Author: Robert Hyatt
Date: 09:31:13 08/26/98
Go up one level in this thread
On August 26, 1998 at 06:26:14, Thorsten Czub wrote: >On August 25, 1998 at 22:44:34, fca wrote: > >>>Only materialists do. >> >>Agreed. But many chess programs (being materialists or whatever) *do* take the >>poisoned pawn, because search is not enough. >> >>So, it is fair to ask why does Gandalf not play Qxg7?. > >It is no materialist. I you would have seen the game: > >ChessTiger 11.2 vs. Gandalf from Paderborn, last round > >[Event "7th IPCCC"] >[Site "Paderborn"] >[Date "1998.02.15"] >[Round "7"] >[White "ChessTiger 11.2"] >[Black "Gandalf3"] >[Result "0-1"] > >1. Nf3 d5 2. g3 e6 3. d4 Nd7 4. Bg2 f5 5. O-O Bd6 6. b3 Qe7 7. c4 c6 >8. a4 a5 9. e3 Ngf6 10. c5 Bc7 11. Re1 b6 12. cxb6 Bxb6 13. Ba3 c5 >14. Bf1 O-O 15. Nc3 Bb7 16. Rc1 Rac8 17. Bb5 e5 18. Re2 Qe6 19. Nxe5 Nxe5 >20. dxe5 Qxe5 21. Bb2 Qe6 22. Rec2 Ng4 23. Qe2 Qh6 24. h4 Rcd8 25. Nd1 d4 >26. exd4 cxd4 27. Qe7 Be4 28. Re2 Qf6 29. Bc4+ Kh8 30. Qxf6 Rxf6 31. Rd2 Ne5 >32. Be2 Rff8 33. Kh2 Bf3 34. Bb5 Bd5 35. Be2 f4 36. gxf4 Rxf4 37. Kg3 Rdf8 >38. Ba3 d3 39. Bh5 Rf3+ 40. Kh2 R8f5 41. Rc8+ Bg8 42. Be8 Bxf2 43. Bd6 R3f4 >44. Rxd3 Rxh4+ 45. Kg2 Nxd3 46. Nxf2 Nxf2 47. Bd7 Ne4 48. Bxf5 Nxd6 >49. Rxg8+ Kxg8 50. Be6+ 0-1 > >you would not believe that Gandalf is materialist. I have seen this game >live, and was able to study the evaluations of gandalf DURING the game/moves. >I can tell you as a witness that it has constructed this ATTACK by static >knowledge. The way it made its pieces attack the white king, although the centre >was blocked, and later opened it... so that the hidden pieces had the diagonals >and files to kill white, is typically for the FIGHT of a SEARCH versus a >KNOWLEDGE program. > >In the same way Gandalf or CSTal would plan an attack without ANY real chances >to see in the tree, just because they have static evaluation DOING the plan, in >the same way one can estimate HOW they would react in the game we discuss about >(Qxg7??). > >I have ordered Gandalf. I hope i will get it soon. You can order it in Steen >Suurballes company and maybe also later at gambit and other distributors. > >>I now rephrase: >>**Perhaps** (I do not know, I do not have Gandalf) it rejects Qxg7 (correctly >>so, of course) but **for the wrong reason**... (i.e. on average bad tuning, >>though in this position correct). > >There is no WRONG reason is you reject Qxg7. >White is a pawn up. CSTal e.g. evaluates Qxg7 in MINUS because black has >completely development advantages, and all white has it 2 pawns and a queen run >away from different attacking pieces, and a king that will not survive. > >There is no other reason. Program having king-safety terms will not play Qxg7... >Thats all. IMO this is a very very good game because it polarized and shows in >example how stupid todays fast and strong chess programs are, BECAUSE they try >to substitute knowledge with search. >+ Here I disagree. I have versions of Crafty that won't touch Qxg7... the ones playing in Paris are good examples... but they don't play that well either, because they get over-concerned about king safety... My interest in such positions would be for programs that don't play Qg7 when the attack on the g-file will work, but which do play it when the attack won't work. I'd bet that you can move a piece or two around, so that the g-file ends up being useless, and these same programs will still avoid taking the pawn. Now for the *wrong* reasons... IE it is possible to be lucky in this position and not take for the wrong reasons and avoid a problem, but then be unlucky in the next game by not taking when the pawn is free... or offering the opponent a pawn you don't think he will take, only to see him snap it up and win... So it isn't a breakthrough to not take such pawns, *unless* it is not taken for the right reasons. Here, I bet that is not the case... And I'm not saying the reasons for not taking are all wrong... but when such "intuition" is embedded into a program, it often leads to problems... because for every case it is wrong, another opponent will push you into such situations and let the error lead to a loss... If anyone is interested, we could take the position, and start subtly changing it and see how the programs fare then... > > >>And the point being: >>If I modify the position slightly (making it harder for a black bishop to come >>into play, say) so Qxg7 *is* then a correct move, will Gandalf (andor CSTal) >>still reject it once they see the WQ being driven away and a black rook >>controlling the g-file? > >Modify a position and i will set it up in CSTal Win95. >Look, in my tournament programs like zarkov, crafty, cstal, tiger, hiarcs > >or more precise : > > 1 CSTAL WIN95, (6) 4.0 8 13½ 3 > 2-6 CHESS TIGER 11.5, (12) 3.5 10 16 3 > JUNIOR5 BETA JUN.CTG, (5) 3.5 8½ 14 3 > CRAFTY 15.17, (22) 3.5(1) 8½ 13 3 > CHESS TIGER 11.2, (13) 3.5 7 12½ 2 > MCHESS 7.1, (1) 3.5 7 12 3 > 7-11 HIARCS6, (2) 3.0(1) 9 14½ 2 > ZARKOV 4.2C, (23) 3.0 8½ 13½ 3 > NIMZO98 PADERBORN, (14) 3.0 7 12½ 2 > GENIUS5, (8) 3.0 7 11 2 > WCHESS 1.04, (18) 3.0 5½ 11 2 > >play very very strong. If you study these programs, you will see that they all >found a way to implement knowledge into their programs. >Some more, some less. >All these programs are IMO on the right track. They don't trust search alone. >And they all play nice and good chess, or ?! >I guess most of them would reject Qxg7. >OR ? > > > >>Your views, please? >> >>:-) >> >>Kind regards >> >>fca
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.