Author: Thorsten Czub
Date: 03:19:39 12/15/01
Go up one level in this thread
On December 14, 2001 at 18:29:41, Uri Blass wrote: hi Uri, first - thank you for your answer. second: would you please only quote the passage you want to talk about and NOT the whole text ? thank you. ok. >>You can maybe understand how i felt when i saw THIS ugly and strange game of >>fritz7, like fritz7 playing (with 2 joints) under drugs: [Event "Odyssey2001-Tournament"] [Site "k6-400, 40/120"] [Date "2001.12.14"] [Round "12"] [White "Fritz7"] [Black "Gromit 3.8.1"] [Result "0-1"] [ECO "A28"] [PlyCount "56"] [EventDate "2001.12.14"] 1. c4 Nf6 2. Nc3 e5 3. Nf3 Nc6 4. e3 Be7 5. a3 O-O 6. b4 d6 7. Qc2 Re8 8. Bb2 Bf8 9. d3 g6 10. Be2 Bg7 11. O-O Ng4 12. Nd5 Bf5 13. Nd2 Ne7 14. h3 Nh6 15. Nc3 Qd7 16. g4 Be6 17. g5 Nhf5 18. Nde4 Nh4 19. Nf6+ Bxf6 20. gxf6 Bxh3 21. fxe7 Bg2 22. f3 Qh3 23. Rfc1 Rxe7 24. Nd5 Re6 25. Kf2 c6 26. Nc7 Rf6 27. Ke1 Rc8 28. Kd2 Nxf3+ 0-1 >>I stopped the game here. I was completely shocked. >>why is fritz allowing black to completely invade whites king position (move >>22.Qh3) . After 23...Rxe7 fritz7 is reduced to death. >>How can fritz7 evaluate the position after 23...Rxe7 with (when you know my >>critics over the years you will laugh about this running gag) 0.04 in favour of >>white, playing Nd5 after 6'36" computation? >> is it mad ?? >>in the same position gromit says +0.74 for black, even deep fritz says +0.69 for >>black after a few minutes on the same machine. >>Programs with more knowledge, e.g. Chess System Tal 2.03 would evaluate the >>black moves with big advantage: >>18...Nh4 with +2.19, 20...Bxh3 with +2.87, 22...Qh3 with +2.80, and would have >>better played 23...f5 with +5.54 instead of Rxe7 from gromit. this is the game and the passage you want to talk about. you don't like the tal score of 5.54. Why ? Tal sees a big advantage and gives a high score. How much is a win worth ? 0.9 pawns ? 0.8 ? 1 pawn ? 2 pawns ? We talk about chess. what your "chess programs" do is they count the material and say: 0.8 for black. Is this chess ? I don't think so, your "chess programs" do not play chess. they emulate it, and they do it very weak. Why is a won position only worth the material difference ? Can you explain to me WHY a won position is ONLY worth the material difference in plein numbers ? and why you believe a chess program is a good chess program when it is able to subtract material A - material B. ???? my pocket calculator (running with solar-energy) is more precise , and cannot play chess, but can subtract too. Is my pocket-calculator a chess program ? can it play chess ? no-. And so fritz can’t. not in this game. Not in this position. And when you now find other programs , behaving as stupid, i am sure i can explain that those programs do also not play chess, but only subtract a and b. >+5.54? >tal seems to be crazy. >I do not undertand a reason that it is better than 23...Rxe7. >The opposite Rxe7 that brings the rook into the attack seems to me a better >move. This is YOUR opinion. Hiarcs and CSTal e.g. see it different. The question is: how to evaluate?? Black is a pawn in advance. This gives +1.0 for black. Now the programs subtrac that the