Binding

Tarrasch's rule (rooks belong behind passed pawns, yours and your opponent's) is really a binding rule. Look at the diagram. The white rook behind its pawn gains scope with every step the pawn takes: the file opens up behind it. The black rook, blockading on the queening square, loses scope with every step, until on the last rank it has no squares at all and any rook move promotes the pawn.

One rook improves by doing nothing. The other degrades by doing nothing. That is not a proverb, it is geometry, and it is why the pawn cannot promote by itself but the position still wins: your king simply walks up the board and attacks the blockader, which cannot run.

The same shape appears with a knight in front of a pawn, in our “Knight Blockade” drill: the knight holds the draw, but it is glued in place for the rest of the game. Binding is not always fatal. It is always expensive.

The second form of binding is defensive obligation. A piece that must keep protecting a pawn, or covering a promotion square, is a piece that cannot go anywhere else, and in an endgame with three or four pieces on the board that means you are playing effectively a piece up on the other wing.

The mirror image teaches the rule best. In our drill “Defend the Passed Pawn From Behind”, a knight guarding its own pawn from behind costs nothing: capturing it drags the enemy king away from the pawn's path and the pawn simply queens. The same knight guarding from the side would be tied down and vulnerable. Rear for your own pawn, front for the opponent's: two different jobs, two opposite squares.

The practical instruction that falls out of this: when you have the advantage, look for the piece with a job, and start the second front on the other side of the board. When you are defending, look for the defensive setup where your pieces guard things passively but freely rather than being pinned to a single square.

Binding can be symmetrical, and then the position is decided by the kings. Our drill “The Rook in Front of Its Pawn” is the classic: a white rook on a8 shielding a pawn on a7 has zero moves and no future, but the black rook behind the pawn is equally chained to the a-file forever. Two bound rooks, and the entire evaluation comes down to where the defending king stands: reach the g7/h7 zone and it is a draw; stand one file too close and there is a skewer break.

That is the useful shape of the idea. Binding does not evaluate a position by itself: it tells you which pieces are still playing. Count them. If your opponent has a rook that cannot move and a knight that cannot leave a square, you are effectively two pieces up wherever the action is, and the plan writes itself.

And if it is your piece that is bound, then the plan is the opposite: free it, or accept the fortress and hold. Bound pieces are what make the fifty-move rule your friend.