Author: James T. Walker
Date: 11:09:17 12/23/00
Go up one level in this thread
On December 23, 2000 at 13:32:31, James T. Walker wrote:
>On December 23, 2000 at 12:43:58, Uri Blass wrote:
>
>>On December 23, 2000 at 09:59:54, James T. Walker wrote:
>>
>>>On December 22, 2000 at 23:55:46, Uri Blass wrote:
>>>
>>>>On December 22, 2000 at 20:06:11, James T. Walker wrote:
>>>>
>>>>>I recently made the statement that Shredder was not a "Blitzer". I'm issuing a
>>>>>recall of that statement. I just finished a 20 game match of Gambit Shredder vs
>>>>>Chess Tiger 13 (G/5minutes). Gambit Shredder won +8 -6 =6 (11-9). That does
>>>>>not necessarily mean Gambit Shredder is better than Chess Tiger but it does
>>>>>indicate it can play with the Tiger. Below is game 17 of the match in which
>>>>>Gambit Shredder announced "Mate in 67". Of course it's a tablebase win in a
>>>>>very difficult KBBKN endgame.
>>>>
>>>>The question is if it is not a draw by the 50 move rule against the right
>>>>defence.
>>>>
>>>>There are 2 questions:
>>>>1)Can shredder win against itself(does it capture the knight early enough to
>>>>avoid a draw)?
>>>>
>>>>2)Is there a win for white against the right defence?
>>>>
>>>>The 2 questions are different because the best move for the side with the
>>>>bishops may be not the shortest way to force mate when you consider the 50 move
>>>>rule.
>>>>
>>>>I guess that you can answer only 1 because I do not know of a program that use
>>>>the right KBB vs KN tablebases of distance to conversion and not distance to
>>>>mate.
>>>>
>>>>Uri
>>>
>>>Here is the position and the "Autoplay" of Shredder 5. So I believe this
>>>answers the question. I think since Shredder captures the Knight on its 50th
>>>move the game is won even under the 50 move rule which is not a mandatory rule
>>>if I understand correctly. I believe if you are playing computer/computer games
>>>with tablebases involved you would wave the 50 move rule. The only reason for
>>>the 50 move rule in the first place is that humans could play on forever and not
>>>make progress. This does not apply to computers with tablebases. I made the
>>>diagram with black on the move since white has just captured the pawn. This
>>>does bring up the question of when to start the count if using the 50 move rule.
>>> I assume it starts on whites next move after the capture (move 2 below). The
>>>Knight is captured on the 50th move if my count is correct.
>>>Jim
>>>
>>>[D]8/6B1/8/8/1k6/8/K7/3B4 b - - Mate in 67 for white
>>>
>>>[Event "5 Minutes/game"]
>>>[Site "?"]
>>>[Date "2000.12.23"]
>>>[Round "?"]
>>>[White "Gambit Shredder"]
>>>[Black "Chess Tiger 13"]
>>>[Result "*"]
>>>[SetUp "1"]
>>>[FEN "8/2n4B/8/8/K2B1k2/8/8/8 b - -"]
>>>
>>>1... Ne6 {-M66/4 0s} 2. Bb2 {+M66/4 0s} Ng5 {-M65/4 0s}
>>>3. Bc1+ {+M65/4 0s} Kg4 {-M64/4 0s} 4. Bd3 {+M64/4 0s} Ne6
>>>{-M63/4 0s} 5. Kb5 {+M63/4 0s} Nf4 {-M62/4 0s} 6. Bc2
>>>{+M62/4 0s} Kf3 {-M61/4 0s} 7. Kc4 {+M61/4 0s} Ne6 {-M60/4
>>>0s} 8. Bb2 {+M60/4 0s} Nf4 {-M59/4 0s} 9. Kd4 {+M59/4 0s}
>>>Kf2 {-M58/4 0s} 10. Ke5 {+M58/4 0s} Kf3 {-M57/4 0s}
>>>11. Be4+ {+M57/4 0s} Kg3 {-M56/4 0s} 12. Kf5 {+M56/4 0s}
>>>Ng2 {-M55/4 0s} 13. Be5+ {+M55/4 0s} Kf2 {-M54/4 0s}
>>>14. Bc7 {+M54/4 0s} Ne1 {-M53/4 0s} 15. Bb6+ {+M53/4 0s}
