparov foresaw was over any computer's horizon. After an exceptionally long pause DB played a move that strong GMs would play. Later in the same game we had a situation where DB played a move (not forced) that would have permitted Kasparov to secure a stalemate had he spotted the correct sequence. The question is "How could a computer program play a move such as this???" Kasparov could miss a tactical forced sequence, but nor a computer! Kasparov smelt a rat and played the rest of the match in an uncharacteristic fashion, particularly the final and deciive game where Kasparov played an inferior line of the Caro-Kann, an opening he never normally plays. In addition tto the evidence of the second game we must note the following: 1. IBM were desperate to win the match. Victory led to a massive 15% rise in the value of its shares. 2. The DB team included an assortment of strong GMs such as Joel Benjamin. 3. The program looked like a totally different program in the second game to the one that played in the first game. The DB team said they made "adjustments" but our experience of computer programs before and since suggests this is very unlikely to explain such a difference in a single day. 4. The computer was kept in a separate room and both Kasparov's aides and independent observers were denied access to the computer and its operation at all times. 5. The DB team refused to publish the computer logs showing DB's evaluations. For me this is the clincher. What possible reason could IBM have for refusing even to this day to publish the logs? After all the logs are the only way that IBM could convincingly refute - or confirm - Kasparov's suspicions that he was playing a Centaur. Dave
As far as we all know, todays engines are better (much-much better) in strenght, than Deep Blue in the 1997's. However Kramnik and Kasparov in their recent matches with Fritz and Deep Junior respectivly, didn't lost! So, how do you explain Kasparov loss with Deep Blue? Was this match a fake? Opinions please! Let us attempt to resolve the mystery! The conditions were unfavourable for Kasparov as he had to play every day, and his mental balance was not normal after the second game. But I do not think these are the real reasons for the result. In my judgement the evidence points strongly to the possibility of Kasparov playing not Deep Blue but a Centaur. Kasparov won the first game easily in the manner he had won his match against DB the previous year. DB looked like a very ordinary program, rather like the one an early Fritz had defeated in the 1995 WCCC. Then came the infamous second game. Kasparov had an inferior position but played a move he knew from experience would cause the computer to play a move leading to a draw because the later position Kasparov foresaw was over any computer's horizon. After an exceptionally long pause DB played a move that strong GMs would play. Later in the same game we had a situation where DB played a move (not forced) that would have permitted Kasparov to secure a stalemate had he spotted the correct sequence. The question is "How could a computer program play a move such as this???" Kasparov could miss a tactical forced sequence, but nor a computer! Kasparov smelt a rat and played the rest of the match in an uncharacteristic fashion, particularly the final and deciive game where Kasparov played an inferior line of the Caro-Kann, an opening he never normally plays. In addition tto the evidence of the second game we must note the following: 1. IBM were desperate to win the match. Victory led to a massive 15% rise in the value of its shares. 2. The DB team included an assortment of strong GMs such as Joel Benjamin. 3. The program looked like a totally different program in the second game to the one that played in the first game. The DB team said they made "adjustments" but our experience of computer programs before and since suggests this is very unlikely to explain such a difference in a single day. 4. The computer was kept in a separate room and both Kasparov's aides and independent observers were denied access to the computer and its operation at all times. 5. The DB team refused to publish the computer logs showing DB's evaluations. For me this is the clincher. What possible reason could IBM have for refusing even to this day to publish the logs? After all the logs are the only way that IBM could convincingly refute - or confirm - Kasparov's suspicions that he was playing a Centaur. Dave
Dave , I never earlier considered the possibility of such a massive fraud but I have always been worried by your point 5. Now when you put it this way I wonder . . . Azaad
