Sunday, November 16, 2008

It's hard to be a penquin.

I'm experimenting a bit with templates since blogger keeps messing up my texts.


























Let's see if we can find some reasoning behind trading towards an endgame.

For reasons of convenience I start with the most common trades. That are the trades of pieces of equal value.


B x B
B x K
K x B
K x K
R x R
Q x Q

Underlying pawnending is a win.
Since the value of the pieces is equal, there must be something else that determines if a trade is beneficial. That beneficiallity is based on the underlying pawn ending. You have to ask yourself first "is this position won if there were no pieces on the board". I said a few things about that here and here some time ago. If the pawnending is won, you have a sustainable advantage and you can start to trade pieces until there are no pieces left.

Underlying pawnending is a draw.
If the pawnending isn't won but a draw, there are other matters that play a role to determine if a trade is beneficial:


  • activity of the pieces.
  • does the trade alter the pawnstructure.

Piece activity.

In the ending the piece activity is two fold: attacking, can a piece help to clear the road to promotion, and defensive, can your piece restrain or block an enemy pawn to prevent it's promotion. This piece activity is mainly influenced by the pawnstructure. Your own pawns block your own pieces, and squares that are under attack by hostile pawns are usually forbidden. The idea is to trade your pieces in such way that you keep the active ones while your opponent is saddled with the passive ones.

But before you can decide on trading you have to ask yourself wether the pawnstructure is fixed or volatile. If the position is very volatile because the pawns are not restrained or blocked, the piece activity can easily change by just moving pawns.

Pawn structure.

If a piece is defended by a pawn, the trade of equal valued pieces alters the pawnstructure. This can effect the outcome of the underlying pawnending and/or the piece activity. This should tell you if the trade is beneficial or not.

It is all so logical that I wonder why I haven't thought of this before.

8 comments:

  1. Good stuff, very helpful: it would have helped me a lot in a couple of recent tournament games! I'll have to print this and tattoo it on my arm.

    ReplyDelete
  2. "If the pawnending is won, you have a sustainable advantage and you can start to trade pieces until there are no pieces left."

    Nope.

    Consider, for example, the case where you have three pawns against two on the same side of the board, with a Rook and a Knight each. You do NOT want to trade the Knights.

    The question is always, "trade _which_ pieces, and why?" Nothing in Chess is automatic.

    ReplyDelete
  3. Ed,
    you clearly have knowledge I haven't. Can you tell me why I shouldn't trade the knights or where I find such information myself?

    Maybe you can give an exact position so that I can find out why my reasoning is wrong?

    ReplyDelete
  4. I tested Ed's statement with the aid of the Nalimov tablebases.

    KRppp vs KRpp = draw
    KNppp vs KNpp = draw
    Kppp vs Kpp = win

    So that seems to support my (highly theoretical) presumption. Pieces on the board are an impediment for promotion.

    How to play it without the aid of a computer is another matter.

    ReplyDelete
  5. You can't trade Knights because R+3P vs R+2P all on the same side is normally a trivial draw.

    A position: White, Kg1, Ra1, Nf3, Ps f2,g2,h3. Black, Kg8, Rf8, Nf6, Ps g7, h6. A Pawn ending is a fairly easy win. A Knight ending is a good chance for a win. A Rook ending is a dead draw.

    ReplyDelete
  6. I'm also curious about Ed's claim. Clearly he is right that it would sometimes be good to keep some material on the board (there are no exceptionless generalizations about chess).

    Yes, I intended the joke in my last sentence.

    ReplyDelete
  7. Also, if N+2 vs N+2 is too problematic a win, consider R+N+3 vs R+B+2, where the minor-piece ending is very superior while the Rook ending is drawn, or R+N+4 vs R+minor+3.

    Even trading Queens isn't automatic: in Q+R+3 vs Q+R+2, trading Rooks leaves some winning chances; trading Queens, not so much.

    Even with Pawns on both sides of the board and an extra Pawn, in a R+minor or Q+R ending you'd often rather not wind up with just the Rooks left. (There will be an example in my next post, which should be up Wednesday.) Not to mention endings including opposite Bishops.

    And so on! You always need to consider what the best pieces are to trade off, and what to keep, based on the particular position. Chess doesn't lend itself to very simple operational rules like "winning pawn ending , start trading pieces".

    ReplyDelete
  8. Yes, R+3p vs R+2p on one wing is not hard to hold to a draw, unless the attacker can create a passed pawn on the knight or bishop files (eg, if the pawns are advanced via a typical breakthrough.) For the stronger side, it would make sense to keep minor pieces on-board so as to complicate the defense.

    In "How to Play Chess Endgames", Mueller encourages caution anytime you trade, but especially when entering a pawn or rook ending. Of course, you can't fully implement this unless you are familiar with a variety of endgame positions.

    So develop that knowledge! Barring specific knowledge, I think tempo's version of "When ahead (enough to win the pawn ending), trade pieces not pawns" is a helpful maxim in most cases. :)

    ReplyDelete