Rook endgames are the most common ones in chess. Unfortunately, they are also complicated both in terms of evaluation and in terms of knowing how to play them. Therefore, it is not surprising that Tarrasch used to say that “all rook endgames are drawn.” Sometimes having an extra pawn (or even two) is not sufficient for a victory. While in queen endgames one can often escape with a perpetual check (thanks to the queen’s mobility and power), in rook endgames the coordination between all the pieces is more important. Rook endgames are really must-know, so don’t forget to study them and get some practice. One of the most respected authors on rook endgames is Mark Dvoretzky. His books are arguably the best on this topic.
So, how does one study rook endgames? First of all, review the theoretical positions, i.e. those endgames where the result is well-known. For example, positions with five pieces: two kings, two rooks, an extra pawn for one of the sides. The defending player must make sure he knows the Philidor position, and the attacker – the Lucena position. These are two most common techniques. All the theoretical positions should be memorized and played out against an opponent. You can practice against your coach, friend, or even set up the position on a PC (that IS challenging). Regarding theoretical positions: you have to KNOW them, i.e. keep in mind how to play move per move. Relying on general principles may let you down; you have to memorize precise variations.
Comprehensive knowledge of basic theoretical positions allows us to understand which transformations are beneficial for us, and which are not. For example, should you trade into a certain endgame or not? Will it be won/drawn/lost? If you keep those in mind, it will be easier for you to study theoretical positions that involve more pawns, but where we know only principles and plans, not the precise move order. For example, in most cases one can’t win a rook endgame with just one extra pawn on one side of the board (e.g. 2 vs 1, 3 vs 2, 4 vs 3). However, a lot depends on the particular position and the players’ endgame technique. Rook endgames are very demanding both in terms of calculation and understanding.
After having studied the theoretical positions one can take on standard ones. In those we don’t know for certain what the outcome is, but we still can use some defensive/offensive plans. For example, a rook endgame with 3 pawns vs 3 pawns on one side, and an extra pawn on the other side. Such positions are tricky. Nonetheless, if you know the typical ideas, playing them will be easier. Don’t forget not only to read about those, but also to practice at home!
Once you are done with all the theory, you can go through different examples from masters’ games. E.g. in Dvoretzky’s books there are many interesting and insightful cases that will help you grasp the essence of rook endgames.
Naturally, no matter how good your memory is, you will have to review the manuals from time to time, as many positions are very difficult to memorize for any player.
[ed note: there is a lot of good material on rook endgames on chess.com as well: this guide to videos on the subject, and here you can also find a bunch of rook endgames to practice against a computer opponent]
Rook endgames are really complicated. The good news is that if you study them well, you will often save worse endgames and win equal ones. Being confident in your endgame technique will greatly improve your overall mastery and make you a more universal player. Also, keep in mind that this will be noticeable not only in endgames, but in other stages of the games as well, as you will boost your calculation skills; adopt new opening lines; get a sense of what you are heading for in the middlegame.
The following game was played in the 2nd round of the European Club Cup vs WGM Olga Girya. The endgame that happened there was winning, but I failed to convert it.
After the opening a position with mutual chances occurred. I missed a few interesting continuations and ended up in a position in which I could press for a win only by taking risks. On move 29 I played an adventurous continuation and got a worse position, but my opponent didn’t take advantage of that misstep. Later she herself blundered twice in time trouble, and I got a winning rook endgame. Most rook endgames with extra connected passers are 1-0. In this case I had just one extra pawn, but the connected passers and Black’s lack of counter-play were enough for a victory. Nonetheless, I managed to spoil the game by blundering in return. Rook endgames require lots of precision and skill.