Author: Robert Hyatt
Date: 08:15:08 07/23/99
Go up one level in this thread
On July 23, 1999 at 08:27:47, Jeroen van Dorp wrote:
>At this very very bad 5,0 game of mine on Chess.net I was analyzing move 37.f3?
>and playing until I came at 44...Kxf3. I thought it to be a logical move -that's
>what you'll get with a patzer like me- but analysis afterwards with Hiarcs 7.32
>(which found some other stupid analysis of me :-( ) showed almost instantly mate
>in 23 moves.
>
>As it isn't a tablebase situation I'm astonished again and again by the
>possibilities of these chess programs.
>If you stronger chess players out there view this situation, would you think
>it's possible as a human to rate this as a mate in about 20 moves or so?
you are not understanding tablebases. This is definitely a tablebase mate.
Here is the output from crafty on my PII/300 notebook:
clearing hash tables
time surplus 0.00 time limit 166:39 (166:39)
s depth time score variation (1)
1 0.00 2.47 1. Nc3
1 0.00 3.59 1. Nd2+ Ke3 2. Nxc4+
1-> 0.00 3.59 1. Nd2+ Ke3 2. Nxc4+
2 0.00 3.69 1. Nd2+ Ke2 2. Nxc4
2-> 0.00 3.69 1. Nd2+ Ke2 2. Nxc4
3 0.00 ++ 1. Nd2+!!
3 0.00 4.33 1. Nd2+ Ke3 2. Nxc4+ Kd4 3. Nxa5
3-> 0.13 4.33 1. Nd2+ Ke3 2. Nxc4+ Kd4 3. Nxa5
4 0.14 4.28 1. Nd2+ Ke3 2. Nxc4+ Kd4 3. Nxa5 Kc5
4-> 0.14 4.28 1. Nd2+ Ke3 2. Nxc4+ Kd4 3. Nxa5 Kc5
5 0.29 4.48 1. Nd2+ Ke3 2. Nxc4+ Kd4 3. Nxa5 Kc5
4. Kg5
5-> 0.30 4.48 1. Nd2+ Ke3 2. Nxc4+ Kd4 3. Nxa5 Kc5
4. Kg5
6 0.49 ++ 1. Nd2+!!
6 0.52 Mat23 1. Nd2+ Ke2 2. Nxc4 Kd3 3. Nxa5 <EGTB>
6-> 0.53 Mat23 1. Nd2+ Ke2 2. Nxc4 Kd3 3. Nxa5 <EGTB>
You seem to be assuming that for a position to be a tablebase mate, you have
to start off with 5 pieces or less. This is wrong. You have to _end up_ with
5 pieces are less. Notice that after the first capture this is an instant mate
(the following comes from crafty on my quad xeon which has _all_ of the 5 piece
files):
clearing hash tables
time surplus 0.00 time limit 30.00 (3:00)
s depth time score variation (1)
1 0.00 2.47 1. Nc3
1 0.00 3.59 1. Nd2+ Ke3 2. Nxc4+
1-> 0.00 3.59 1. Nd2+ Ke3 2. Nxc4+
2 0.00 3.69 1. Nd2+ Ke2 2. Nxc4
2-> 0.00 3.69 1. Nd2+ Ke2 2. Nxc4
3 0.00 ++ 1. Nd2+!!
3 0.00 4.33 1. Nd2+ Ke3 2. Nxc4+ Kd4 3. Nxa5
3-> 0.04 4.33 1. Nd2+ Ke3 2. Nxc4+ Kd4 3. Nxa5
4 0.10 ++ 1. Nd2+!!
4 0.10 Mat23 1. Nd2+ Ke2 2. Nxc4 <EGTB>
4-> 0.11 Mat23 1. Nd2+ Ke2 2. Nxc4 <EGTB>
my notebook requires 2 captures before it sees the mate because I don't have
the knpkp file (which is on the xeon) so that I have to see this getting to a
4 piece ending first.
This isn't remarkable at all... just an endgame database win...
>
>And if not, how does a program like Hiarcs or Fritz calculates that so fast?
>It didn't do a 23 ply deep brute force search. (so it's possible it overlooks an
>escape route, but....)
