Today we will continue with the article series on endgames from the US Championship. The topic of today's article is the endgame from Shulman - Robson. Usually, I prefer to look at more than one endgame at a time but this endgame proved so full with ideas and critical moments that it will easily fill today's article. We will break the endgame into critical positions and try to find plans for each critical position. The ideas that will be covered include pawn breaks, open files for rooks, exchanges, passed pawns, and rook placement when there are passed pawns. We will start analyzing from the following position.
Robson is down an exchange for a pawn. However, the position is more of a closed type with few half-open files and only one open h-file. Therefore, white cannot fully exploit the rook's potential, unless he opens more files. The knight on c4 not only protects the weakness on b6 but also blocks the c-pawn and is ready to relocate to a better square on e4. Both bishops seem not to play a major role now but this might change. It is black to move and if you evaluate the position as worse for black-- which is the correct evaluation-- then it is time to find a defensive plan.
The first question that comes to mind is what is white's plan? I tried hard to come up with one but wasn't successful. It is very hard for white to improve the position. Exchanging the bishops might be one of the plans but what comes after that ? White can get a rook to the 8th-rank through the h-file but then all the black weaknesses are defended with the rook on the 6th and the knight on c4. Overall, white cannot be too relaxed because after Rc6 and a knight move, the c3-pawn will be in danger. So, I propose the plan for black of waiting, to do nothing and to see what white can show. The following variation illustrates possible ideas when black chooses a waiting strategy.
The above variations do not exhaust the position but serve as illustrations of possible plans that white could choose. Instead of following the passive strategy, Robson decided to create a passed pawn at the cost of a pawn. The resulting a-pawn is dangerous but white also has a newly-created b-pawn to push forward.
Up until now white's play deserves only praise. Shulman managed to open the e-file to get the rook to the 6th-rank and to threaten to push the b-pawn. Now, he is faced with a tough decision, whether to give up the b-pawn but get a piece back for it or to keep the b-pawn but enter some tactical complications. Let us look at the second option first.
Shulman chose the second path, where he takes the g-pawn but has to find quite a few precise moves to win the endgame. For example, on move 42 it is very hard to find intricate maneuver Rh4-Rh5 with the sole purpose of putting the black rook on the "bad" square c5. Right away Rh5 will not achieve the goal as the black rook gets to be active on b2. White's task is hard here, it is easier to play this position for black as his plan is simple - queen the a-pawn.
Another critical position. White managed to win a piece but black still has the dangerous a-pawn. If white managed to win the f-pawn the game would be over. On the other hand trading the g for the f-pawn will result in a drawn endgame. I remember watching the postmortem and somebody suggesting Ra6 right away. After shuffling the pieces for a while, the grandmasters came to the conclusion that the endgame is winning for white. Here, I show one possible variation after Ra6. Shulman played Kf3 - which is a good move in itself but it allows black to get the rook to the 1st-rank, from where it can always defend the f-pawn from behind, while securing the 1st-rank for the a-pawn promotion. Robson's play is impeccable in the rest of the game.
Today we analyzed yet another game from the US Championship. Shulman did not win being a piece up, but this only proves the fact that the endgame stage is not a simple part of the game. The game is a good example of the price of passed pawns. Robson sacrificed a pawn to get a passed pawn. Shulman sacrificed his passed b-pawn but won a piece, which was enough for a win but he didn't see the critical Ra6. In the end the a3-pawn secured black a draw.
Next week we will look at another endgame from the Championship.