Chess RU

Part IV · Strategy Chapter 28 of 40

Pawn structure

Pieces move from square to square, but pawns cannot go back. So the placement of the pawns changes slowly and sets both sides' plans many moves ahead. We look at the main kinds of pawns and structures: islands, the isolated pawn, doubled, backward and hanging pawns, and chains — where their weakness lies, where their strength lies and how to play with them.

Amateur 35 minutes
  • pawns
  • evaluation

What to know first: 27 · How an engine judges a position

What you will learn

  • recognise isolated, doubled, backward and hanging pawns and pawn chains
  • understand how each structure is weak and how it is strong
  • choose a plan from the pawn structure: what to attack and where to put the pieces

A game on lichess.org, 2014. The black bishop on b7 stands on the long diagonal, but Black's own pawn on c6 blocks it. Behind it, on the same diagonal, stand the white queen on f3 and the rook on h1. Black to move.

Black to move, a game on lichess.org, 2014. Win material.

…c5: the pawn leaves the diagonal, and the b7 bishop attacks the queen on f3. The queen moves away, and …Bxh1 takes the rook. On top of that, the c5 pawn attacks the central pawn on d4. Full game.

One pawn move opened a diagonal and won a rook. Pawns change a position more than it seems: they open and close lines, take squares away from pieces and create weaknesses. And unlike a piece, a pawn cannot come back. A weakness created by a pawn move usually stays until the end of the game.

What pawn structure is

Pawn structure is the placement of both sides' pawns. It shows which files are open for the rooks, where each side has space, which squares are weak and which pawns will need defending. Strong players looking at a new position often start with the pawns: pieces can be rearranged in a couple of moves, pawns cannot.

The pawns decide where the rooks will play. A file without pawns is open, and a rook goes down it into the opponent's camp. A file where only the opponent's pawn is left is half-open: a rook presses along it on that pawn. So every pawn exchange also changes how the pieces play: lines open, bishops get diagonals, knights get outposts from which a pawn can no longer chase them.

Structures repeat. The same pawn placement arises from different openings: an isolated pawn on d4 comes from the Caro-Kann, from the Queen's Gambit and from the Sicilian. So once you have studied a structure, you know its plans for many openings at once. It is the same trick as in the chapter on the repertoire: learning shared structures pays better than learning separate variations.

In the last chapter you saw that the engine evaluates pawns as a separate group of terms. You also saw that each term is small: an isolated pawn costs the engine about a tenth of a pawn. That does not make the weakness trivial. It shows itself later: twenty moves on, when the pieces have been traded, a weak pawn can decide the game.

The first idea is the pawn island: a group of pawns on neighbouring files, separated from the other pawns of the same side by an empty file. The fewer islands, the better: the pawns of one island protect each other, while at the edge of every island a pawn has to be defended by pieces.

How many pawn islands do White and Black have?

White: a2, c3 and the group e2–h2, three islands. Black: a7–b7, d5 and f7–h7, also three. Black's d5 pawn is isolated as well, and White's a2 and c3 are isolated.

The isolated pawn

An isolated pawn is a pawn with no friends on the neighbouring files. It has two troubles. It cannot be defended by a pawn, only by pieces. And the opponent can occupy the square in front of it with a piece that no pawn can chase away: this square is called the blockading square.

The best-known isolated pawn is the queen's pawn, on d4 for White or on d5 for Black. It arises in many openings: in the Panov Attack against the Caro-Kann, in the Tarrasch Defence, in some lines of the Sicilian. See how it appears.

The starting position. Watch the c- and e-pawns: both sides will soon trade them. White has traded the e-pawn for the c-pawn. Now White has no e-pawn, and Black has no c-pawn. White attacks d5 with the c-pawn: the Panov Attack. After the exchange on d5, White's d4 pawn has no neighbours: White has no pawns on the c- and e-files. The square d5 in front of it is the blockading square, and a black knight stands on it.