>In fact after some time it still gave 7/19 (I think that means 7 ply brute
>force, 19 ply selective search, correct me if'I'm wrong)
>
>This is the FEN
>8/8/8/p7/P1p1N2K/5k2/8/8 w - - 0 1
>
>and this is the game
>
>[Event "mmeics rated blitz game"]
>[Site "mmeics, Boston, MA USA"]
>[Date "1999.07.23"]
>[Round "?"]
>[White "****"]
>[Black "Jeroen"]
>[Result "0-1"]
>[ECO "B20"]
>[WhiteElo "1318"]
>[BlackElo "1414"]
>
>1. e4 c5 2. Bc4 Nc6 (2... e6 3. Qe2 Nc6 4. c3 Be7 5. d3 d5 6. Bb3 Nf6 7. Nf3
>O-O 8. O-O) 3. d3 Nf6 4. c3 g6 5. Nf3 Bg7 6. Na3 d5 7. exd5 Nxd5 8. Bxd5 Qxd5
>9. c4 Qd8 10. Ng5 O-O 11. Qf3 Bf5 12. g4 Qxd3 ? (12... Bxd3 -+) 13. gxf5 Qxf5
>14. Qxf5 gxf5 15. O-O Rad8 16. Be3 b6 ? (16... Bxb2 17. Rab1 Bxa3 -+) 17.
>Rad1 Bxb2 18. Rxd8 Rxd8 19. Bc1 Bxc1 20. Rxc1 Rd2 (20... h6 21. Nf3 Rd3 !) 21.
>Rc2 Rxc2 ? (21... Rd3 22. Nb5 h6 !) 22. Nxc2 Ne5 23. Ne3 f6 24. Ne6 Ng4 ?
>25. Nxg4 fxg4 26. Nc7 Kf7 27. Nb5 a6 28. Nc7 a5 29. Nd5 e5 30. Nxb6 f5 31. Nd5
>(31. Nd7 ! +-) 31... f4 32. Kg2 Ke6 33. h3 gxh3+ 34. Kxh3 e4 35. Nxf4+ Ke5
>36. Kg3 Kd4 37. f3 ? (37. Ne6+ Kxc4 38. Kf4 Kd5 39. Ng5 c4 40. Nxe4 h5 41. f3
>h4 42. Kg4 Ke5 43. Kxh4 Kf4 44. a4 Kxf3 ?? {Hiarcs 7.32 #23!!} 45. Nd2+ Ke3 46.
>Nxc4+ Ke4 47. Nxa5 Ke5 48. Nb7 Ke6 49. a5 Kd7 50. a6 Kc7 51. Nc5 Kb8 52. Kg4
>Ka8 53. Kf4 Kb8 54. Ke4 Ka7 55. Ke5 Ka8 56. Ne6 Kb8 57. Kd5 Kc8 58. Kd6 Kb8 59.
>Nc7 Ka7 60. Kc6 Kb8 61. Kb6 Kc8 62. a7 Kd8 63. a8=Q+ Ke7 64. Qf3 Kd6 65. Qf4+
>Kd7 66. Qf7+ Kd8 67. Qe8#) 37... exf3 38. Kxf3 Kxc4 39. Ne2 ? Kd3 40. Nc1+ Kd4
>41. Nb3+ Kd5 42. Nxa5 c4 43. Nb7 c3 44. Ke2 h5 45. a4 ? (45. Na5 Kc5 46. Nb3+
>+/=) 45... h4 46. a5 c2 47. Kd2 h3 48. a6 c1=Q+ 49. Kxc1 h2 50. a7 h1=Q+ 51.
>Kb2 Qh8+ 52. Kb3 Qa8 53. Na5 Qxa7 54. Kb4 Qb6+ 55. Ka4 Kd4 56. Nb3+ Kc4 (56...
>Kc3 57. Na1 Qb4#) 57. Ka3 Qxb3# {White checkmated} 0-1
>
>
>
>Thanks for your comment.
>
>Jeroen ;-}
>
>FYI I have a PII@400 oc'd to 450 (4x112) with 128 Mb
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.