What Is a Plan in Chess, and Why Does Yours Not Work?

"Play with a plan" is one of the most repeated pieces of chess advice and one of the least useful, because it almost never comes with an explanation of what counts as a plan. The result is players who think they have one and actually have an intention.

What is not a plan

None of these sentences is a plan:

• "I am going to attack the king."
• "I am going to improve my position."
• "I am going to apply pressure."
• "I am going to wait and see what they do."

They are directions, not plans. They cannot be checked, they do not tell you what to play right now, and above all they do not tell you when they are finished.

What is a plan

A real plan has three properties:

1. It is concrete. It names pieces and squares. "Get the knight from f3 to d5 via e3" is a plan. "Activate my pieces" is not.

2. It is short. Two to five moves. Longer than that almost never survives, because someone on the other side is playing too.

3. It can be checked. When it is over, you can say whether it worked. That is the property most often missing: a plan you cannot evaluate is a plan you cannot correct.

Examples of real plans:

• "Trade off their dark-squared bishop with Bh6, because it defends the castled king."
• "Open the c-file with c4-c5 so the rook has somewhere to go."
• "Get the knight to d5, a square none of their pawns can attack."
• "Trade queens to reach an endgame where their a6 pawn is weak."

Each one either happens or does not within three or four moves, and then you pick another.

Where the plan comes from

Almost always from two places, and neither is "what I feel like doing":

From the pawn structure. Pawns change least, so they dictate what is and is not possible. Is there an open file? That is where the rooks go. Is there a pawn chain? You play towards where it points. Do you have a majority on one wing? That is the wing to advance on.

From your opponent's weaknesses. An isolated pawn, a square no pawn can defend, a king with a broken shield, a piece with no squares. The plan is to attack that, and the first step is almost always putting a piece in front of or on top of the weakness, not capturing it immediately.

The three errors that ruin a good plan

1. Executing it when it no longer applies

The most common one. You pick the plan on move fifteen, your opponent does something that changes the position, and you carry on with the plan anyway because it was the plan.

A plan is a hypothesis, not a commitment. Every two or three moves you have to check whether it still makes sense.

2. Attacking with too few pieces

An attack on the king needs three pieces pointing at the same place. With two it almost never works, and an attack that dissolves leaves your pieces badly placed for the rest of the game.

If you have two pieces and want to attack, the right plan is to bring the third, not to attack.

3. Forgetting your opponent

The most expensive error. A five-move plan of yours against a three-move plan of theirs loses, however good yours is.

Before starting any plan: what are they doing and how long does it take?

How to pick a plan in practice

When the position is quiet and you do not know what to do, this sequence produces a reasonable plan nearly every time:

1. Which is my worst piece? Improving it is a valid and sufficient plan.
2. What is my opponent's clearest weakness? Putting a piece in front of it is the first step.
3. Which pawn break is available? Preparing it is a three-move plan.
4. Is there a trade that favours me? Trading their good piece for your bad one is a complete plan.

None of them is spectacular. All four win games.

Five plans that cover most positions

In practice, almost every plan you will need below 1800 is a variation of five. Knowing them by name speeds the choice up enormously, because the question stops being "what do I do?" and becomes "which of these five fits here?".

Occupy the open file. Rook to the file, double the other rook behind it, invade the seventh. It is the most frequent plan in chess and the one that wins the most games among club players, because a rook on the seventh attacks pawns and boxes the king in at the same time.

Plant a knight on an outpost. Find the square in enemy territory that no pawn can attack, and walk the knight there in three moves. A well-planted knight paralyses a whole position.

Attack the base of the pawn chain. When your opponent has a chain, the link that matters is the rearmost one: it is the only one not defended by another pawn. The plan is to bring pieces and a pawn of your own against that square.

The majority on one wing. If you have three pawns against two on one side, the plan is to advance them and create a passed pawn. It is slow and it is one of the soundest plans there is, because the passed pawn still counts in the endgame.

The pawn storm against a castled king. When you have castled on opposite sides, you push the pawns on the wing where their king is. It is the most aggressive plan and the most dangerous to handle, because every pawn that advances also weakens something.

If none of the five applies in a position, almost certainly the position is not asking for a plan: it is asking you to improve your worst piece and wait.

What a written plan looks like

An exercise that teaches far more than it appears to: write the plan out in one sentence, during the game. Not in your head — written, on a piece of paper beside you.

A good written plan looks like this:

"I take the knight from f3 to d2 and from there to c4, to attack the pawn on b6 and occupy

d6."

It has three parts and all three are needed: which piece, where to and what for. If, when writing it, you cannot complete the third, it was not a plan.

A bad written plan looks like this:

"I attack on the kingside."

It does not say with what, it does not say against what, and therefore it cannot be checked for whether it worked. That is the most useful test of all: a plan that cannot be evaluated within four moves is a wish.

After twenty games writing the plan out, the habit is internalised and the paper is no longer needed. And something better than the habit remains: a record of your own plans, which on re-reading shows with uncomfortable clarity which ones were wishes.

What to do when your opponent has a plan and you do not

It is the most frequent real situation and almost never explained: the other side is doing something coherent and you are reacting. Three things, in this order:

Name their plan. Before moving, say in one sentence what they are doing. Half the time, naming it reveals it is far less dangerous than it looked; the other half, at least you now know what you are playing against.

Look for the move that stops it, not the one that dodges it. If their plan is to land a knight on d6, the question is what controls d6. A concrete plan is stopped at a concrete place, and that place is almost always a square.

And if it cannot be stopped, counterattack elsewhere. It is the classic answer and the right one: a position where your opponent advances on one wing is defended by creating problems on the other, because pieces cannot be in two places.

What does not work — and it is what almost everybody does — is defending passively move by move without naming anything. That hands your opponent the one resource that genuinely decides a game between evenly matched players: the initiative.

Chaining short plans

Strong players do not have a master plan lasting twenty moves. They have a chain of short plans, each chosen from the position the previous one left:

I get the knight to d5. Now that it is there, my opponent had to weaken e6. Now I attack e6 with the rook. Now that the queen defends e6, their kingside has no defender: now I attack.

From outside, that chain looks like one long, brilliant plan. From inside it is four three-move plans, chosen one after another.

And that is probably the most important part: you do not need to see twenty moves ahead. You need to choose the next three well, over and over.

Related questions

How many moves should a plan last? Three or four. Ten-move plans do not exist outside books: your opponent plays too, and any long plan breaks along the way. Chaining short plans is what, in practice, resembles having a long one.

Can I play well without plans, just reacting? Up to a point yes, and you can see exactly where it stops being enough: in quiet positions, where there is nothing to answer. There, somebody without a plan starts moving pieces without criteria and weakens themselves.

How do I know whether my plan is too ambitious? If it depends on your opponent not doing something obvious, it is a wish. A legitimate plan works even against their best reply, even if it works less well.

What do I do if the plan fails halfway through? You abandon it without emotional cost, re-evaluate the position and choose another. Insisting on a plan that stopped applying is the commonest way of turning an equal position into a worse one.