|
< Earlier Kibitzing · PAGE 3 OF 4 ·
Later Kibitzing> |
| Dec-31-06 | | Happypuppet: <barbababa> An uneducated guess is that the slightest mistake by either player in a computer-computer match is exploited and taken to a full point by the other side. In human games mistakes that are equally subtle are probably frequently made by both sides, while losing or winning is caused by ones much more severe. |
 |
| Dec-31-06 | | actinia: I agree with <Happypuppet>. A human can get a +1.0 advantage but still draw or lose, but I imagine a computer would likely win with that kind of advantage. |
|
| Jan-02-07 | | SniperOnG7: Rybka rocks! What a machine!! :S |
|
| Jan-02-07 | | somitra: <SniperOnG7: Rybka rocks! What a machine!! :S> Slight modification : What an engine !!! (Hydra is a machine, Rybka is a software) |
|
| Jan-02-07 | | Rocafella: <somitra> I apologise on behalf of <SniperOnG7>'s foolish, idiotic mistake =) |
|
| Jan-02-07 | | 14 Dog Knight: Does anyone know why the strongest program in the world, the 1970's version of Chess Challenger, did not play? |
|
| Jan-02-07 | | Rocafella: Lowel. |
|
| Jan-03-07 | | strobane: Why didn't Fritz play? |
|
| Jan-03-07 | | Jole: the missing of tactics by humans may be important, especially after gaining an advantage where likely alot of time has been used, as is the fact that humans are better with the initiative so in the absense of blunders in the first moves, the position may look really even (also the players may not feel up to fighting it out etc) |
|
| Jan-03-07 | | gokuson: Fritz didnt play because she's tired and needed a rest after the match with Kramnik |
|
| Jan-03-07 | | CrystalFrost: I don't know why Fritz didn't play though he had almost 99.90% chance to win the tournament |
|
| Jan-03-07 | | Shajmaty: <strobane: Why didn't Fritz play?> As already posted, Fritz had nothing to win here, but a lot to lose. <CrystalFrost: I don't know why Fritz didn't play though he had almost 99.90% chance to win the tournament> You have to be kidding: Rybka 2.2 is, by far, better an engine than Fritz (even Deep 10)! |
|
| Jan-03-07 | | 14 Dog Knight: Deep Junior is pretty good...
|
|
| Jan-03-07 | | 14 Dog Knight: of course the greatest chess engine of all time is...the Turk |
|
| Jan-03-07 | | SniperOnG7: <Rocafella> lol...um...thanks. BTW: i didnt know hydra was an actual machine :< |
|
| Jan-03-07 | | s4life: <CrystalFrost: I don't know why Fritz didn't play though he had almost 99.90% chance to win the tournament> As they say, ignorance is bliss.. |
|
| Jan-04-07 | | code13: Fritz's people have said that Fritz is designed to play against humans. If they were designing software to play against other computers they would do it differently. |
|
| Jan-05-07 | | s4life: <code13: Fritz's people have said that Fritz is designed to play against humans. If they were designing software to play against other computers they would do it differently.> How are the two approaches different? Personally I think that statement is pure marketing, a load of BS. |
|
| Jan-05-07 | | Jakobicek: <s4life> there is definitely a lot of marketing and hype on the side of the Fritz developers nonetheless it is true that finetuning the evaluation function and mainly adapting the opening book for use against silicon might bring several elo-points to a program... i imagine it is basically the same as human preparation... if u are to play against a tactically skilled oponent u would like to reach positions that dont allow carnage whereas if the other guy is a master of endgame technique u gonna try to smash him before he can utilize his strengths |
|
| Jan-05-07 | | Jakobicek: that said... rybka is insanely strong... and this fritz finetuning i mentioned would give fritz very little hope |
|
| Jan-05-07 | | s4life: <Jakobicek: <s4life> there is definitely a lot of marketing and hype on the side of the Fritz developers nonetheless it is true that finetuning the evaluation function and mainly adapting the opening book for use against silicon might bring several elo-points to a program... i imagine it is basically the same as human preparation... if u are to play against a tactically skilled oponent u would like to reach positions that dont allow carnage whereas if the other guy is a master of endgame technique u gonna try to smash him before he can utilize his strengths> Still, Hydra didn't boast any 'special' features but it demolished (5.5 - 0.5) an elite master a year ago...nowadays computers are playing in a whole different league, just check the games Rybka played in the last championship. I think it makes little difference whether the comp has been tuned to play better certain types of positions or not, when the depth of analysis of current computers is reaching what some would call 'positional' understanding. |
|
| Jan-05-07 | | jackpawn: One thing I immediately noticed is that Black scored very well in this event. I've always thought that people overrate the advantage of having White. Perhaps this is a small confirmation of my theory?! |
|
| Jan-05-07 | | Happypuppet: <jackpawn> In human games I doubt the advantage White has is extremely overrated. As mentioned in my previous post playing a mistake against a computer has tremendous consequences compared to playing a mistake against a human. In this case I think that results in Black having more of a chance if he's a better player in a computer-computer match. In human-human games, on the other hand, Black often has to play for a draw even if he's a better player. |
|
| Jan-05-07 | | s4life: <happypuppet> before you jump to conclusions, I would like to see statistics in a larger set of games. I am lazy to go and check it, but I'd be pretty surprised if this trend (black > white) would also hold there. According to an old corollary of game theory there exist a sequence of moves in which white cannot lose no matter what black does.... but that would apply only if the game was solved. |
|
| Jan-07-07 | | percyblakeney: <<CrystalFrost: I don't know why Fritz didn't play though he had almost 99.90% chance to win the tournament> You have to be kidding: Rybka 2.2 is, by far, better an engine than Fritz (even Deep 10)!> Even the free Rybka 1.0 Beta version seems to be better than Deep Fritz 10: http://www.husvankempen.de/nunn/40_... |
|
|
|
< Earlier Kibitzing · PAGE 3 OF 4 ·
Later Kibitzing> |
