Today we will look at two endgames from the recent and very strong Washington International chess tournament that took place in Maryland. The two endgames that I chose are similar. Both feature a position where one side has an extra pawn but the other side has enough piece activity to compensate for the lost material. In the first game Kamsky managed to exchange a pair of pieces and activate his king safely. Onischuk in the second game could not achieve the same and the game ended up in a draw. Let 's proceed to the examples.
Looking at this position I thought white has to be winning - an extra pawn in endgame is a big material advantage. Being up material is only one notch on the evaluation scale. There are permanent and temporary advantages, material advantage being permanent. White's king is weak, the space on the kingside is nice to have but it is a tremendous disadvantage when the king's position is considered. The f-pawn requires constant attention, and the e-pawn is no better because of the g5-break. White's passed pawn is still on b3, he has to come up with a way to push it forward, while keeping the kingside together. The computer's evaluation is close to equal in this position. Kamsky is probably one of the best or the best chessplayers in the world who can convert super-small endgame advantages into a full point. His opponent - Gareev is a very strong grandmaster too, which makes this encounter even more interesting.
Kamsky's strategy was and is to exchange a pair or all major pieces. Gareev is in such a situation where there is no way one can avoid trades. On Qd1 there is Qc2 and on Qh1 as happened in the game there is Rd2. The king from f2 takes important e3-squre from the rook. The question is what to exchange and what to keep on the board? Intuitively, it seems black should aim for the rook trade, keeping the queens on the board due to poor placement of white king. However, as the game shows this is not the case. Keeping the rooks on the board and undermining the e5-pawn right away gave black more chances.
The next example is similar to the previous one because one side is also up a pawn but this passed-pawn is at the starting rank. White pieces are very active, plus rooks with a bishop coordinate better than rooks with a knight. The black king can get in trouble - being defended only by a knight, so Onischuk has to be careful converting the extra pawn. The game ended as a draw in three moves; it is understandable that black did not want to risk by placing the knight on e8 but in my opinion it was the best option.
The main line we considered started with the move Kg6. GM Kaidanov came up with 28...Nd5, taking away the e7-square from the rooks. After the game I talked to Gelashvili and asked his opinion on the resulting endgame. He thought that at least he had the endgame--rook with three pawns vs. the knight + rook and the h-pawn. Gelashvili thought that he will hold that endgame without much trouble. Saying "at least" does not mean that he would go for this continuation in the game, I am sure both would have found the continuation after 29.h4.
In today's two examples we saw that having active pieces sometimes compensate lost material. In the first example Gareev could have sacrificed the second pawn but kept the queen active which would lead to an about equal endgame. Instead he locked his queen on h2 which ended the game immediately. In the second example it was very hard for black to untangle but there was a possibility of placing the knight temporarily on a passive square but get the pawn moving.
Next week we will look at few more endgames from this tournament.