>>>Kg3 {-M52/4 0s} 16. Bd5 {+M52/4 0s} Ng2 {-M51/4 0s} 17. Kg5
>>>{+M51/4 0s} Ne1 {-M50/4 0s} 18. Bc7+ {+M50/4 0s} Kf2
>>>{-M49/4 0s} 19. Kg4 {+M49/4 0s} Ke3 {-M48/4 0s} 20. Bb6+
>>>{+M48/4 0s} Kd3 {-M47/4 0s} 21. Bf7 {+M47/4 0s} Nc2 {-M46/4
>>>0s} 22. Bg6+ {+M46/4 0s} Kc3 {-M45/4 0s} 23. Ba5+ {+M45/4
>>>0s} Kb3 {-M44/4 0s} 24. Kf4 {+M44/4 0s} Nb4 {-M43/4 0s}
>>>25. Bf7+ {+M43/4 0s} Ka4 {-M42/4 0s} 26. Bb6 {+M42/4 0s}
>>>Kb5 {-M41/4 0s} 27. Be3 {+M41/4 0s} Kc6 {-M40/4 0s} 28. Ke5
>>>{+M40/4 0s} Nd3+ {-M39/4 0s} 29. Kd4 {+M39/4 0s} Nc5
>>>{-M38/4 0s} 30. Bd5+ {+M38/4 0s} Kb5 {-M37/4 0s} 31. Bf2
>>>{+M37/4 0s} Kb4 {-M36/4 0s} 32. Bf7 {+M36/4 0s} Nb3+
>>>{-M35/4 0s} 33. Ke4 {+M35/4 0s} Nc5+ {-M34/4 0s} 34. Kd5
>>>{+M34/4 0s} Na4 {-M33/4 0s} 35. Kd4 {+M33/4 0s} Nc5 {-M32/4
>>>0s} 36. Be1+ {+M32/4 0s} Kb5 {-M31/4 0s} 37. Kd5 {+M31/4
>>>0s} Nb3 {-M30/4 0s} 38. Be8+ {+M30/4 0s} Kb6 {-M29/4 0s}
>>>39. Kd6 {+M29/4 0s} Ka6 {-M28/4 0s} 40. Bd7 {+M28/4 0s} Kb6
>>>{-M27/4 0s} 41. Bc3 {+M27/4 0s} Ka6 {-M26/4 0s} 42. Bg4
>>>{+M26/4 0s} Kb5 {-M25/4 0s} 43. Be2+ {+M25/4 0s} Ka4
>>>{-M24/4 0s} 44. Bd1 {+M24/4 0s} Ka3 {-M23/4 0s} 45. Kd5
>>>{+M23/4 0s} Ka2 {-M22/4 0s} 46. Kc4 {+M22/4 0s} Nc1 {-M21/4
>>>0s} 47. Bg7 {+M21/4 0s} Ka3 {-M20/4 0s} 48. Kc3 {+M20/4 0s}
>>>Na2+ {-M19/4 0s} 49. Kc2 {+M19/4 0s} Ka4 {-M18/4 0s}
>>
>>The last move is a big mistake.
>>
>>Black can draw by the 50 move rule with 49...Nb4+
>>
>>I think that the position with all the game is a good test position for chess
>>programs.
>>
>>The main question is if the program can find 49...Nb4+ 50.Kb1 Nd5.
>>
>>I am interesting if there is a program with tablebases that can find these
>>moves.
>>
>>I suspect that a lot of programs are going to do Shredder's mistake(49...Ka4)
>>and I suspect that a lot of programs are going to play 50...Nc2 that is the only
>>losing move by the rules of chess after 49...Nb4+ 50.Kb1(the 50 move rule is one
>>of the rules).
>>
>>Uri
>
>Hello Uri,
>You make a good point. But I still believe the 50 move rule can be waived
>before the start of any tournament.
>Anyway the "autoplay" function of Shredder is probably not a good way to test
>the position because I'm sure since it's switching sides each move it is losing
>the 50 move count. So if it were an actual game it would not play the Ka4 move
>but play a move to claim the 50 move draw. This brings up another question.
>Would Shredder play the exact same line if playing another opponent and saw the
>50 move rule coming? As you know sometimes there are several moves available
>that lead to mate in the same number of moves. Maybe in a real game Shredder
>would be forced to take one which avoids the 50 move rule and wins after all. I
>will have to see if I can make this happen against Fritz or some other program.
>Jim
Well I just tried it and Shredder plays the same moves in a game as it played in
the autoplay mode. I made a book out of the game using the first 95 moves as
book and then let Shredder play Gambit Shredder in auto232 mode on two
computers. I guess when they get locked into the tablebases they forget about
the 50 move rule. I wonder if all programs do this?
Jim
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.