Two questions Dave . One slightly and the other completely OT. 1.Why do almost all chessbase engine vs super GM matches end in tie . 2. Can you find the name of the British player ( actually more of a trainer /administrator ) who gave a simultaneous performance in USSR just after the war and went on to lose almost all his games . I tried to get it by googling but could find only the name of another player (Kreicek) who did even worse (0/25 ) . You know how it is . You forget something and you can't get it out of your mind . Azaad
I spoke many times with people about match Kasparov-Deep Blue in 1997 with my friends (GM and IM).That was very strange.First game Kasparov won very easily.Some say even;"he plays like with child".Kasparov show in this game his class.In the second game (Ruy Lopez-Smyslov's Variation)position black become weak move by move.Endgame still could be finish in draw.But unexpectedly Kasparov resign.Even ordinary Fritz 5 on Pentium 120 Mhz(in this time that was best program in the world, and one of the best hardare)show after 2 minutes simply solve in external situation.Kasparov after second game admit that really overlook rescue move.After draw games become game number six.Quite passive opening for black (Kasparov as black played Caro-Kann)and in the opening phase big big mistake.My friend IM A.Sznapik told me:No kids,in order to the Kasparov didn't know theory here" He means Kasparov's blunder in the early stage of opening.This defect is known even players under 2000 Elo FIDE.Is include in all Caro-Kann's books.Kasparov just gave to understanding, that reason defeat in this game(but consequence in all match)is simply.Money,money and once again money.There are no other explanation in this situation.A lot of great players told the same! SAD BUT TRUE. Deep Blue mark about 2770 Elo FIDE, but this is not true.Later Karpov draw twice and told about Deep Blue:"Nothing great it does not play by it" Obviously in 1997 and 1998 Deep Blue played with top comercial programs, even with Fritz on Pentium 90.We can find this in Megadatabase.But Deep Blue's results with top programs was quite poor.Hydra also show power but only on the fast FPGA processors(now 64 and calculate much more than 200 million positions per second).Really difficult to say witch program is best on the same hardware. Greeting IM from Poland
IBM have published all the log files for the last Kasparov v Deep Blue match, they are still on their website and have been publicly available for ages!
Yes - all the IBM Deep Blue log files are available here... http://www.research.ibm.com/deepblue/watch/html/c....
Yes they published the logs - but why did it take them 3 years to do it ? Enough time to fabricate anything . What Dave meant of course was the IBM did not publish the logs shortly after the match even though Kasparov had hinted at human assistance to Deep Blue and by publishing the logs the IBM could have taken wind out of his sails but they chose not to do it . Why ? Azaad
Deep Blue is a Hoax,no wonder why it's top secret.
Yes they published the logs - but why did it take them 3 years to do it ? Enough time to fabricate anything . What Dave meant of course was the IBM did not publish the logs beveg the match even though Kasparov had hinted at human assistance to Deep Blue and by publishing the logs the IBM could have taken wind out of his sails but they chose not to do it . Why ? Azaad
Azaad, Yes. I knew that the logs were not published after the match and had not been published a couple of years later though I had not realised that a version, real or otherwise, was published three years after the event. However, as you say it was imperative to have the logs immediately after the match in order to be reasonably sure of their legitimacy. Could you imagine a match today between, say, Kramnik and Fritz in which ChessBase hid its computer away and refused to allow access to computer evaluations for three years? Dave
Do not forget also that this match was difficult for Kasparov for the mere fact that he had no idea how Deep Blue played. Deep Blue was kept in secret and Kasparov had no idea how to prepare for the match since he did not know the playing style of Deep Blue. Compare this to the recent matches between Junior and Kasparov, as well as Fritz and Kramnik, where both Kasparov and Kramnik got copies of the program monthes ahead and were able to prepare.
Do not forget also that this match was difficult for Kasparov for the mere fact that he had no idea how Deep Blue played. Deep Blue was kept in secret and Kasparov had no idea how to prepare for the match since he did not know the playing style of Deep Blue. Compare this to the recent matches between Junior and Kasparov, as well as Fritz and Kramnik, where both Kasparov and Kramnik got copies of the program months ahead and were able to prepare.
In the text you can use Wiki or HTML tags.
