Author: Laurence Chen
Date: 13:01:54 12/16/98
I want to give thanks to Thom Perry for his initiative and help in analyzing
Capablanca's sacrifice. Much light has been shed in this position, and we now
can gain a better understanding of the position. For those of you who have not
read the previous posts I present the initial position again.
r1bq1rk1/pp2nppp/4p3/2n5/8/2NBPN2/PP3PPP/R2Q1RK1 w - - 0 12
Capablanca played 12. Bxh7+ and won the game, the question has been that many
chess engine failed to find this sacrifice and that some chessplayers concluded
that the move must be unsound because the chess engine gives a negative PV.
Well, is the engine correct in the assessment of this position then? Karate or
Kendo?
That is the title for this heading which my friend suggested me to use. His
insight was because karate is a hands free art of martial arts and kendo is a
sword art of martial arts, he asked me the question who wins the fight the sword
or the man behind? A fair question, are we totally dependent on chess engines
analysis, or do we use the engine to validate our theories and ideas? Karate or
Kendo, that is the choice !!!
Remember that chess engines suffer the deficiency of not being able to foresee
long ranges plans, it suffers what is known as horizon blindness. It was
interesting to see when the sacrifice was played manually the initial assessment
of the chess engines has been a negative PV which indicates that the engines
favour Black's position. As the forced moves are played out, and the situation
becomes more clear to the chess engines, the PV evaluation starts to change
dramatically, Black no longer has the positional advantage, the PV becomes 0 or
slightly positive. Therefore, White has enough compensation for the material
sacrificed. In testing this position it was found that the games ended up in a
draw with correct defense of Black's part, and that if White or Black try to
force a win in this dynamic balanced position they lose the game. The position
started as a quiet even position and transformed into a dynamic balanced
position with the sacrifice of a bishop. Does that mean that the sacrifice has
been refuted? Some of you may conclude that, perhaps you are right if you
believe that sacrifices must always yield a win. I don't think so, I've seen
games where GM make sacrifices in difficult or "lost" positions in order to save
the game, and just because the sacrifice yielded a draw does that make such a
sacrifice unsound? The whole point was Capablanca played a REAL SACRIFICE, and
he got a win, but in my analysis and Thom Perry's showed that Black can draw the
game by accepting the sacrifice with correct defense. If Black refuses the
sacrifice, White's stands better and is winning. Hence not all real sacrifices
are unsound, the same is true to pseudo-sacrifices (combinational). Not all
combinations are sound, I've seen examples where a chessplayer played a combo
and it turned out that he missed an intermediate move and lost. So if you claim
that real sacrifices are unsound, then combinational sacrifices falls in the
same category, because there are combo's which are unsound. Of course, chess
engines won't make such mistakes, and it is rare to find a chess engine playing
a unsound pseudo-sacrifice with today's engine. In the past, I've played with
some chess engines which missed some intermediate move in their combo and they
lost the game. I expect with the great advancement in microprocessor technology,
the CPU is getting faster and faster each year, the chess engine software
designers will improve their engines and be able to incorporate real sacrifices
in their engine, and perhaps solve the horizon blindness problem. Below are the
games which show the results of my findings and Thom Perry's. You will see that
CM 6000 did not find the proper defense and was easily punished by Junior, and
in another game Junior pressed to win and played a weak move 20. fxe5??,
although it showed a positive PV, I looked at position and concluded that the
position favours Black and White is hard pressed for a draw, and using CM to
play out the position, CM punished Junior. Of course, I must give credit to Thom
Perry for the analysis of the position using MCP8, MCP8 finds the correct move
which forces a draw. Notice that in such position White cannot force a win, and
Black cannot cash in his material advantage without being mated. So it was a
short and exciting draw. MCP also evaluates the position in favour of Black,
however as the forced moves are played out, its evaluation changes gradually to
0.00. To conclude I want to leave with these words of GM Pachman, "if both
players have equal chances we speak of the equilibrium of the position... this
concept has nothing to do with that of a drawn position... when we mantain that
the equilibrium can only be disturbed by a mistake, we do not mean that a player
cannot fight to achieve this result. To bring about an advantage, he MUST CREATE
COMPLEX STRATEGIC AND TACTICAL PROBLEMS for his opponent. Even in so called
drawn positions it is sometimes possible to find a continuation which makes it
hard for the opponent to formulate the correct strategic plan, or which even
induces a tactical error." (Note: the bold lettering are mine). So it can be
concluded that Capablanca sacrifice was sound because it create problems for his
opponent and his opponent was not able to solve, and with the help of chess
engines (Junior 5, MCP8), it was shown that Black could save the game only with
accurate moves. I enjoyed this litte discussion and I hope you chessplayer can
also benefit from it. Thanks Thom for helping.
