Plans and prophylaxis

A game on lichess.org, 2014. White to move. The black queen on c4, the bishop on d3 and the rook on a1 have crept up to the white king. White can take the d3-bishop with the c2-bishop, but first it is worth asking: what is Black going to do?

The move Bc3 takes nothing and threatens nothing. It is needed because White first looked at the position through the opponent's eyes. This way of thinking is called prophylaxis, and it is the second half of this chapter. The first half is about what to do when there is no mate and no tactic on the board: how to choose a goal and make a plan.

How to make a plan

A plan is a goal several moves ahead and the way to reach it. It follows from the evaluation of the position. Look at the pawns first: where the opponent has weaknesses, where the open files are, on which wing you have more pawns. Then at the pieces: which stand well, which stand badly, where the kings are. From this comes a goal: attack a weak pawn, occupy an outpost, carry out a break, steer the game into a favourable endgame.

A good plan is concrete. “Attack on the queenside” is not a plan yet. “Put a rook on b1, march the b-pawn to b5 and, by exchanging on c6, leave Black a weak pawn” is a plan: it has a goal and the moves that lead to it. A plan like that can be checked. If the opponent has time to defend or to strike back, you will see it in advance and choose another.

A plan does not have to be long. Masters rarely work out a plan twenty moves deep. Usually it is two or three immediate tasks: reroute a knight, double the rooks, prepare a break. Once the tasks are done the position has changed, and it has to be evaluated again. The only thing worse than a bad plan is no plan at all: aimless moves scatter the pieces at random while the opponent gets on with their own business.

Compare your plan with your opponent's. They have a goal too, and often the two plans are a race: you attack on one wing, they attack on the other. Then you have to count who gets there first. If the opponent is faster, it may pay to hinder them first and come back to your own plan later. If your plan needs five moves and the opponent threatens something in three, the plan has to change.

One of the most reliable plans is to trade into a favourable endgame. If you have the better pawn structure, the more active king or an extra pawn, every exchange of pieces brings the win closer: in the endgame these small things become decisive. The side with the weaknesses, on the contrary, tries to keep the pieces on and look for play.

The worst piece

When it is unclear what to do, there is a simple rule: find your worst piece and improve it. The worst piece is the one that does least: a knight on the rim, a bishop running into its own pawns, a rook behind a closed file. Move it to a good square and the whole position gets better, even if nothing concrete is threatened.

The engine can measure this directly. The term “Piece placement” scores how good the squares of the pieces are, and “Mobility” how many squares they can reach. On the board below the position is almost level, but the white knight stands on h3, on the edge of the board. Play Nf4 and see how the evaluation and its terms change. Then try other moves.

While the knight stands on h3, the evaluation is about half a pawn in Black's favour: from h3 the knight reaches few squares and stands far from the centre. After Nf4 both rows have gone up, and the evaluation is almost level. White has not won a single pawn, but the position is half a pawn better. For comparison, press “Back to the position” and play Ng5. The knight also leaves the rim and the evaluation rises, but by less: f4 is closer to the centre than g5.

Sometimes improving one piece takes several moves. A knight goes from b1 via d2 and f1 to g3, a bishop switches from one diagonal to another, a rook lifts to the third rank and swings to the other wing. That is fine as long as the opponent cannot do anything in the meantime. Before such a manoeuvre ask yourself: where will this piece stand in three moves, and what will it do there?

The rule works for the opponent too. Their best piece is often worth trading: if an enemy knight has settled on an outpost in your camp, trade it off, even if it costs you your good bishop. And their worst piece is sometimes best left on the board. The opponent's bad bishop, shut in by their own pawns, works for you. Do not help them get rid of it.

A bad bishop can be improved in two ways. The first is to bring it outside its own pawn chain before the chain closes: in the French Defence Black often does exactly that, taking the c8-bishop via d7 to b5, or to a6 after …b6, to trade it for White's bishop. The second is to move the pawns to squares of the other colour so that the diagonals open. The second way is slower and creates weaknesses, so it is used less often. More about bishops is in the chapter “Bishop against knight”.