An isolated pawn is not only a weakness. It gives space in the centre and two strong squares for the knights: for a white pawn on d4 these are e5 and c5. Next to it are the half-open c- and e-files, along which the rooks go. So the side with the isolated pawn usually plays actively: attacks the king while there are many pieces, and prepares the break d4–d5. The opponent wants to trade pieces, blockade the pawn with a piece on d5 and win it in the endgame.

The break d4–d5 is the isolated pawn's main weapon. The pawn is exchanged for Black's e6 pawn, the lines in the centre open, and White's active pieces suddenly get room: the rooks on the d- and e-files, the bishop on the diagonal to f7. If the break comes at the right moment, the isolated pawn disappears and the attack remains.

Against an isolated pawn the rule formulated by Aron Nimzowitsch applies: first blockade, then attack. A piece on the square in front of the pawn stops it, and the break is no longer possible. A knight blockades best: the blockade does not hinder it, and from the blockading square it still hits important squares in the enemy camp. Then the rooks press on the pawn along the file, and after the pieces are traded there is nobody left to defend it.

White has an isolated pawn on d4, and there are many pieces on the board. Which plan suits White?

While there are many pieces, the isolated pawn gives room: the squares e5 and c5, the half-open files. Its strength is in activity. If the pieces are traded, only the weakness remains, so the side with the isolated pawn avoids trades and looks for play.

Doubled pawns

Doubled pawns stand on the same file. The rear one blocks the front one, and they do not protect each other. In the endgame two doubled pawns are often worth no more than one. On the other hand, doubled pawns arise after a capture, and that capture opens the neighbouring file for a rook. Besides, doubled pawns cover more squares than one pawn.

Often the choice comes already at the capture: towards the centre or away from it. The rule says: capture with a pawn towards the centre. A pawn that comes to the centre controls more central squares, and the neighbouring pawns stay connected.

The Ruy Lopez: White has traded the bishop for the knight on c6. How does Black usually recapture?

After …dxc6 Black has doubled pawns on c7 and c6, but the d-file is half-open for them, the c8 bishop's diagonal is free, and they have two bishops against a knight and a bishop. This is the Exchange Variation of the Ruy Lopez: White hopes for a better pawn endgame, Black for the bishops. Here the rule "towards the centre" gives way to development.

The example shows that the rules about pawns are not laws. Doubled pawns are a weakness worth accepting if you get something tangible in return: an open line, the bishop pair, quick development. The engine, by the way, penalises doubled pawns very little, by hundredths of a pawn: in the middlegame they rarely decide.

Doubled pawns also occur in solid openings. In the Nimzo-Indian Black often plays …Bxc3+, and after bxc3 White has doubled pawns on c3 and c4. In return White gets the bishop pair and the centre, Black gets a target: they attack the c4 pawn with a knight from a5, a bishop from a6 and a rook along the c-file. Who turns out right is decided by the rest of the game: the structure by itself does not settle the result.

The backward pawn

A backward pawn has fallen behind its neighbours: they have moved forward and can no longer protect it, and the square in front of it is hit by an enemy pawn. It cannot move without being lost, and it remains a target. Even more important is the square in front of it: no pawn defends it, and the opponent puts a piece there.

The classic example is the Najdorf Sicilian with …e5. Black's d6 pawn is backward: Black has no c-pawn, the e-pawn has gone to e5, and the square d5 is hit by White's e4 pawn. The square d5 becomes the key to the game: White wants to put a knight there, Black wants to play …d5 and get rid of the weakness.

Why does Black agree to a backward pawn? Because …e5 gives them space and takes the squares d4 and f4 away from White's knight. The d6 pawn is defended, and if Black manages …d5, the weakness disappears. Pawn structure is always a trade: a weakness in one place for an advantage in another.

How do you play against a backward pawn? First occupy the square in front of it. In the Najdorf White brings a knight to d5 via c3 or f4 and tries to trade the pieces that can guard that square, above all Black's e6 bishop and f6 knight. When a knight stands on d5 and there is nothing to trade it for, and the d6 pawn is under pressure from the rooks on the d-file, Black has a hard time. That is why in this variation the fight for d5 starts from the first moves.

