Monday, April 2, 2012

A Rook Ending from the Denker

Here's a position from Moskwa v. Kogen that illustrates the differences between rook endings and queen endings discussed in the last post.

Robert's pawn is nearer to queening but that's not nearly as big an advantage in a rook ending because the Black rook will sacrifice itself for the new queen after which the White rook won't be able to handle the Black king and connected pawns alone and may be forced to sacrifice itself in return.  A queen could handle the Black king and pawns with ease.

The game went 39.Rc8 Rd3 40.Ke6 b3 41.d7 a4 42.d8=Q+ Rxd8 43.Rxd8 Kc5 and in order to draw White will have to hustle his king over to help with the defense and sacrifice his rook at the right moment.  On the other hand, 39.e5! would have won after 39....Rd3 40.e6 Rd5+ 41.Ke4 Rxd6 42.e7 42.Rxe7 42.Rxe7  because the White king is close enough to assist the rook before the Black pawns advance too far.

A Queen Ending from the Denker

One of the places that I thought Robert Moskwa might run into trouble at the Denker Qualifier due to the relatively short time he has been playing serious chess was in the endgame.  Although his calculation skills are excellent, you often don't have enough time left to calculate as thoroughly as needed when you reach the endgame.  As a result, the kind of general understanding of endgame principles that you can only get by experience, especially painful experience, can be very important.

Nowhere is the lack of time felt more than in queen endings.  On nearly every move, the players must try to visualize the consequences of an avalanche of checks.

In most endings, a player must carefully weigh the value advancing a pawn closer to the queening square against the possibility that it becomes weak due to lack of support.  Hence, the general endgame principle of pieces before pawns, i.e., unless the ending is a pure pawn race, players should try to improve the position of their pieces before they advance their pawns.

In queen endings, however, advanced pawns are much less likely to become a liability.  This is because a queen has the ability to protect a pawn and control the squares in front of the pawn at the same time without hindering the pawn's advance.  A king can blockade a pawn that is only supported by a knight, a rook, or a bishop, but it is helpless to stop the advance of a pawn supported by a queen.  In fact, it often has to worry about being mated. Even another queen acting alone cannot hold back a pawn supported by a queen.

The flip side of the queen's ability to advance a pawn without assistance is its ability to oppose a king and a pawn without assistance.  Acting alone, a knight, a rook, or a bishop has difficulty holding back a pawn supported by a king and can often be forced to sacrifice itself to prevent a pawn from queening.  A queen on the other hand has very little trouble holding back a pawn by itself.  In fact, it can often hold back several pawns.

The upshot of all this is that having the farthest advanced pawn is of paramount importance in queen endings.  It is not unusual to see a deficit of several pawns can be offset by a single pawn that has reached the sixth rank because the extra pawns can be easily rounded up if the defender can be forced to sacrifice his queen.

All of this brings us to the following position from Moskwa v. Meduri.

Fearing Black's g-pawn, Robert played 59.d7+? Qxd7 60.Qxg3 and the game ended in a draw.  However, despite the fact that Black's g-pawn is as near to its queening square as the White's d-pawn, the Black pawn is not supported by its queen and Black queen is needed to defend against mating threats.  After 59.c5!, Black can't play 59...g2 because of 60.Qb8+ Kf7 61.Qb7+ Kf8 62. Qxg2.  If Black tries 59...Qd7, 60.c6! is devastating.