Most openings ask you to remember what your opponent does. The London System asks you to remember what you do. White plays 1.d4, develops the dark-squared bishop to f4 before it can be shut in by e3, and builds the same small, solid structure against almost anything Black tries. It is the reason a club player can reach a comfortable middlegame against a stronger opponent without studying twenty move-orders — and, in the last decade, the reason world-championship players keep wheeling it out to sidestep an afternoon of preparation.
This is the encyclopedia reference: the idea, the move-orders, and the structures. For a practical, plan-by-plan guide to actually playing it, see the tutorial How to Play the London System.
A system, not a variation
The word that matters is system. A variation is a specific sequence of moves inside a named opening; a system is a formation White aims for regardless of what Black plays. The London formation is the bishop on f4, pawns on d4 and e3, the light-squared bishop on d3, a knight on d2, a pawn on c3, and short castling. White can reach it against a Queen’s-Gambit-style …d5, against a King’s-Indian-style …g6, and against a Benoni-style …c5, changing only the details.
That is the London’s great practical virtue and its theoretical limit at once. The virtue: you learn one structure and a handful of plans, and you are never out of your book, because your book is a shape rather than a line. The limit: because White commits to the same setup early, a well-prepared Black can choose the exact structure he is happiest to meet it with. The London does not claim an opening advantage by force. It claims a sound, familiar game — which, across thousands of amateur games and a surprising number of elite ones, is worth more than a theoretical half-tempo nobody remembers how to use.
Origins
The name comes from the 1922 London tournament, where several leading masters — including Capablanca — used the d4/Bf4 setup. For most of the twentieth century it lived on the respectable margins: a solid, slightly dull choice, the kind of opening a strong player used to make a quick draw or to avoid an opponent’s preparation. Its reputation was “safe but toothless.”
Two things changed that. First, engines demonstrated that several London structures carry more venom than their sleepy reputation suggested — particularly the lines where White delays e3, keeps options open, and meets an early …c5 or …Qb6 with concrete play rather than passive defence. Second, Gata Kamsky adopted it as a main weapon in the 2010s and showed it could be played for a win at elite level, and after him a wider group — Magnus Carlsen, Wesley So, and later Gukesh Dommaraju among them — used it to take strong opponents out of preparation. What had been a drawing weapon became a way to make a full game happen on White’s terms.
The main tabiyas
Three Black set-ups define modern London theory.
Against a classical …d5, the reference position arises after 1.d4 d5 2.Nf3 Nf6 3.Bf4 e6 4.e3 c5 5.c3: the textbook London triangle, with Black striking at the base of the chain.
Against a King’s-Indian …g6 — the move-order that gives this entry its A48 label, 1.d4 Nf6 2.Nf3 g6 3.Bf4 — White is happy: with Black’s bishop headed to g7 and biting on the d4-granite, the f4-bishop and an early h3/Be2 or Bd3 give a comfortable, flexible game.
Against an early …c5 and …Qb6, hitting b2 and d4 at once, White must know a concrete answer — usually a timely Nc3, Qb3, or the pawn sacrifice that follows …Qxb2. This is the one place the “just play your setup” advice breaks down, and where a London player earns the point by having looked at the details.
Plans and pawn structure
The London is a game of three recurring ideas.
The first is the knight to e5. Supported by the f-pawn’s absence of counterplay and often by Nbd2-f3xe5 or a direct Ne5, the outpost anchors White’s attack. If Black takes it, the recapture opens lines toward his king; if he leaves it, the knight is a permanent thorn.
The second is the kingside attack. With the centre closed by the c3/d4/e3 triangle, White frequently castles short and then throws the h-pawn forward — h4-h5 against a fianchettoed king — using the f4-bishop and a queen on the b1-h7 diagonal (after Bd3 and Qc2 or Qe2-e3) to generate real threats.
The third, quieter plan is the queenside squeeze: Ne5, f4 to support it, and pressure down a half-open c-file after the centre pawns are exchanged. This is the Capablanca-flavoured London, where White wins by accumulating small structural pluses rather than by attacking.
Black’s counterplay almost always involves the same levers in reverse: …c5 and …Qb6 against b2, the trade of the “London bishop” via …Nh5 or …Bd6, and a timely …Bf5 to neutralise White’s best minor piece before it can be traded on his own terms.
The Jobava cousin
Worth knowing is the London’s aggressive relative, the Jobava–Rapport System: 1.d4 Nf6 2.Nc3 and 3.Bf4, developing the queen’s knight to c3 instead of d2. Named for the Georgian Baadur Jobava and the Hungarian-turned-Romanian Richárd Rapport, it keeps the same bishop on f4 but adds ideas of Nb5 and a quick e4, trading the pure London’s solidity for sharper, more forcing play. It is a different opening in temperament, but it grows from the same bishop move, and many London players keep it as a change of pace.
How to study it
Learn the setup first, not the lines. Play the pieces to their London squares — d4, Bf4, e3, Bd3, Nbd2, c3, 0-0 — in a dozen of your own games until the shape is automatic. Then study the three or four places where the automatic pilot must switch off: the …Qb6 hit on b2, the trade of your f4-bishop, and the moment to commit to Ne5 or to the h4-h5 attack.
For models, look at Kamsky’s London games from the 2010s for the ambitious treatment, and at Carlsen’s for the way a small edge is nursed into an endgame. Then read the companion tutorial, How to Play the London System, which walks through those plans one position at a time. The opening will not teach you tactics or force an advantage. It will teach you a structure — and, like the Italian with its patient bishop on c4, it lasts precisely because understanding a structure outlives memorising a line.
— Editor’s desk, 5 July 2026