The Van Geet Opening: Dougherty Gambit

This gambit is about speed and activity over material. You have surrendered two central pawns (the e4 pawn and the f3 pawn, which Black can take immediately), but in return you get open lines for your pieces and a lead in development. The engine's recommended path for Black is 3...exf3 4.Nxf3 g6, preparing to fianchetto the dark-squared bishop. Notice that Black is already thinking about safety, not just counting pawns. Your job as White is to make every tempo count: develop quickly, castle, and create threats before Black can consolidate. If Black panics or plays passively, your active pieces can overwhelm them.

The most-played move by a huge margin — 27,937 games out of 36,768 — is 3...exf3, simply taking the offered pawn. After 4.Nxf3, the engine likes 4...g6, heading toward a kingside fianchetto. White scores 47.2% against this line, which is actually the best Black can do. That is crucial: Black's strongest reply still gives you nearly a 50-50 chance. The game becomes a real fight. You will have the half-open f-file, your knight on f3 is active, and you can follow up with moves like d4, Bc4, and O-O to build pressure. The pawn is gone, but the initiative is real.

Here is where the Dougherty Gambit earns its keep. Many Black players do not know the best response, and the punishment is severe. Three common choices are all losing errors according to the engine: - 3...Nc6 is an inaccuracy, costing about 0.6 pawns. Black develops a piece but ignores the tension in the centre. - 3...e3 is a mistake, losing roughly 1.2 pawns. Blockading with ...e3 seems logical but it is too slow. - 3...Bf5 is also a mistake, costing around 1.4 pawns. Developing the bishop early looks natural but lets you seize the initiative. These three moves together account for over 2,600 games where Black has already slipped. If you play actively after any of them, your winning chances jump significantly.

Across 36,768 games in the Lichess database, the overall results are remarkably balanced: White wins 47.0%, Black wins 49.7%, and draws are rare at just 3.3%. Despite the engine giving Black a clear theoretical edge, practical play tells a different story. Black's winning percentage drops noticeably after some replies: 3...e5 scores only 41.6% for White (meaning Black wins even more), while 3...Nc6 gives White 43.5%, and 3...e3 gives White 44.7%. But against 3...Bf5 — the second-most popular mistake — White scores a healthy 53.0%. That is a winning percentage despite the engine evaluation, a clear sign that Black's position is harder to handle than it looks.

If you enjoy unbalanced, tactical positions where you have the initiative and your opponent must find precise moves, this opening is for you. The typical middlegame after 3...exf3 4.Nxf3 features an open f-file, a pawn centre you can build with d4, and easy development for your bishops (Bc4, Bf4 or Bg5). Black often spends time defending their extra pawn or castling queenside. You are playing for an attack, not a pawn-up endgame. That is the trade-off: you accept a theoretical disadvantage in exchange for an entertaining, aggressive game where your opponent is far more likely to blunder than you are.