Thursday, July 19, 2007

Endgame strategy. Thinking Out Loud

This is a post that will grow in the next few days. So please come back regularly and comment, because feedback helps me to think.

Update 1 black
Update 2
Update 3

To keep matters simple I describe the strategic idea's from the position of an attacker (white) who plays against a defender (black). The strategy for the defender is the negative from the attacker. To improve readability I drastically cut down on information that is considered common knowledge. For instance that having 2 pawn islands is better than having 3 pawn islands etc.. This is no scientific paper or so thus there is no need to be complete.

At the transition from the middlegame to the endgame the method to try to win changes for the attacker. In the middlegame you try to win a piece with a tactical shot, or to attack your opponents king. That crude methods don't work in the endgame. In the endgame there is only one method by which you can win:

Queen a pawn.

Ok, there can be an occasional matingnet or tactical shot, but that's not the meat of the matter. The fact that your king can walk out in the open without being harassed is a clear sign that the threat of enemy pieces has diminished drastically.

With the ultimate goal of the endgame in mind it is possible to define the task of the remaining pieces and of the king:

The task of the pieces is to assist the pawns with queening.

(remember I promised you triviality?)
Let's see what subtasks can be derived from this.

Look at the following diagram

In a balanced sitiuation the first subtask becomes evident:

Create a passer.

First the enemy pawns that impede the journey to the promotion square must be removed by all possible means: trade, decoy, conquest. This kind of reasoning tells you where you want your pieces and your king: behind the enemy pawns to stab them in the back.

You have to penetrate with your king and/or piece into the enemy camp in order to attack the pawns.

At the same time you must prevent of course that the opponent penetrates into your camp. Here you find the underlying reason for the maxim centralize the king in the endgame. In the center you are closer to the enemy camp. At the same time you are ready to defend.
When we talk about piece activity or king activity we always talk in relation to what do they do to the enemy pawns. If there are no enemy pawns in the center then your king is not active when he is in the center.

In order to help penetration special care has to be taken when you have a bishop. Take for instance the following diagram:

You always have to place your pawns at the opposite color of your bishop. No matter what piece your opponent has, knight or bishop from whatever color. Look what happens:
  • Your bishop and pawns work together. The bishop covers the light squares, the pawns cover the dark squares. Together they make it impossible for the opponents king to penetrate.
  • They tend to fix the black pawns on the light squares, what makes them potential targets for your bishop.
  • They tend to fix the black pawns on the light squares, where they impede their own bishop
  • They enable your bishop to move freely over the board, which helps when you want to penetrate your opponents territory to stab his pawns from behind.
You can see what happens to black when he places his pawns at the same color as his bishop: the white king can penetrate via the black squares.
When you replace the lightsquared bishop of black with a darksquared bishop, you get bishops of opposite colors. It is easy to see why such position tends to be drawish: any change in the pawnstructure that is beneficial for your bishop is automatically beneficial for the opposite colored bishop. Besides that they live in a different universe, so they cannot influence each other.

One of the problems with the pawns that are standing abreast is that they are flexible.

You can't shoot a moving target. So you have to fixate the pawnstructures.

King and pieces are much too volatile in the endgame. They are too fast to hunt them down. So if we talk about targets, or about weaknesses, we talk about pawns. When pawns are fixated they form chains. The backward pawn from a chain is a natural target.

An extra weapon that is seldom or never mentioned is the sole distance to the promotion square:

With advanced pawns a whole bunch of new threats come into play for the sole reason that the promotion square is nearby. In the diagram above a simple breakthru is enough to secure the win. Are the pawns advanced and fixed, than often a piece sac is enough. This leads to the following maxim:

Move your frontier ahead.

Whenever you can do it safely. Closely related to this is a method to get a local advantage with equal material or even with less material. Take for instance the following position where white is a pawn down:

At the right side only one pawn keeps three pawns under control. This is caused by the fact that the g-pawn is closer to promotion than the 3 pawns it keeps under control. On the left side white has a local majority which secures the win. Of course white must watch out not to end up in the same situation as black. 1.b4? would be a major blunder which gives blacks b-pawn the chance to block both white queenside pawns with 1. ... b5. In stead white should play 1. a4 in order to create an outside passer. This very important priciple is called by Capablanca:

One unit holds two.

An important technique to create a passer is when you have a local pawnmajority. That is just what happened in the diagram above. White has a local majority of 2 vs 1 on the queenside. The pawns are traded and you are left with a passer, Here you see the downside of a double pawn: for attacking purposes you have to reackon with 1 pawn less when you have a double pawn. For defending purposes there is no difference.

