Sunday, December 16, 2007

Pawns, initiative and tempos

Disclaimer: my thoughts haven't cristallized yet. Changes of view can happen during the writing. To simplify matters I'm talking only over the opening and the middlegame.

My previous post lead me to the conclusion that there can be only one reason for a pawnmove: it's effect on the relative piece activity. So far as I can tell there can't be no other justification for a pawnmove. To simplify matters I'm not talking about pawn promotion. A pawn is the trickiest piece since a move has such complex effect on the activity of all pieces on the board. There is a simple method to assess the effect of a pawnmove though. By looking to all pairs of pieces from you and your opponent. Compare the dark squared bishop of yourself and of your opponent before and after the pawn move. Is the activity of your bishop improved relatively or not?

If the above is true, then automatically another question is raised. "what exactly is the effect of the initiative?" In the begin position all pieces except the knights are restricted by the pawns. If white makes reasonable moves and black mirrors them exactly, there can't be a difference in piece activity on both sides. Yet the game can be decided quickly if black continues to copy white's moves. Take for instance the following nonsensical sequence:

1.e4 e5 2.Nf3 Nf6 3.Nxe5 Nxe4 4.Qe2 Qe7 5.Qxe4 Qxe5 (see diagram)

White to move.

Despite identical piece activity black is lost. Since white had the initiative. But maybe that is not the reason. If white on move 4 would had put his knight en prise with 4.Nc6, black would have won the game. Sorry for the nonsense moves but extreme examples help to clear the picture.

None of the given examples can be forced though. Usually the flow of a won game goes something like: opponent A manages to improve his relative piece activity. At a certain moment opponent B cannot keep up with all posed threats and must give up material. So the act of forcing has a relation with the act of improving your piece activity. Hence with the moves of your pawns. Yet anytime the outcome of the game can go in any direction if one of the opponents makes a gross mistake.

There seem to be 3 stages.
First stage, the opening.
Initially there is no contact between the forces. The first task is to bring your pieces to the frontline. This is the mobilization of your army and is called development. During development every move counts. If you trade a piece with 4 invested tempos against a piece that hasn't moved yet, you will be behind in development.

Second stage, the middlegame.
After the opening tempos seem to play a different rol. If you trade off a knight in which you have invested 21 moves against a piece that has only moved once, you can't say beforehand that you are worse. It depends solely on the position on the board. Yet tempos do play a role.
Take for instance the following position from the Polar Bear (see diagram):

Black to move.

The development of both sides is about equal. But if you look at the pawnstructure, then you see that the black pieces are very restricted by the white pawns. The knight on c6 is developed. But it doesn't do anything useful. It only can go to a5, from where it is quite restricted by white pawns too. You can be sure that it will take several tempos before that knight does something usefull. This means that tempos are related to something usefull. If white can do something usefull with all his pieces (which is not the case in this position) and black needs 5 tempos before all his pieces swing into action, than white is better. This raises the question "what is something usefull?" The answer is probably something like "inducing or threatening a weakness"
Here you find an argument to stop treating the opening as a different phase as the middlegame. What use is it to develop a piece to a square where it still does nothing? A piece is only usefull when it does something usefull. No matter how actively placed it is. Tempos indicate how much moves you need before a piece does something usefull. And now follows what I was after: pawnmoves have an influence on how many tempos pieces are away from usefullness.

Third stage, a forced end.
All of a sudden it is over. The scale has tipped and you can force a win. When you are a queen up you can win from Kasparov in most positions. The end is not really forced in the way that there is no choice of moves by your opponent, but no matter how he plays you can force the win. To enter this sudden death you need a bad move. If this bad move isn't forced, it is called a blunder. If it is forced, we have outplayed our opponent in the middlegame.

I have the idea that there is something very important in all this, especially in this tempo stuff in relation to pawnmoves, but it appears not quite clear to me yet.


  1. These are the types of revelations that make you realize how little you actually understand about chess, especially the middlegame. In my last comment I was trying to point out that although relative piece activity is something to think about it is impossible to actually be able to wrap your head around the effect of a pawn move to the point where you will be able to objectively evaluate this activity. Except of course in critical or forcing situations,
    there is a subjective aspect that makes you choose any such continuation over another. Speaking of extreme examples, try to evaluate objectively the difference between 1 e4 and 1 d4.

    Just being aware of the ideas that you are presenting here is the achievement, if you are concious of these ideas during your play you will find the epiphany you are looking for.

  2. Drunk,
    I understand now what you mean. Yes, it's true that there is an element of subjectivity in the valuation how much each pieces is improved, worsened. And how an improved knight weights against a worsened bishop etc..

    If the criteria is "piece activity" you can imagine a computerprogram that expresses it in a number and simply calculates it. But in this post piece activity is related to "something usefull" and how do you express that in a number?

    Understanding the difference between 1.e4 and 1.d4 is a good training. Pawnmoves decide the character of the game.

  3. My previous post lead me to the conclusion that there can be only one reason for a pawnmove: it's effect on the relative piece activity.

    Some rambling thoughts:

    There are direct tactical considerations, pawn forks and the like. But I suppose that if you interpret "piece activity" broadly enough then winning a piece falls into "piece activity." As does checkmate by moving a pawn.

    Conversely King safety is a reason to move a pawn, or not. Again, with a broad enough definition "piece activity" is covered by that.

    But I would prefer to say that piece activity is an important consideration.

    Also, if piece activity is the reason to gambit a pawn then material is the reason to accept the pawn and *less activity* and those all likely involve pawn moves.