Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Radio Shack 1650 vs Chess Tiger

Author: Christophe Theron

Date: 10:03:41 05/21/98

Go up one level in this thread


Hi Mark

Tiger plays:
46...c2 (time=0.43s, score=+15.35, time left=11.74s)


[Event  "Correspondence game"]
[Site   "Internet"]
[Date   "1998.02.07"]
[Round  "?"]
[White  "Radio Shack 1650 Level 7"]
[Black  "Chess Tiger K5-100Mhz, 256Kb hash, 1min/game"]
[Result "*"]

1. d4 d5 2. c4 e6 3. Nc3 Nf6 4. Bg5 Be7 5. Bxf6 Bxf6 6. e4 dxe4
7. Nge2 Nc6  8. d5 Ne7 9. dxe6 Qxd1+ 10. Rxd1 Bxe6 11. Nxe4 Bxb2
12. Rd2 Ba3 13. N2c3 Bc1 14. Rd1 Bf4 15. g3 Be5 16. Nd5 c6
17. Ne3 O-O 18. Bg2 Nf5 19. Nxf5 Bxf5 20. O-O Rad8 21. c5 Rfe8
22. Nd6 Bxd6 23. Rxd6 Rxd6 24. cxd6 Kf8 25. Rd1 Bd7 26. Kf1 b6
27. Re1 c5 28. Rxe8+ Kxe8 29. Ke1 b5 30. Be4 f5 31. Bd5 g6
32. Kd2 Kd8 33. Bg8 h6 34. Ke3 Bc8 35. Bf7 g5 36. f4 Kd7 37. fxg5 hxg5
38. h4 f4+ 39. Kf3 Bb7+ 40. Kg4 f3 41. Bg6 gxh4 42. Bd3 c4 43. Bf5+ Kxd6
44. Kxh4 f2 45. Bh3 c3 46. g4 c2 *



>Graham's question was interesting... After that I was wondering if some
>sort of algorithm be added to a program to play moves which delay the
>inevitable long enough to try to make the other computer run out on
>time.  This is obviously useful when programs play blitz.

Yes. The evaluation function should take into account the "distance from
win". Usually the eval only measure the probability of winning the game.
But this is not enough to ensure that the program is going to win. For
example, if the program sees the mate in X moves, the probability of
winning is 100%. There is no way to increase this value, so the program
could simply choose a move that leads to mate in X+1 or X+2 (this is
also 100% win) if we do nothing to speed up things.

That's why most programs use a ( mate_value - depth_to_mate ) to
evaluate a mate sequence.

For example, mate in 1 can be evaluated 9999 (10000-1). Mate in 2 can be
9998 (10000-2). This way, the program chooses the fastest way to the
mate.

This is easy to do for the mate case. Others cases are much harder to
handle. If we modify too much the eval to speed up an endgame, we could
weaken the program.


    Christophe



This page took 0.01 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.