Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Game 2, IBM Kasparov vs. Deep Blue, move 36.axb5

Author: Geo Disher

Date: 13:55:48 03/08/06

Go up one level in this thread


You may be wrong about Qb6 being the best move.  Rybka finds axb5 as best also
after 17 hours of eval.

New game
r1r1q1k1/6p1/p2b1p1p/1p1PpP2/PPp5/2P4P/R1B2QP1/R5K1 w - - 0 1

Analysis by Rybka 1.0 Beta 32-bit:

1.Qb6 Rd8 2.Be4 Be7 3.axb5 Rd6 4.Qa5 Bd8 5.Qa4 a5
  +-  (1.57)   Depth: 4   00:00:00
1.Qb6 Rd8 2.Be4 Be7 3.axb5 Rd6 4.Qa5 Bd8 5.Qa4 a5
  +-  (1.57)   Depth: 4   00:00:00
1.Qb6 Rd8 2.Be4 Be7 3.axb5 Rd6 4.Qa5 Bd8 5.Qa4 a5
  +-  (1.82)   Depth: 5   00:00:00
1.Qb6 Rd8 2.Be4 Be7 3.axb5 Rd6 4.Qa5 Bd8 5.Qa4 a5
  +-  (2.06)   Depth: 6   00:00:00  5kN
1.Qb6 Rd8 2.Be4 Be7 3.axb5 Rd6 4.Qa5 Bd8 5.Qa4 a5
  +-  (1.73)   Depth: 7   00:00:00  10kN
1.Qb6 Rd8 2.Be4 Be7 3.axb5 Rd6 4.Qa5 Bd8 5.Qa4 a5
  +-  (1.83)   Depth: 8   00:00:00  13kN
1.Qb6 Rd8 2.Be4 Be7 3.axb5 Rd6 4.Qa5 Bd8 5.Qa4 a5
  +-  (1.73)   Depth: 9   00:00:00  17kN
1.Qb6 Rd8 2.Be4 Be7 3.axb5 Rd6 4.Qa5 Bd8 5.Qa4 a5
  +-  (1.95)   Depth: 10   00:00:00  37kN
1.Qb6 Rd8 2.Be4 Be7 3.axb5 Rd6 4.Qa5 Bd8 5.Qa4 a5
  +-  (1.86)   Depth: 11   00:00:01  44kN
1.Qb6 Rd8 2.Be4 Be7 3.axb5 Rd6 4.Qa5 Bd8 5.Qa4 a5
  +-  (1.77)   Depth: 12   00:00:02  116kN
1.Qb6 Rd8 2.Be4 Be7 3.axb5 Rd6 4.Qa5 Bd8 5.Qa4 a5
  +-  (1.68)   Depth: 13   00:00:03  199kN
1.Qb6 Rd8 2.Be4 Be7 3.axb5 Rd6 4.Qa5 Bd8 5.Qa4 a5
  +-  (1.65)   Depth: 14   00:00:04  298kN
1.Qb6 Rd8 2.Be4 Be7 3.axb5 Rd6 4.Qa5 Bd8 5.Qa4 a5
  +-  (1.63)   Depth: 15   00:00:05  450kN
1.Qb6 Qe7 2.axb5 Rab8 3.Qxa6 e4 4.Qa7 Qe5 5.Qe3 Re8
  +-  (1.65)   Depth: 16   00:00:10  1369kN
1.Qb6 Qe7 2.axb5 Rab8 3.Qxa6 e4 4.Qa7 Qe5 5.Qe3 Re8
  +-  (1.65)   Depth: 17   00:00:46  5181kN
1.Qb6 Qe7 2.axb5 Rab8 3.Qxa6 e4 4.Qa7 Qe5 5.Qe3 Re8
  +-  (1.55)   Depth: 18   00:01:54  12780kN
1.Qb6 Qe7 2.axb5 Rab8 3.Qxa6 e4 4.Qa7 Qe5 5.Qe3 Re8
  +-  (1.75)   Depth: 19   00:03:47  26970kN
1.Qb6 Qe7 2.axb5 Rab8 3.Qxa6 e4 4.Qa7 Qe5 5.Qe3 Re8
  +-  (1.70)   Depth: 20   00:06:27  46469kN
1.Qb6 Qe7 2.axb5 Rab8 3.Qxa6 e4 4.Qa7 Qe5 5.Qe3 Re8
  +-  (1.59)   Depth: 21   00:12:58  89868kN
1.Qb6 Qe7 2.axb5 Rab8 3.Qxa6 e4 4.Qa7 Qe5 5.Qe3 Re8
  +-  (1.53)   Depth: 22   00:32:35  215285kN
1.Qb6 Qe7 2.axb5 Rab8 3.Qxa6
  ±  (1.39)   Depth: 23   01:11:34  454452kN
1.Qb6 Qe7 2.axb5
  +-  (1.52)   Depth: 24   02:27:32  937211kN
1.Qb6 Qe7
  ±  (1.26)   Depth: 25   07:55:22  2876850kN
1.axb5 axb5
  ±  (1.28)   Depth: 25   17:10:17  2876850kN

