Chess Improvement: Pattern Recognition / Endgames

The endgame pattern is tougher to see than that in the the previous post, but the pattern is essential to understanding the position and many others like it. White is up a pawn and Black is faced with a difficult defense to hold the game. Through a little pattern recognition, he sacs the Knight for a much easier route through to the draw.



Note that White can save the h-pawn, but it doesn't matter: 1.dxe6 fxg3 2.Bxe4 Kxe6 3.Kd4 g2 4.Bxg2 Kf5 5.Bf3 Ke6 6.Bg4+ Kf7 7.Ke4 and there's no way to prevent the Black King from controlling h8, the promotion square. This Rook pawn and wrong-colored Bishop endgame arises often.

Chess Improvement: Pattern Recognition / Endgames

This is a good example of a simple position that should be recognized instantly. It is easily calculable, but it should be a position that requires no calculation on your part. That's not to say that you won't confirm what you know if there's time available, but presume that you have virtually no time, and you can play a short series of forcing moves to arrive at this position. Do you play them? Does it matter who's to move in this position? Include such positions in your tactics drill set.



Suggested Reading in order of recommendation:

Silman's Complete Endgame Course--Jeremy Silman
100 Endgames You Must Know--Jesus de la Villa
Dvoretsky's Endgame Manual--Mark Dvoretsky (available in book and software)
The ABC of Endgames--ChessBase software

Semi-Slav 101: Bayonet; 7...h6

This Semi-Slav variation is marked by the moves 6.Qc2 Bd6 7.g4. Black has four traditional responses to 7.g4: 7...h6, Nxg4, Bb4, and dxc4. This variation is credited to Alex Shabalov, with Alexei Shirov playing it for the first time in 1992. The variation is thus often called the Latvian Gambit (somewhat confusing since that name is given to 1.e4 e5 2.Nf3 f5) or the Shabalov-Shirov Gambit. Since the opening's beginning, the most popular response has been 7...dxc4, but at top levels in recent years, the move has been 7...h6. Filtering a largely up-to-date Chessbase Mega 2009 database, I find 50 games with both players greater than 2500 in games played from 2006-2009. The results are (with a Rybka evaluation at 18 ply tacked on):

7...h6, 31x, W: 48.4%, Perf. Elo +29, (0.00) Rybka 3
7...Nxg4, 10x, W:70%, Perf. Elo -147, (0.00) Rybka 3
7...dxc4, 4x, W:37.5%, Perf. Elo +50, (0.09) Rybka 3
7...Bb4, 4x, W:37.5%, Perf. Elo +140, (0.32) Rybka 3

Draw your own conclusions on this sparse sample, but it at least indicates the top players' preference. I have not been able to confirm this preference with the aid of Rybka; however, and I've stood by 7...dxc4 with good results. We'll look at each of Black's responses in turn, beginning here with 7...h6. The conclusion on 7...h6 can be summed up as "solid."

The endgame is highly instructive. Take a look at the position after White's 56th and compare it to the 1949 game that follows.