Thursday, June 08, 2006

Target Stats

In seven games, I have identified 45 significant targets. Roughly half of them are pawns and half of them are pieces. About 90 percent of the targets were not protected at all, or underprotected. About 10 percent were just exposed to an attack or in danger of being trapped.

Thus, the most likely target will be a piece or pawn not protected at all, or not sufficiently protected. No protection and insufficient protection occur in roughly equal amounts. The insufficient protection, if present, will be by another piece, and here K and Q are overrepresented because of their limited protective power. Pawns are underrepresented, because they normally do a god protecting job if they are not pinned or under attack themselves.

Of course, K and Q are targets all the time, but not always significant ones. For a target to become significant it must be vulnerable to an attack, and this attack must gain advantage. Just any unprotected pawn or any piece only protected by the Queen are no significant targets. Nor are Kings in safe castles nor Queens exposed to fruitless attacks.

A target most likely becomes significant by combination. That is, forming a geometrical pattern (fork, skewer, pin) with another target or being «guarded» by an overworked piece. If a single target becomes significant, it is most likely the King under a mating attack or a trapped piece.

In conclusion: A piece or a pawn have about equal odds to become a significant target. Any piece or pawn not protected by a pawn is a good target candidate. The higher the value of the protector, the higher the odds of becoming a significant target. Two targets on squares of same color have higher odds to become significant. If significant targets are on squares of different colors, most likely an overworked piece is involved. I have the impression that best candidates of overworked pieces are King and Queen, but my data base is too limited so far.


At 3:52 PM, Blogger Temposchlucker said...

That's an interesting way of looking at it. It remembers me of Nunn's "LPDO": Loose Pieces Drop Off.

At 4:43 PM, Blogger Mousetrapper said...

Exactly. And it pays to define a Loose Piece in the widest sense. Not only a piece undefended at all, but also a piece attacked as many times as defended.


Post a Comment

<< Home