Author: Fabio Barrettone
Date: 00:48:04 04/27/01
Yesterday I was making a game against Junior 6a and Gambit Tiger II in the pause for dinner at my job, for passing a bit of time and see an interesting game. Well it sure was. I put GT 2 with aggressive style (SS=6 as normal) and Junior 6a with "Favour sacrifice" setting on. This the game: Date: 26/4/2001 White: Junior 6a Black: Gambit Tiger 2.0 Opening: A56 Benoni Defense 1. d4 Nf6 2. c4 c5 3. d5 g6 4. Nc3 Bg7 5. e4 d6 6. Bd3 e6 7. h3 Na6 8. a3 O-O 9. Nf3 exd5 10. cxd5 Nc7 11. O-O b5 12. Bd2 Re8 13. Re1 Rb8 14. Bg5 h6 15. Bd2 a6 16. Na2 Nd7 17. Qc1 g5 18. Bc3 Bxc3 19. Nxc3 Qf6 20. b4 c4 21. Be2 Qg7 22. a4 Nf6 23. axb5 axb5 24. Bf1 Rb7 25. Ra3 Qg6 26. Qb1 g4 27. hxg4 Bxg4 28. Nd4 Bd7 29. Re3 Ng4 30. Rh3 Qf6 31. Nf3 Kh8 32. Qc1 Rg8 33. e5 Qf5 34. Rxh6+ Nxh6 35. Qxh6+ Qh7 36. Qxd6 Ne8 37. Qe7 Qh5 38. Nd4 Bc8 39. Qc5 Rc7 40. Nc6 f5 41. Ra7 Bb7 42. Nxb5 Rh7 43. f3 Qh4 44. Ra2 Qh2+ 45. Kf2 Qg3+ 46. Ke2 Rh1 47. Nbd4 Qh2 48. Nxf5 Bxc6 49. dxc6 Rxg2+ 50. Bxg2 Qxg2+ 51. Ke3 Qg5+ 52. Ke4 Re1+ 53. Ne3 Qh4+ 54. f4 Qh7+ 55. Kf3 Qh3+ 56. Kf2 Qh4+ 57. Kg2 Qxf4 58. Rf2 Qg5+ 59. Kf3 Qh5+ 60. Ng4. Gambit Tiger lost, but the problem is this: move 43: GT played Qh4? here. Looking at the game (since I was moving manually beetween the two programs with PB off) I wondered why GT 2 played that move. So I let it rethink to the position with more time (after the game) and suddenly GT2 finds the winning move 43 ... Qh1+!. This move wins the game. What is weird now is that at any time control, now, GT2 plays Qh1+, always, not ever thinking about Qh4? no more. I've restarted the computer (thinking that maybe it was storing in the hash the better move) but, alas, GT never plays Qh4? no more. I'm wondering why it played Qh4? in that game since it had plenty of time (about 15 minutes left in the clock). Now, as I said, if I repeat all the moves GT2 doesn't make no more the move Qh4? that loses the game, ever. It is this some sort of Positional learning? Strange, since Gt2 (and Junior 6) discovered much later that Qh4? was losing, or, what else? I'm currently starting to think that maybe GT could have pshicological troubles (as humans) in difficult positions ;-) (well, this would be interesting). I'm wondering why this strange behaviour. Curiosly, Fabio.
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.