A plan from the pawn structure

Pawns move slowly and never move back. So a pawn structure lasts a long time, and the plan that follows from it often holds for many moves. For the typical structures such plans have been known for a long time. Knowing them is like having a hint: you see familiar pawns and know at once where to play. Structures were covered in the chapter “Pawn structure”.

A classic example is the minority attack in the Queen's Gambit. After the exchange on d5 White has two pawns on the queenside against Black's three. You would think the side with more pawns should advance. But White goes forward with the minority: the b-pawn reaches b5 and exchanges on c6. After the exchange Black is left with a weak pawn on c6 on a half-open file, and the white pieces attack it.

Notice that White's plan took many moves, and almost none of them threatened anything. But every move led towards the goal. Black is not idle in this structure either. Their typical plan is play on the kingside: a knight goes to e4, the f-pawn goes forward. Whoever carries out their plan faster and more precisely gets the advantage.

Other structures have their own plans. With an isolated pawn on d4 White usually attacks: the pawn gives the pieces room, and the break d4–d5 opens the game. The side playing against the isolated pawn, on the contrary, trades pieces and puts a knight on d5, in front of the pawn. No pawn can drive it away from there. In the endgame the isolated pawn becomes simply a weakness. The same structure suggests a different plan to each side.

Prophylaxis

Prophylaxis is a move that hinders the opponent's plan before they have carried it out. The idea was brought into chess by Aron Nimzowitsch. Tigran Petrosian, the ninth world champion, built his whole style on prophylaxis: he noticed his opponents' threats so early that he never had to parry them.

Prophylaxis starts with a question worth asking before every move: what does my opponent want to do? Which of his moves is the most unpleasant for me? If it were his move, what would they play? Many losses begin with a player so busy with their own plan that they missed a threat. That is what would have happened to White in the first position of the chapter if they had simply taken the bishop.

The question is useful even when the opponent threatens nothing right now. What will they do if I spend a move on my plan? Where will their worst piece go? Which break are they preparing? It often turns out that the opponent has only one good idea. Take it away and they are left without a plan, which almost always means aimless moves from then on.

Prophylaxis takes many forms. You can block the line along which an enemy piece will come, as with Bc3. You can defend in advance the pawn they are going to attack, or take your king off the line of a future attack. You can stop an enemy pawn before it becomes dangerous. In the next position Black is a bishop up, but the white pawn on d6 is two steps from promotion.

Against a passed pawn the best prophylactic remedy is the blockade: put a piece directly in front of the pawn. While the pawn is far away this is easier: you have time to bring a piece over. The closer the pawn is to promotion, the fewer ways there are to defend. That is why a dangerous pawn is stopped early.

When there is nothing to do

There are positions where you have neither a threat nor a clear plan, and neither does your opponent. Then the best move is a quiet, useful one. It takes nothing and threatens nothing, but makes the position slightly better. A rook goes to a half-open file, the king steps off a diagonal where it could be pinned against, the h-pawn makes luft. Players say such a move is made “for nothing”: without a concrete aim, simply because the piece stands better there.

Such a move is also useful because it hands the move to the opponent. Sometimes they have no good moves themselves, and any move they make spoils something. Then the quiet move becomes a waiting move, and in the endgame it can put the opponent in zugzwang.

What you should not do is make moves that create weaknesses for no reason. A pawn pushed forward “just because” will never come back, and the squares behind it are weakened for good. If you do not know what to do, move a piece: a piece can come back, a pawn cannot.

Quiet moves matter most when your position is better. The side with the advantage need not hurry. If the opponent cannot undertake anything, first put all your pieces on their best squares, take your king out of reach of possible checks, deprive the opponent of their last active moves. Only then go for the decisive action. A hasty break in a good position often gives the opponent exactly the counterplay they were missing.

How do you learn to plan? Go through annotated master games: the notes explain not only the moves but the idea behind them. Before every move in your own games take two steps. First: what does my opponent want? Second: which of my pieces stands worst, and where should it go? These two questions will save you from many mistakes.

What's next

A plan often leads to an endgame: the pieces are traded, and kings and pawns remain. In a pawn ending every tempo decides the game, and its own rules apply. They are the subject of the next chapter.