Hanging pawns

Hanging pawns are two pawns side by side, usually on c4 and d4 or on c5 and d5, with no neighbouring pawns. While they stand side by side they are strong: they hold four central squares and are ready to advance. But only pieces can defend them, and if one of them advances, the other becomes backward or isolated.

Hanging pawns have their own break too. The advance d4–d5 at the right moment opens the centre, and a pawn on d5 takes the squares c6 and e6 away from the opponent. So the side with hanging pawns keeps them side by side and waits for the moment to break, while the opponent presses on them and tries to force an advance too early.

White has hanging pawns on c4 and d4. Which plan suits Black?

Hanging pawns are strong while they stand side by side. If Black attacks them with pieces (rooks on the c- and d-files) and with the break …b5, White will have to either defend or advance one pawn. Then the other is left without support.

Pawn chains

A pawn chain is a set of pawns standing on a diagonal and protecting each other, such as White's d4 and e5 in the French Defence. The chain holds itself up, but it has a base: the rear pawn, which no pawn protects. Nimzowitsch's rule: attack a chain at its base. In the French, Black strikes at d4 with …c5, and White sometimes at e6 with f5. A chain can be longer, such as c3–d4–e5: then the base is c3, and it is attacked with …b5–b4.

The French Defence, Advance Variation. Which pawn of White's d4–e5 chain should Black attack first?

The d4 pawn is the base of the chain: no white pawn protects it. If it falls or is exchanged, the e5 pawn is left without support. That is why in the French Black nearly always plays …c5 and then attacks d4 with a knight, the queen and a bishop. Later, when d4 has weakened, the time comes for …f6 too.

A chain divides the board. The side whose chain points towards the opponent's king usually attacks on that wing. In the French the d4–e5 chain points at Black's kingside, and White attacks there. Black replies with play on the queenside, where they have more space.

The King's Indian is the same, only the other way round. White's d5–e4 chain points at the queenside, and White advances there with the c- and b-pawns. Black's d6–e5 chain points at the kingside, and Black undermines the base of White's chain with …f5, then marches the pawns at the white king. Both sides attack where their chains point.

The pawn majority

If on one wing you have more pawns than the opponent, that is a pawn majority. It can create a passed pawn: advance the pawns of the majority, starting with the one that has no enemy pawn in its path. Three pawns against two on the queenside is a common picture after exchanges in the centre. In an endgame, where the kings are far from that wing, such a majority often decides the game: the passed pawn lures the opponent's king away, while your king picks up pawns on the other wing.

How to choose a plan

The pawn structure suggests the plan. Look at which of the opponent's pawns are weak: isolated, backward, doubled. Put rooks against them, and knights on the squares in front of them. Look at your own weaknesses: they have to be defended or exchanged. And look at which breaks are possible: a pawn move that changes the structure in your favour is often stronger than any piece move.

One more rule: do not make pawn moves without a reason. Every such move weakens the squares the pawn no longer covers. If you do not know why you are moving a pawn, play a piece instead. This matters most in the opening and in front of your own king.

Trading all the pieces in your head also helps. Imagine that only the kings and pawns are left on the board. Whose endgame is it: who has more islands, who has weak pawns, who has a passed pawn or a majority? If you would be better in that endgame, trades suit you. If worse, look for play while the pieces are on the board.

Which piece move is usually strongest against the opponent's isolated pawn on d5?

A knight on d4 stops the pawn, no pawn can chase it away, and it stands firmly in the centre. The d5 pawn will advance no further, although all its strength lies in advancing. Blockade, pressure from the rooks along the d-file and trades: the classic plan against an isolated pawn.

What's next

Of all pawns, the most dangerous for the opponent is the passed pawn: one whose path is not blocked by enemy pawns. It heads for promotion, and only pieces can stop it. How to create one, how to support it and how to stop it is the subject of the next chapter.