I have been writing this column about endgames for more than two years - a genuine surprise for me because it feels like I just started. It is time to move on to a different topic but before going forward we need to look back and have a wrap up of what has been done and how one has to keep studying endgames independently.
Looking over the material presented over these two years I can summarize it all by dividing it into these few categories:
- theoretical endgames (ex: R+B vs. R or R+3P vs. R+4P)
- planning and typical endgame structures (ex: endgames from French, KID, QID etc.)
- unbalanced endgames (ex: Q vs. 3 pieces, R vs. 4P)
- endgame virtuosos (ex: games by Karpov, Andersson, Sokolov)
- tactics and initiative in endgames.
- modern endgame treatment (ex: new ideas from recent games)
I am sure that this list can be extended but it captures general ideas that we have been studying over this period of time. You can divide your time of endgame study between these categories but how to do it efficiently I would try to address it here. First of all, one should start by analyzing one's own endgames. Let me give you an example of my recent loss to GM Shankland without analysis for now.
Let us place this endgame into a few of the above mentioned categories. Due to the limited amount of material this endgame can be considered theoretical, so Category 1. It is not 3 - the material balance is common, probably very few tactics are present and rook endgames are very common and have been studied for many-many years so one should not dig much into 6. modern treatment. It can be number 2 as it is very important to know some kind of plans here.
Now that we know where it belongs we can look up some study material. Dvoretsky's Endgame Manual is an excellent reference for theoretical endgames. Let us start by looking up the chapter on R vs. R with extra pawn. Aha, there is a chapter called "A Rook and Two Pawns vs. a Rook and Pawn" with a sub-chapter "Disconnected Pawns, One of them is Passed". The reason I like this book is that Dvoretsky right away gives you gist of the position in one sentence. For example, in this chapter he says "if one or two files separate pawns of the stronger side, the position is most often a draw." In my endgame two files separate the pawns, so it should be a draw.
However, there are no examples with one or two files because in practice it is more often seen the pawns being at larger distances, therefore Dvoretsky concentrates on those examples. Although, we haven't gotten any specific plans of defense we know that the endgame should be a draw and the defence mechanism is rather trivial.
At this point one should either ask a more experienced player or look for other sources. Going to an engine wouldn't give one general rules or understanding that could be applied in the future. The other endgame reference that I happened to have is John Nunn's "Chess Endings 1 and 2". In a similar chapter Nunn states that the most common situation is indeed when the pawns are close together and he chooses to concentrate on these positions (research pays off for us!!).
The defensive idea consists of keeping the king in a way so it defends its own pawn and tries to stop the opponent's pawn. Here, one should invest time into studying the examples that GM Nunn gives and then use the knowledge for the analysis of the position from their own game. Here, I present partial analysis of GM Nunn's book of the position from Shapovalov - Estrin, which he evaluates as equal if white plays it correctly. The white king will be able to stop the e-pawn but it is cut off along the 2nd-rank for now.
Now, that we have some experience with this type of position it is time to apply this knowledge to the analysis of the game.
If I knew that the starting position is a straight-forward draw then I would not resort to unorthodox methods. Keeping the pawn on e3 and the king on f3 with the rook on the a-file is sufficient. There is no need to force the draw right away, one should have confidence in his/her endgame skills. It is important to analyze one's own endgames and to categorize them. Today's endgame was clearly theoretical, but with the next and last article in endgame series we will look at other practical endgames that belong to different categories and the methods that we can use to improve there.