[Event "?"]
[Site "?"]
[Date "1998.??.??"]
[Round "?"]
[White "Junior 5.0"]
[Black "CM 6000"]
[Result "1-0"]
[SetUp "1"]
[FEN "r1bq1rk1/pp2nppp/4p3/2n5/8/2NBPN2/PP3PPP/R2Q1RK1 w - - 0 12"]
[PlyCount "125"]
12. Bxh7+ Kxh7 13. Ng5+ Kg6 14. Qg4 f5 15. Qg3 f4 16. exf4 Nf5 17. Qg4 Nh6 18.
Qh4 Rh8 19. Rad1 Qf6 20. Rfe1 b6 21. Rd5 Nf7 22. Qg3 Nxg5 23. Rxg5+ Kf7 24. b4
Na6 25. Qf3 Rb8 26. Ne4 Qe7 27. Qc3 Kf8 28. Nd6 Bd7 29. Rf5+ Kg8 30. Rf7 Qxf7
31. Nxf7 Kxf7 32. Qd3 Bc8 33. f5 Re8 34. fxe6+ Rxe6 35. Qf5+ Ke7 36. Rxe6+ Bxe6
37. Qg5+ Kd6 38. Qg3+ Ke7 39. Qxg7+ Bf7 40. a3 Nc7 41. h4 Ne6 42. Qe5 Rf8 43.
f4 Bg6 44. g3 Bd3 45. h5 Bc2 46. Qe2 Rc8 47. g4 Kf7 48. Qb5 Rd8 49. f5 Nd4 50.
Qc4+ Ke8 51. h6 Bb3 52. Qc7 Nf3+ 53. Kf2 Ng5 54. Qe5+ Kd7 55. Qe3 Nf7 56. Qxb3
Ke7 57. Qe6+ Kf8 58. h7 Rd2+ 59. Ke1 Rd8 60. Qf6 Ke8 61. g5 Rd6 62. Qg7 Rd7 63.
g6 Re7+ 64. Kf2 Nd6 65. h8=Q+ Kd7 66. Qd4 Re4 67. Qh7+ Re7 68. g7 Kc6 69. g8=Q
Ne4+ 70. Kg1 Nc5 71. Qhg6+ Kb5 72. a4+ Kxa4 73. Qa2+ Kb5 74. Qdc4# 1-0
[Event "?"]
[Site "?"]
[Date "1998.??.??"]
[Round "?"]
[White "Junior 5.0"]
[Black "CM 6000"]
[Result "0-1"]
[SetUp "1"]
[FEN "r1bq1rk1/pp2nppp/4p3/2n5/8/2NBPN2/PP3PPP/R2Q1RK1 w - - 0 12"]
[PlyCount "80"]
[EventDate "1998.??.??"]
12. Bxh7+ Kxh7 13. Ng5+ Kg6 14. Qg4 f5 15. Qg3 f4 16. exf4 Nf5 17. Qg4 Nh6 18.
Qh4 (18. Qg3 Nf5 19. Qg4 Nh6 20. Qg3 20... Nf5 {0.5-0.5 Draw. MCP8 Analysis.})
18... e5 19. Rad1 Nd3 20. fxe5 (20. f5+ Nxf5 21. Qh7+ Kxg5 22. Ne4+ Kf4 23. Qh5
Nd4 24. Nc3 Qg5 25. Nd5+ Kf5 26. Ne3+ Kf6 27. Nd5+ Kf5 28. Ne3+ Kf6 29. Nd5+
29... Kf5 {0.5-0.5 Draw. MCP8 Analysis.}) 20... Qxg5 21. Qxg5+ Kxg5 22. Rxd3
Kg6 23. Nb5 Bf5 24. Rb3 Rfc8 25. Re1 Be6 26. Ra3 Nf5 27. b3 a5 28. Nd6 Nxd6 29.
exd6 Bd7 30. h4 Rc6 31. Rd1 Rca6 32. Rd2 b5 33. b4 axb4 34. h5+ Kxh5 35. Rxa6
Rxa6 36. Kf1 Kg6 37. g3 Kf6 38. Ke2 Ke6 39. Ke3 Rxd6 40. Rc2 Ra6 41. Rb2 Ra3+
42. Ke2 Kd6 43. Rd2+ Ke7 44. Rb2 Ra4 45. Kd3 Be6 46. Ke3 Kd6 47. f3 Bxa2 48.
Kf4 Kc5 49. Ke5 b3 50. Re2 Rc4 51. Rh2 Rc2 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.