(,  08.03.2006)


On March 04, 2006 at 01:10:21, Uri Blass wrote:

>On March 04, 2006 at 00:50:36, Stuart Cracraft wrote:
>
>>On March 03, 2006 at 22:39:47, Lar Mader wrote:
>>
>>>Ok, I know this topic has been discussed a lot (beaten to death?!), but I am
>>>still curious about this move:
>>>
>>>Game 2, IBM Kasparov vs. Deep Blue Rematch
>>>Deep Blue played
>>>     36.axb5
>>>
>>>I know that the move 37.Be4 is a move that is found by some current chess
>>>engines.
>>>
>>>But this 36.axb5, to my knowledge, is never selected by current chess engines.
>>>
>>
>>Nothing for me but that doesn't say much considering my own program:
>>
>>.sd 99
>>maxdepth = 99 maxtime = 9999999 timed = 0
>>.ts
>>position file? [wac.epd] tests.epd
>># of test positions to test? 1
>>maxtime = 9999999
>>Interrupt current ply and return move at timeout
>>Testsuite: Tests/tests.epd 1 positions
>>*** Problem   Solution(s): axb5 (bm)
>>BR ** BR ** BQ ** BK **
>>** -- ** -- ** -- BP --
>>BP ** -- BB -- BP -- BP
>>** BP ** WP BP WP ** --
>>WP WP BP ** -- ** -- **
>>** -- WP -- ** -- ** WP
>>WR ** WB ** -- WQ WP **
>>WR -- ** -- ** -- WK --
>>mv 1 stage 0, white to move, computer plays white
>>hash=13ae1dd545f1642d
>>pawnhash=38fa81d61563fda8
>>0 0 0 0 0 0
>>Alpha=1457 Beta=2258 Maxdepth=99 MaxTime=9999999 xboard=0
>>Itr/Max Mv   Score   Time Nodes PV
>> 1/ 4< a4b5 1457  0.00       37 a4b5
>> 1/ 4  g1h2 288  0.01       82 g1h2
>> 2/ 6> g1h2 688  0.02      269 g1h2 e5e4 g2g3
>> 2/ 7  a4b5 689  0.02      388 a4b5 a6b5 a2a7
>> 3/ 9  a4b5 698  0.05     1010 a4b5 a6b5 a2a7 e8f8
>> 4/10  f2b6 349  0.10     2932 f2b6 d6c7 b6f2
>> 5/13  f2b6 629  0.25     8081 f2b6 d6c7 b6e6 e8e6 f5e6 b5a4 a2a4 c7b6 g1h2
>> 6/14  f2b6 720  0.54    18844 f2b6 d6c7 b6e6 e8e6 f5e6 c8d8 a4b5
>> 7/16  f2b6 848  1.74    60879 f2b6 d6b8 b6e6 e8e6 f5e6 b8a7 g1h2 c8f8 a4b5
>> 8/19  f2b6 1001  3.19   115760 f2b6 d6c7 b6e6 e8e6 f5e6 c7b6 g1h2 c8e8 a4b5
>>a6b5 a2a8
>> 9/21  f2b6 1175  8.05   308035 f2b6 e8e7 a4b5 c8b8 b6c6 b8c8 g1h2 c8e8
>>10/21  f2b6 1125 16.45   639593 f2b6 d6c7 b6b7 e8d8 a4b5 c8b8 b7c6 b8b5 a2a6
>>a8a6 a1a6 b5d5 c6c4
>>11/23  f2b6 1415 37.50  1488255 f2b6 d6c7 b6e6 e8e6 f5e6 a8b8 a4b5 a6b5 c2g6
>>c7d6 a2a7
>>12/27  f2b6 1450 91.88  3725097 f2b6 e8f8 a4b5 c8b8 b6a5 d6c5 b4c5 f8c5 g1h2
>>b8b5 a5a4 c5d5 a1d1
>>13/33  f2b6 1499 208.56  8436384 f2b6 d6c7 b6e6 e8e6 f5e6 a8b8
>>14/
>>
>>People have consistently underestimated Deep Bluer's evaluation function.
>>
>>Bob Hyatt can speak more to your concerns than any other person on this
>>forum.
>>
>>Stuart
>
>I think that Qb6 is the right move and deeper blue's move was not correct.
>
>Evaluation function is a lot about having the right parameters and not only
>about knowledge and it is known that the deep blue team had not enough time to
>tune their evaluation function so I see no reason to assume that deep blue's
>evaluation was better than the evaluation function of the programs of today.
>
>Uri



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.