The Trompowsky Attack (1.d4 Nf6 2.Bg5) skips the main-line d4 theory and immediately pressures Black's knight — either forcing a concession or baiting a trade that leaves Black with a doubled pawn. Stockfish rates the starting position +0.15, and the real edge comes from facing opponents who haven't prepared for it.
On this page
Play the Trompowsky Attack against the engine
Free, no signup — you play white, the engine adapts to your level.
You just played the Trompowsky Attack against the engine. Create a free Chessy account for move-by-move AI coaching that explains the right plan in every structure.
Find your own #1 mistake — free →The idea behind 2.Bg5
White's bishop goes to g5 before developing the knight, threatening Bxf6 on the next move. If Black ignores it White gets to double Black's f-pawns; if Black plays 2...Ne4 to kick the bishop, White retreats and exchanges anyway — either way Black's structure or development is slightly disrupted. The Trompowsky is a system you can learn as a set of plans, not a tree of forced variations.
Black's most common replies
- 2...e6 — the most popular (1,127,046 games); Black sidesteps the trade and builds a solid center. White scores 47.9%.
- 2...d5 — the engine's top choice; Black stakes the center immediately. White scores 48.3%.
- 2...g6 — fianchetto setup (621,638 games); Black aims for a King's Indian structure. White scores 47.1%.
- 2...Ne4 — grabs the bishop before it doubles the pawns (478,868 games). White scores 47.2%.
- 2...h6 — asks the bishop where it wants to go (364,805 games). White scores 50.9% — the most favorable reply for White.
How to play it as White
After Bxf6 (wherever Black allows it), push for a central majority: e3, c4, Nf3 and chip away at Black's doubled f-pawns in the endgame. When Black avoids the trade with 2...e6, pivot to a Torre-like setup — Nf3, e3, Bd3, and a slow central build. The consistent thread is avoiding Black's preferred pawn structures: don't let the game become a Nimzo-Indian or King's Indian by transposition without a fight.
What 3.7 million games say
White scores 48.2% across 3,716,726 games — slightly under 50%, yet the engine gives White +0.15. The practical gap reflects that the Trompowsky attracts sharper responses from prepared opponents. The bright spot: Black's least popular main reply, 2...h6, is also its worst — White scores 50.9% there. The opening's real value is in the preparation asymmetry: you've studied it, they probably haven't.
Results across 3,716,726 Lichess games
| Most-played continuation | Games | White wins | |
|---|---|---|---|
| 1 | e6Main line | 1,127,046 | 47.9% |
| 2 | d5 | 652,590 | 48.3% |
| 3 | g6 | 621,638 | 47.1% |
| 4 | Ne4 | 478,868 | 47.2% |
| 5 | h6 | 364,805 | 50.9% |
| 6 | d6 | 187,178 | 49.1% |
Frequently asked questions
Is the Trompowsky Attack a sound opening?
Yes — Stockfish evaluates the position after 1.d4 Nf6 2.Bg5 at +0.15, essentially equal. It's a legitimate alternative to main-line d4 systems favored by GMs including Nigel Short and Julian Hodgson.
What is the best response to the Trompowsky?
2...e6 is the most popular reply (1,127,046 games) and keeps a solid structure. 2...d5 is the engine's first choice. Both hold White to under 49%.
How does the Trompowsky differ from the Torre Attack?
Both open 1.d4 and develop the bishop to g5, but the Torre waits for 1...Nf6 2.Nf3 3.Bg5 — a less forcing move order. The Trompowsky plays 2.Bg5 immediately and often aims to trade the bishop for the knight sooner.
Why does White score below 50% if the eval is positive?
Players who choose 2...e6 or 2...Ne4 tend to know what they're doing — the Trompowsky selects for prepared opponents. Across all 3.7M games White scores 48.2%, but against the weakest reply (2...h6) it reaches 50.9%.