The London System: a complete guide, its plans, and how not to play on autopilot

Short answer

The London System sets up the same structure against almost everything — pawn on d4, bishop on f4, e3, c3, Bd3, Nbd2 — but the plan that follows changes completely with Black's reply. Against ...c5 it is the b-file and the trade on c5; against ...Bf5 it is Qb3 hitting b7; against ...g6 it is the attack with h4. Playing the same continuation in all three is what turns the London into autopilot and drains its strength.

The essentials in a few lines

• The modern order is 1.d4 and 2.Bf4, getting the bishop out before Nf3 — which changes which Black defences are available.
• ...c5 followed by ...Qb6 is the sharpest test, and the correct answer is not b3: it is 3.e3 Qb6 4.Nc3, leaving b2 as bait.
• Against ...Bf5, Black solves their own bad bishop; the reply is c4 and Qb3, hitting b7 and d5 at once.
• Against the King's Indian setup with ...g6, the f4 bishop already eyes the right diagonal and h4-h5 is a real plan, not an adventure.
• The standard London attack — Ne5, f4, Rf3-h3 and the sacrifice on h7 — only works after Black castles short and with no counterplay in the centre.

What the London System is and why it works

The core idea is to solve, on move two, the problem that plagues half of all 1.d4 openings: where to put the dark-squared bishop. After 1.d4 d5 2.Bf4 it stands outside the pawn chain on its most useful diagonal, and only then does e3 come to support the centre. Reversing that order leaves the bishop stuck and produces a worse version of the Colle System.

What you gain in economy you pay for in ambition. The London rarely produces an opening advantage against accurate play; it produces a playable position with a clear plan, where the player who knows the structures decides the game in the middlegame. That is excellent for anyone with two hours of study a week and poor for anyone expecting pressure from move eight.

The modern move order matters more than it looks. Playing 2.Bf4 before Nf3 avoids some lines where Black pins the knight with ...Bg4, but it opens the door to an immediate ...c5 and ...Qb6. It is a conscious trade of one nuisance for another — which is why the next section exists.

Move order and the ...Qb6 problem

This is the sharpest test of the modern move order, and it is worth understanding the mechanics rather than memorising them. The queen on b6 hits b2 and d4 at once, exploiting the fact that the bishop has left c1 and the b2 pawn has no natural defender. The answer 4.Nc3 accepts the material loss because after 4...Qxb2 the black queen is far from the game and short of squares: 5.Nb5 threatens Nc7+ forking the a8 rook, and every move Black spends escaping is a move White spends developing.

The honest alternative deserves a mention: anyone who would rather avoid that discussion can simply play 2.Nf3 before Bf4, delaying the bishop until ...Qb6 loses its sting. The cost is allowing ...Bg4 and some lines with ...Nf6 and ...e6 where Black equalises comfortably. There is no free move order.

Against ...c5 without ...Qb6, White's plan is different and more positional: support d4 with c3, keep the tension, and capture on c5 at the moment Black's recapture leaves a weakness — usually with the d6 pawn or the bishop, which opens the b-file for pressure with Rb1 and a4-b4.

Three plans, not one

This is the answer to the criticism that 'the London is always the same'. The piece setup is the same, but a setup is not a plan. What changes is which piece takes the lead: in the king attack it is the e5 knight and the d3 bishop; in queenside pressure it is the b1 rook; against ...Bf5 it is the queen on b3, doing two jobs at once; against the King's Indian it is the h-pawn.

Recognising which of the four plans is on the table is the real skill in this opening, and it is trained in a specific way: by reviewing your own games and classifying each one by Black's reply before looking at the result. After twenty classified games the pattern appears by itself — almost everyone plays two of the four scenarios well and the other two badly.

The standard king attack, and when it is not worth it

The knight on e5 is the key piece and deserves attention. It stands advanced, in the centre, protected by the f4 pawn and eyeing the two squares that matter around Black's king — g6 and f7. While it is there, the d3 bishop has real value; if Black manages to trade it with ...Nxe5 and the recapture is with the d4 pawn, White's structure gains space but loses the attacking piece.

The mistake that costs the most games in this plan is starting it too early. Black almost always has ...c5, and a flank attack with an unstable centre is the textbook definition of a failed attack. The correct sequence is to lock or stabilise the centre first — c3 and, if necessary, accepting the trade on c5 — and only then move pieces to the kingside.

There is a version of this attack that works without the sacrifice and pays better at club level: trade the dark-squared bishops (with Bh6 against the King's Indian setup, or Bg5 after ...Nf6) and invade the dark squares with queen and knight. It is less spectacular and far easier to calculate correctly with the clock running.

• Ne5 supported by f4, not a loose Ne5 — without the pawn, the knight is chased off by ...Nd7 or ...f6.
• Bd3 already aimed at h7 before the queen comes out; the reverse order gives time for ...h6 and ...Nh5.
• Rf1-f3-h3 is the rook's route; it only makes sense once f2-f4 has vacated the square.
• The h7 sacrifice requires the f6 knight to be deflected or traded — it is not automatic.
• With ...c5 in the air and the centre loose, postpone the attack and settle the centre first.

The mistakes that turn the London into autopilot

The f4 bishop is the piece that gives the opening its name and its point, and trading it for a knight without structural compensation is the quietest mistake on the list. After ...Nh5, for instance, retreating to g3 and then allowing the trade on g3 hands Black the doubled pawns White wanted to avoid and solves Black's development problem. Retreating to e5, playing g3 with h3 prepared, or simply recapturing with the h-pawn are decisions that must be made consciously.

The second quiet mistake is automatic castling. Against the King's Indian setup, castling short puts the king on the side where Black will push pawns, and the h4 plan (which only exists while the king is still on e1) becomes unavailable. It is the same reasoning the Caro-Kann demands of Black in the Advance Variation: with a closed centre, castling stops being urgent and becomes a choice.

Finally, the mistake that shows up as early as move two in club games: playing 2.e3 before Bf4. It looks like a detail and it is the whole opening — with the bishop still on c1, Black plays ...Bf5 or ...Bg4 and solves their own problem first, leaving White with a Colle without the Colle's advantages.

• Bf4 before e3, always. The reverse order gives the opening away.
• Decide in advance what to do about ...Nh5 — retreat, allow it, or prevent it with h3.
• Classify Black's reply before choosing the plan: ...e6, ...c5, ...Bf5 or ...g6.
• Against ...g6, consider not castling short: h4 is the plan, and it needs the king on e1.
• Flank attacks only with the centre settled; otherwise ...c5 arrives before your Rh3.

How to train the London in two weeks

The London's advantage as a study object is that there is a single variable: Black's reply. That makes self-assessment simple and honest, which is rare in opening study. Across twenty classified games, how your score distributes among the four scenarios tells you exactly where the next hour of study belongs.

The sign that the problem is one of plans rather than tactics is this: you reach move twelve with a normal position and move twenty with nothing to do. When that repeats in the same scenario, what is missing is not calculation — it is knowing which piece is the protagonist in that structure.

One last repertoire note: the London covers 1...d5 and 1...Nf6 well, but anyone playing 1.d4 will also meet the King's Indian, the Grünfeld, the Dutch and the Benoni. It is worth deciding from the start whether the London is the entire repertoire — with the anti-...g6 plans properly studied — or only the answer to ...d5 defences.

Frequently asked questions

Openings mentioned in this guide

• London System (D02)
• Trompowsky Attack (A45)
• Colle System (D05)
• King's Indian Defense (E60)