Author: Bob Durrett
Date: 05:55:07 12/11/02
Go up one level in this thread
On December 10, 2002 at 20:08:21, Laurence Chen wrote:
>Hope this position will help to clarify Space advantage
>
>1. e4 c6 2. Nc3 d5 3. Nf3 g6 4. d4 Bg7 5. h3 a6
>6. Bf4 Nf6 7. e5 Ng8 8. Qd2 b5 9. Be2 h6 10. O-O-O e6
>
>[d]rnbqk1nr/5pb1/p1p1p1pp/1p1pP3/3P1B2/2N2N1P/PPPQBPP1/2KR3R w kq - 0 10
>
>Andy Soltis in his book, "The Art of Defence" comments as follows:
>"It doesn't take long to conclude that White has a very strong game. He has
>developed nearly all his pieces while Black's only developed piece, his king's
>bishop, bites on granite. Black's queenside is full of holes on dark squares
>and he has just locked in his queen's bishop. A quick mating attack is assured,
>you might conclude. And you'd be right!"
I read Soltis' book years ago. I distinctly remember this position. I laughed
when I first saw it, and I laughed again when reading this bulletin. You must
admit that it is a very amusing game.
You must realize that Soltis was "setting up" the reader when he said "It
doesn't take long to conclude that White has a very strong game." He was
joking! The gullible reader accepts his statement at face value. That's why
Soltis' little joke was so effective!
Of course, it is NOT true that White had a very strong game. Soltis was
demonstrating the pitfalls of *superficial* evaluation. Assessing the position
as "giving White a huge <snip> advantage" is a false conclusion of superficial
evaluation.
>What does your chess engine think about this position?
>
>11. g4 Nd7 12. Bg3 Bf8
>13. Rdf1 Nb6 14. Nd1 a5 15. Ne1 b4 16. Nd3 Nc4 17. Qe1 Qb6 18. b3 Qxd4 19. bxc4
>Qa1+ 20. Kd2 dxc4 21. Nf4 Qxa2 22. Ke3 Bb7 23. Qd2 g5 24. Nh5 c3 25. Qd3 Rd8
>26. Qe4 Bc5+ 27. Kf3 Rd4 28. Qe3 Qd5+ 0-1
>Andy Soltis: "Yes, Black delivered the mate. And in less than 20 moves from the
>diagram!"
>
>Below is the complete game.
>
>[Event "Latvian Championship"]
>[Site "."]
>[Date "1961.??.??"]
>[Round "?"]
>[White "Khliavin"]
>[Black "Zhdanov"]
>[Result "0-1"]
>[ECO "B15"]
>[Annotator ""]
>[PlyCount ""]
>[TimeControl ""]
>
>{256MB, Elect2002.ctg} 1. e4 c6 2. Nc3 d5 3. Nf3 g6 4. d4 Bg7 5. h3 a6
>6. Bf4 Nf6 7. e5 Ng8 8. Qd2 b5 9. Be2 h6 10. O-O-O e6 11. g4 Nd7 12. Bg3 Bf8
>13. Rdf1 Nb6 14. Nd1 a5 15. Ne1 b4 16. Nd3 Nc4 17. Qe1 Qb6 18. b3 Qxd4 19. bxc4
>Qa1+ 20. Kd2 dxc4 21. Nf4 Qxa2 22. Ke3 Bb7 23. Qd2 g5 24. Nh5 c3 25. Qd3 Rd8
>26. Qe4 Bc5+ 27. Kf3 Rd4 28. Qe3 Qd5+ 0-1
>
>Here Black chooses to under-develop his pieces, and move some of his pieces back
>to its original squares, and several times. This gave White a huge space
>advantage, but such advantage was not good at all. Why? Because the pieces
>which White had developed did not meet the demands of the pawn structure.
>Hence, the space advantage was useless, it had no value. White's developed
>pieces are unable to take advantage of his space advantage, therefore, White has
>to loose Time re-developing his pieces. What good is to have an advantage which
>cannot be exploited?
A very interesting question. It points to a "special problem" faced by chess
programmers, IMHO. It is not enough to evaluate a position by adding up the
superficially apparent "advantages" in a position. The GOOD position evaluation
code must also be able to roughly estimate the "exploitability" of the
advantages. Perhaps "exploitability" is something that can be fully assessed
only by a search algorithm. I don't know.
I have a lot of faith in the experienced chess programmers here, so my
expectation is that this special problem has already been "properly dealt with"
in the better engines.
>Just because one has space advantage does not translate
>into a winning advantage. Because Black pieces are back to its original
>squares, Black is able to regroup and put his pieces in squares which are more
>useful. Hope this position helps demystifying some ideas about Space as an
>advantage. Look at the Modern Defence as an example where Black chooses
>deliberately to have a space disavantage and allow White to build a huge space
>advantage. These games are see-saw battles where White tries to mantain his
>space advantage and Black tries to destroy it. See Gruenfeld Defence, too.
Your example above clearly demonstrates that a single type of "advantage" may
not exist in isolation within a single chess position. I am still searching for
such a position and have not found it.
Bob D.
>
>Regards,
>
>Laurence
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.