Skipping 2.c4: Seven Systems White Uses to Dodge Theory, and the Plan Black Needs

At round seven of the national chess championship I was sitting in the third row, behind board four. White, a sixteen-year-old, placed 1.d4 on the board. Black answered 1...d5. The kid played 2.Nf3, skipping 2.c4. His opponent — a player with more than ten years of tournament chess behind him — reached for the clock, pressed it, and then sat still.

Forty minutes later he still had not played move six.

I wrote that timestamp in my notebook next to a short line: "nothing to remember." Black in that game did not collapse because his position was torn apart. He collapsed because the entire body of theory he had prepared — the Slav, the Semi-Slav, the Queen's Gambit Accepted, the Orthodox Queen's Gambit — suddenly went unnamed.

The cluster of openings that the Bulgarian coach IM Valeri Lilov devotes six full videos to, outlining a plan for Black in each, is exactly what showed up in that game and around it: the Colle System, the Trompowsky, the Torre Attack, the London System, plus the Catalan, the Blackmar-Diemer Gambit and the Richter-Veresov Opening.

A heavy body of theory and one way out

1.d4 d5 opens one of the densest theoretical complexes in chess. At club level and above, a serious player may have spent hundreds of hours dealing only with 2.c4 and its branches — from Slav structures with ...c6 and ...Bf5, through the Semi-Slav with ...e6 and ...dxc4, to the Queen's Gambit Accepted with ...dxc4 on move two.

Skipping 2.c4 is a deliberate decision. White declines the theoretical fight, accepts a less ambitious position, and trades it for two things: preparation time and discomfort for the opponent. At national championships, where preparation time is shared across many games and many opponents, the second sometimes outweighs the first.

What makes this cluster worth discussing is that it is nowhere near as harmless as its reputation. The Colle, the Trompowsky, the Torre and the London each carry their own independent theory, and at many levels they deliver better results than systems considered textbook-correct.

Seven systems, seven plans

Colle System. White plays 1.d4 d5 2.Nf3 Nf6 3.e3, then 4.Bd3, 5.c3, 6.Nbd2, 7.O-O, 8.Re1. The idea is to get e4 in with full support. Black has to claim space before White completes the setup: ...c5, ...Nc6, ...Bd6, then either ...b6 with ...Bb7, or ...dxc4 followed by ...b5 to change the queenside pawn structure. If Black just plays ...e6 and waits, White gets e4 for free and the position gets squeezed in silence.

Trompowsky Attack. 1.d4 Nf6 2.Bg5. Black has 2...Ne4 — the main line, accepting a bishop for the centre — or 2...e6, 2...d5. The general plan: do not let the g5 bishop sit on the diagonal for long, and if White trades on f6, Black's pawn structure must be chosen in a direction that favours ...c5.

Torre Attack. 1.d4 Nf6 2.Nf3 e6 3.Bg5. Black has 3...h6 4.Bh4 g5 — driving the bishop away at a structural cost — or 3...c5 4.e3 Qb6, hitting b2 directly. The plan is to take kingside space and pressure the centre before White gets in Bd3 and e4.

London System. 1.d4 d5 2.Nf3 Nf6 3.Bf4. This is the system I used to underrate the most. Black needs ...c5, ...Qb6 aimed at b2, ...Nc6, and usually ...Nh5 to trade off the f4 bishop. Once the f4 bishop is gone, the London loses most of its attacking idea and turns into a dull symmetrical structure — exactly what White does not want.

Catalan. 1.d4 Nf6 2.c4 e6 3.g3 d5 4.Bg2. Unlike the systems above, the Catalan keeps c4, but Black faces the long-term g2 bishop. The practical choice: ...dxc4 and hold the pawn with ...b5 or ...a6, conceding White the centre in exchange for a flexible pawn structure.

Blackmar-Diemer Gambit. 1.d4 d5 2.e4 dxe4 3.Nc3 Nf6 4.f3. Black should not get greedy. 4...exf3 5.Nxf3, then ...e6, ...c5 or ...g6 with ...Bg7. Trying to keep the extra pawn is the mistake; giving it back at the right moment and developing is enough.

Richter-Veresov. 1.d4 Nf6 2.Nc3 d5 3.Bg5. Black has 3...Nbd7, 3...Bf5 or 3...e6. The plan is to accept a slightly cramped structure in return for trading the bishop or gaining a tempo of development.

What all seven have in common is that they delay or drop c4 entirely. In the Queen's Gambit, Black has two natural freeing moves ready-made in ...c5 and ...dxc4. When c4 never appears, those moves lose their target. Black is forced to create tension rather than react to tension White creates — and that is a completely different skill from remembering variations.

A lesson from the stands at Hang Day Stadium in 2026 still holds when you put it on a chessboard: a setup only looks beautiful when the opponent agrees to stand still.

A plan does not replace judgement

Six videos, one system each, one plan each — the packaging is effective as instruction. It also carries a specific risk: the learner absorbs the plan as a fixed formula, when the real value of a plan lies in knowing when it has expired.

Based on my experience watching games at national championships and open tournaments, the most repeated error by Black in this cluster is that players know the plan but apply it to a position that has already changed. You are taught ...Qb6 hitting b2 in the London; you play ...Qb6 when White already has c3 and Nb1-d2, and the move becomes a wasted tempo.

No tactic is ever old; only the way we read the game expires.

In the summer of 2026 I wrote that the London System was the opening of the lazy. I take that back. The result tables at the open tournaments I follow show White's win rate in this cluster is not significantly lower than in the main-line Queen's Gambit at the same Elo level. A system dismissed as boring that wins consistently is not boring.

A repertoire chart can lie, but five consecutive games squeezed by move twenty cannot.

Data limits

The source for this piece is the description of a six-part video course by IM Valeri Lilov, in which each video handles one system and proposes a plan for Black. This is course copy; it comes with no dataset attached. Claims such as "extremely dangerous" are qualitative, with no sample, no game count and no Elo band attached. I keep them qualitative and do not give them the weight of data.

One remaining limit: six videos for seven systems means some systems were merged. That makes sense inside a course, but it does not make sense as a preparation map for a specific tournament.

What I will check at the next championship

I will log every game that opens 1.d4 d5 with White skipping 2.c4 in the first round of the next national championship, along with three variables: minutes Black spends in the first ten moves, how often Black plays ...c5 on his own initiative, and the final result. If thinking time goes up while the win rate stays flat, the problem lies somewhere else, not in the opening.

And one question I have not answered: if the plan is what decides, why do we still teach openings as sequences of moves to be memorised?