Author: Kurt Utzinger
Date: 00:10:02 04/08/04
Go up one level in this thread
On April 08, 2004 at 01:02:42, Mridul Muralidharan wrote:
>Hi,
>
> Importance of endgame knowledge has been time and again discussed here.
>But it is quiet frustrating for a chess patzer like me to implement this in the
>engine - much more if I attempt to do this effectively.
>Simple things that other engines statically evaluate as very good/very bad
>require some pretty heavy searching by new engine due to the miserable lack of
>chess ending theory knowledge for both the program and its author.
>In the following game against crafty , my engine gets a very good position and
>material advantage and then throws it - literally - away and , forget draw :
>looses the game in the end !! :(
>
>Lot of people have suggested some good resources on endgame study to me , and if
>anyone has more suggestions on what resources to look at my thanks in advance.
>(THanks to all who did - I have started looking seriously into them now that my
>search/extensions/move ordering/basic eval/interface code/etc/etc is more
>stable).
>
>Also , if someone could analyse and give me good alternatives to the miserable
>goofup moves made towards the end , might help me more in "fixing" the engine.
>Both human and computer analysis are welcome - already starting a analysis by
>some engines here now to get complete list of blunders ....
>
>Thanks in advance,
>Regards all,
>Mridul
>
I have made some comments of this game.
Kurt
[Event "Computer chess game"]
[Site "TITAN"]
[Date "2004.04.08"]
[Round "?"]
[White "Borer ver 2.01"]
[Black "Crafty 19.3"]
[Result "0-1"]
[ECO "A46"]
[Annotator "Utzinger,K"]
[PlyCount "136"]
[EventDate "2004.??.??"]
1. d4 Nf6 2. Nf3 e6 3. Nc3 Be7 {3...d5 to control e4 looks more logical} 4. Qd3
{It's not clear where the queen will find its best place and an important rule
says one should not move the queen too early in the opening phase. Stronger
therefore 4.e4} 4... Nc6 5. a3 {
A waste of time instead of taking control of the centre with 5.e4} 5... O-O 6.
h4 {To start an attack before pieces have been developped is usually not a
good strategy} 6... h6 {Not at all necessary. In the given situation one would
expect 6.. .d6 preparing e6-e5} 7. e4 d5 {
And here 7...d6 was again a more sound plan} 8. e5 Ng4 9. Be2 f5 10. Nd1 Na5
11. Bf4 c5 {An interesting move of Crafty 19.3 but at the end it will be at
Black's disadvantage} 12. dxc5 Nc6 (12... Bxc5 13. b4 {and wins}) 13. b4 Qc7
14. Nc3 a6 (14... Ngxe5 15. Nxe5 Nxe5 16. Qg3 Bf6 17. Nb5 Qb8 18. Rd1 Bd7 19.
Nd4 {and White has clearly the better position}) 15. Na4 a5 16. Nb6 Ra7 17. Nd4
Nxd4 18. Qxd4 axb4 19. axb4 Rxa1+ 20. Qxa1 Bf6 21. Bxg4 fxg4 22. Qd4 Bd8 23.
Be3 h5 {A good and better try was 23...g3} 24. Bg5 Bd7 {White is a pawn up and
has the better position but it's not so easy to make progress as long as the
king is left in the centre and the rook is out of play.} 25. Bxd8 {White should
try to get rid off its double pawns in the c-file with 25.c4 in order to later
have the possibility to obtain a free pawn on the queenside. This would
eventually decide the game} 25... Rxd8 26. O-O Bb5 27. Ra1 Rf8 28. Ra5 Be2 29.
Qd2 Bc4 30. g3 Kh7 31. c3 Qf7 32. Ra7 Bb5 33. Qf4 Qxf4 {A wrong decision:
Black should keep queens on board to put pressure against White's camp on the
half-open f-file} 34. gxf4 Bc6 35. b5 {And this time a wrong decision of White}
(35. Ra5 Rxf4 36. b5 Be8 37. Nc8 {and White has a won position in this endgame}
) 35... Bxb5 36. Rxb7 Rxf4 37. Nxd5 exd5 38. Rxb5 Rc4 {And already from here
on it is no longer clear if White can win the game as Black has a very
actively placed rook} 39. Rb8 (39. e6 Kg6 40. c6 Rxc6 41. e7 Kf7 42. Rxd5 Rxc3
43. Rxh5 Kxe7 44. Rg5 Kf6 45. Rxg4 {and this rook ending should be a draw})
39... Kg6 40. Rf8 Rxc5 41. e6 Rc7 {It's still a draw} 42. f4 {But this is
wrong as Black can later on take possession of the fourth rank with its rook
or bring the king to g4. In both cases White's h4-pawn will go lost. On the
other hand it's true that even after this inaccuracy, the game is still drawn}
42... gxf3 43. Rxf3 Re7 44. Rd3 Rxe6 45. Rxd5 Re4 46. Rd6+ Kf5 47. Rd5+ Kg4 (
47... Re5 48. Rxe5+ Kxe5 49. Kf2 Kf5 50. c4 g5 51. hxg5 Kxg5 {=}) 48. Kg2 {
?? A move that is hard to understand: why not 48.Rg5 with an easy draw} (48.
Rg5+ Kxh4 49. Rxg7 Rc4 50. Kg2 Rxc3 51. Kh2 Rc2+ 52. Rg2 Rxg2+ 53. Kxg2 Kg4 {
draw}) 48... Kxh4 49. Rd4 {???? The worst move possible in this situation
allowing Black to exchange rooks and to get a won pawn ending. To avoid this
mistake no endgame knowledge is required as it can seen by mere calulation
that Black will win the d-pawn and then remain with two more pawns} 49... Rxd4
50. cxd4 Kg4 51. d5 Kf5 52. d6 Ke6 53. Kg3 g5 54. d7 Kxd7 55. Kh2 Ke6 56. Kg3
Kf5 57. Kf3 g4+ 58. Ke3 h4 59. Kf2 Kf4 60. Ke2 h3 61. Kd3 g3 62. Kd2 g2 63. Kd3
h2 64. Kc3 h1=Q 65. Kb3 g1=Q 66. Kb2 Qc1+ 67. Kb3 Qb7+ 68. Ka2 Qcb2# {
Black mates} 0-1
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.