Use your local pawnmajority to create a passer.

If you have managed to inflict your opponent with a weakness and you attack that weakness, usually that bounds the enemy piece to the defense of that weakness. That is very good, since the attacker is always more flexible, since he is the one who can decide when to move his piece. Generally it is not enough to inflict only one weakness and is a second weakness needed to "stretch" the defense beyonds its limits. This is called

The principle of the two weaknesses.

Let's have a closer look at a weakness. The chessworld isn't very consequent in the use of the term weakness. The next definition is an attempt to avoid confusion:

A weakness is something that requires defending recources.

There are 3 kind of weaknesses:
  • Target. Since the king and the pieces are too volatile, only certain pawns can be a target. A target is a pawn that: 1. isn't protected by another pawn and 2. cannot be protected by another pawn in the near future and 3. cannot move into the protection of another pawn and 4. is accessible by an attacking piece or king. Typical targets are: isolated pawns. The backward pawn of a pawnchain.
  • Invasion square. Any square in the enemy camp that gives an attacking piece or king access to a target.
  • Passer. When the attacker has a passer, the road to the promotion square has to be defended by the defender.
It is kinda weird. After 3 month's of middlegame study I formulated that piece activity is the most import in middlegame strategy. After a few months of putting this idea into practice I want to change that in the following maxim: Activate your pieces in order to penetrate in the enemy camp.
If I had to summarize the essence of endgame right now it would be: Activate your pieces and king in order to penetrate into the enemy camp. Yet the methods and subgoals are quite different.

(To be continued/updated . . .)


  1. fyi, im still checking in regulary...

    but very involved in bullet, my databases, and investment.

    if i can get my target bullet rating soon, you will 'see' and surely hear from me sooner rather than latter. a labor of love--and obscession. tons of three min games... that is to say, five or six minutes total. in nearing four hundred, last three weeks... a specialty for sure, among chessics, so no one just 'gives away' mercy points. :)

    keep up the good work, warmest,

    warmest, dk

  2. I wonder is this will distill down to Positional Placement leading to Material Advantage (Queen Promotion). I suspect in the future you will have some posts on Middlegame to Endgame transition the selection thru trades of what actors will appear in this Endgame Play. I suspect I am where you were a few years ago trying to understand the King vs King and Pawn nuances. I am finding this study to be quite interesting.

  3. Tempo, what you are doing is great, specially to me who also struggles in the endgame.

    This can easily become a high-quality source about endgame strategy, a topic almost never addressed by the GMs...

  4. Tak,
    I stopped with pawnendings for one sole reason: it had no connection with my games. I hope this new knowledge will fill the gap. I suspect this is what Capablanca meant by "learn the endgame first". And ofcourse, once this is mastered I will probably end up thinking about how to play the middlegame without frustrating the following endgame.

    But I use the principle of two weaknesses: the endgame is only one of them. The other is "how to treat complex tactics". And I haven't the slightest idea how to approach that. But first things first, that is to say, the end first.

  5. GM,
    welcome back!
    I have no idea why grandmasters don't treat this part of the game. Maybe because you will only become a grandmaster when this part of the game is obvious to you. Because now I start to formulate things, I must admit it is actually rather obvious.

  6. Tempo-
    thanks for writing this up, it is very helpful! I am pretty confused by endgames right now. On one hand, I fully agree with you that the so-called theoretical endings have little in common with the problems I am typically facing in my games. On the other hand, I am afraid that I first have to understand the theoretical stuff to be able to drive the game in a practical situation towards a won theoretical endgame.

  7. Sciurus,

    On the other hand, I am afraid that I first have to understand the theoretical stuff

    I shouldn't worry about that too much. Endgame books are reference books. Say, you are a pawn up in a rookending and you don't know the Lucena. Then you draw and when you come home, you look it up what you should have played. That way you wil never forget it again.

    When I learned th KG I knew only a few moves. When I was out of book, which was pretty soon, I had always reached a nerve wracking position, which I sometimes lost and sometimes won. Later at home I looked up where I got wrong. That way you learn the important lines in no time. I suspect it will work the same in endgames.

    All the so called endgame experts I have spoken were just one-eyed kings. They told me that they had little theoretical knowledge. I suspect they have just a feeling for the part of the game I describe in this post.

  8. Then you draw and when you come home, you look it up what you should have played. That way you wil never forget it again.

    Good point. Why not approaching endgames in a similar way I as I learn openings? I'll think about that.