Author: Enrique Irazoqui
Date: 05:30:59 06/21/00
[D]1r6/1pb1k1p1/4p2p/1p1p4/3Pp2P/1R2P1PB/1P2P1K1/8 b - - 0 1
Yesterday I looked at this position that reveals once again how much trouble
programs have in recognizing the importance of blocked pieces. Some programs
pick and drop 1...b4, but none of them realize that the blocked rook is out of
the game until the search makes them see the consequences many hours later. The
evaluation at the initial position or after 1...b4 2.Rxb4 b5 3.Rb3 b4 is almost
the same. It takes 61 minutes for F6a and 335 minutes for Tiger to pick b4, and
much, much longer to fail high.
[Event "Biel"]
[Site "?"]
[Date "1977.??.??"]
[Round "?"]
[White "Andersson"]
[Black "Torre"]
[Result "0-1"]
[SetUp "1"]
[FEN "1r6/1pb1k1p1/4p2p/1p1p4/3Pp2P/1R2P1PB/1P2P1K1/8 b - - 0 1"]
[PlyCount "7"]
[EventDate "1997.??.??"]
1... b4 $3 2. Rxb4 b5 $1 3. h5 (3. Rb3 b4 {
/\...Bd6, Re8 (>< e6), Kd8-c7-...-c4 -+; >< g3}) 3... Bd6 4. Rb3 b4 $1 $19 {
>< Rb3, g3; /\ 5...Re8 (>< e6), 6... Kd8-c7-...-c4 -+ Andersson-Torre/Biel1977}
0-1
n-Torre/Biel1977}
0-1
The same theme appears in the Scherbakov-Rebel game:
[D]3r4/p3kp2/2p1ppbp/2b5/P3P3/2NB1P2/1PK3PP/7R b - - 0 21
Here Rebel played e5??, the bishop became a big pawn and black lost. Again, all
the programs I tried gave a similar evaluation to this initial position and
after 23.h3, failing to realize that in practice black is a piece down.
[Event "Monthly GM Challenge"]
[Site "?"]
[Date "2000.??.??"]
[Round "7"]
[White "Scherbakov, R."]
[Black "Comp Rebel Century"]
[Result "1-0"]
[WhiteElo "2540"]
[BlackElo "2500"]
[Annotator "Scherbakov,R"]
[SetUp "1"]
[FEN "3r4/p3kp2/2p1ppbp/2b5/P3P3/2NB1P2/1PK3PP/7R b - - 0 21"]
[PlyCount "70"]
[EventDate "2000.??.??"]
21... e5 $4 {Unbelievable!!! Obviously Rebel did not consider seriously
White's next move after which Black is practically a piece down.} (21... f5 $5
{It was better to make almost any other move (or even don't move on the whole!)
. Still was not so bad as after direct} 22. exf5 Bxf5 23. Bxf5 exf5 {
Black has active pieces and good play on the kingside.}) 22. g4 $1 {The Bishop
on g6 is now a "big pawn". The attempt to escape with f6-f5 is pointless.}
22... h5 23. h3
[D]8/4k3/4bpp1/7p/1p2p3/4P1PB/1P2P1KP/8 b - - 0 1
Time ago I posted this one, which is another case of blindness regarding blocked
pieces. Here, a few programs pick the winning f5, but the difference in
evaluation between f5 and next best is insignificant.
[Event "Black to move f5"]
[Site "?"]
[Date "1997.??.??"]
[Round "?"]
[White "Blocking bishop"]
[Black "?"]
[Result "0-1"]
[SetUp "1"]
[FEN "8/4k3/4bpp1/7p/1p2p3/4P1PB/1P2P1KP/8 b - - 0 1"]
[PlyCount "1"]
[EventDate "1997.??.??"]
1... f5 0-1
[D]r4rk1/p1q2ppp/1pn2n2/2p1p3/3PP1b1/P1P2N2/B4PPP/R1BQR1K1 w - - 0 14
Another position I posted bbefore. No program plays dxc5, invariably picking d5
instead and blocking their own bishop on a2.
[Event "Linares14th"]
[Site "Linares"]
[Date "1997.??.??"]
[Round "11"]
[White "Kasparov, Gary"]
[Black "Kramnik, Vladimir"]
[Result "1-0"]
[WhiteElo "2795"]
[BlackElo "2740"]
[Annotator "Psakhis"]
[SetUp "1"]
[FEN "r4rk1/p1q2ppp/1pn2n2/2p1p3/3PP1b1/P1P2N2/B4PPP/R1BQR1K1 w - - 0 14"]
[PlyCount "87"]
[EventDate "1997.02.16"]
[Source "ChessBase"]
[SourceDate "1997.04.10"]
14. dxc5 $1 {Natural 14.d5 will be serious positional mistake- bishop on a2
could be closed for a long time (Psakhis).}
There is a pattern in the first three positions that has to do with a piece
being blocked for as long as the opponent wants to, which effectively means
losing a piece and the game. No program sees it. Unless they find by search they
have no clue, and search often takes too long. Is it that difficult to teach
programs about all this?
Enrique
This page took 0.01 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.