white pawn on e7 can be takan and come to +2 pawns. Now they see that white can play Nd5 and can have a fork on c7. somehow they reduce the black score to +0,70 - 0.80 for black. But what about the king-position of the white king. What about the king attack black can launch ?? nothing ! CSTal indicates this. And you call the program crazy ?? the move is ok. Hiarcs would also play f5. so why is it wrong when a chess program evaluates a better position with +5 ? is there a rule, by john huston or a godlike rule that this position has to be evaluated +0.80 ?? >Here is yace's evaluation after 23...f5(p800 64 mbytes) >It can see only a small advantage for black and it searched for 45 minutes > >Fritz7 - Gromit 3.8.1 >[D]r3r1k1/ppp1P2p/3p2p1/4pp2/1PP4n/P1NPPP1q/1BQ1B1b1/R1R3K1 w - f6 0 1 > >Analysis by Yace 0.99.56: > >24.Kf2 Bxf3 > ± (0.78) Depth: 1 00:00:00 >24.Nb5 Rxe7 > ± (1.24) Depth: 1 00:00:00 >24.Nd5 Nxf3+ 25.Bxf3 Bxf3 26.Nxc7 > ± (1.26) Depth: 1 00:00:00 >24.Nd5 Nxf3+ 25.Bxf3 Bxf3 26.Nxc7 > ± (1.26) Depth: 2 00:00:00 >24.Nd5 Nxf3+ 25.Bxf3 Bxf3 26.Nxc7 > ± (1.26) Depth: 3 00:00:00 >24.Nd5 Nxf3+ 25.Bxf3 Bxf3 26.Nf6+ Kg7 27.Nxe8+ Kg8 > +- (1.66) Depth: 4 00:00:00 >24.Nd5 Nxf3+ 25.Bxf3 Bxf3 26.Nf6+ Kf7 27.Nxe8 > +- (1.77) Depth: 4 00:00:00 >24.Nd5 Qg3 25.Rf1 Bxf1+ 26.Kh1 > ± (1.37) Depth: 5 00:00:00 17kN >24.Nd5 Qg3 25.Rf1 Nxf3+ 26.Rxf3 Bxf3+ 27.Kf1 > ² (0.37) Depth: 5 00:00:00 17kN >24.Nd5 Qg3 25.Bxe5 dxe5 26.Nf6+ Kg7 27.Nxe8+ Rxe8 > = (0.11) Depth: 5 00:00:00 17kN >24.Kf2 Bxf3 25.Nb5 Qh2+ 26.Ke1 Qxe2+ 27.Qxe2 Ng2+ > = (0.12) Depth: 5 00:00:00 17kN >24.Kf2 Bxf3 25.Bf1 Qg4 26.Ke1 > ² (0.60) Depth: 5 00:00:00 17kN >24.Kf2 Bxf3 25.c5 Rxe7 26.cxd6 Rd7 27.Ke1 Rxd6 > ² (0.61) Depth: 6 00:00:00 147kN >24.Kf2 Bxf3 25.Bf1 Qh2+ 26.Ke1 Qg3+ 27.Qf2 Qxf2+ 28.Kxf2 Rxe7 > ² (0.58) Depth: 7 00:00:01 307kN >24.Kf2 Bxf3 25.Bf1 Qh2+ 26.Ke1 Qg3+ 27.Qf2 Qxf2+ 28.Kxf2 Rxe7 29.Ke1 f4 > ² (0.45) Depth: 8 00:00:05 1718kN >24.Kf2 f4 25.Ne4 Nf5 26.Ke1 Qh4+ 27.Kd2 fxe3+ 28.Kc3 Nd4 29.Qd1 Qxe7 > = (0.24) Depth: 9 00:00:17 5625kN >24.Kf2 f4 25.Ne4 Nf5 26.exf4 Qh4+ 27.Kg1 Qh1+ 28.Kf2 Qh4+ 29.Kg1 Qh1+ 30.Kf2 >Qh4+ > = (0.00) Depth: 10 00:00:43 14759kN >24.Kf2 f4 25.Ne4 Nf5 26.exf4 Qh4+ 27.Kg1 Qh1+ 28.Kf2 Qh4+ > = (0.00) Depth: 11 00:04:08 85555kN >24.Kf2 Bxf3 25.b5 f4 26.exf4 Nf5 27.Ke1 Nd4 28.Qb1 Qg3+ 29.Kd1 exf4 30.Kd2 > = (-0.18) Depth: 12 00:19:10 339931kN > >(Blass, Tel-aviv 15.12.2001) as far as i see it yace will play Kf2. After 24...Bxf3 it wants to play b5 ??????? 25.b5 loses the game in an instant. Your program searched for 45 minutes and move 3 of the main line produces a loser-move ??? after 45 minutes of computation ?? uri uri. Maybe you should buy a new machine or another program. Hiarcs says +2.17 when i input b5 and says 25...f4 (that was the sense of the good move f5 before) and cstal says +9.46 for f4 ! Uri - it seems you talk about computers. And not about chess. Your programs compute 45 minutes and they see a disadvantage of -0.18. and come with a main line that is dead ! uri - what you support is old paradigm of computerchess, where chess programs have NO idea about chess at all. I cannot support this. Sorry.
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.