Chessy
The King's Gambit Accepted: Tartakower Gambit — A Subtle Surprise
The King's Gambit is one of chess's most romantic openings, and the Tartakower Gambit (3.Be2) is its sneaky cousin. Instead of the typical 3.Nf3 or 3.Bc4, you develop a piece and tempt Black into overreaching. The resulting position — after 1.e4 e5 2.f4 exf4 3.Be2 — may look quiet, but it hides real bite. In the 7,379 games played from here in the Lichess database, White actually scores a respectable 51.3% despite the engine giving a small edge to Black. That gap between raw results and the computer evaluation tells you something: amateurs mishandle this position. Your job is to learn how Black stumbles — and how you punish it.
Play the King's Gambit Accepted: Tartakower Gambit against the engine
Free, no signup — you play white, the engine adapts to your level.
Ready to test your Tartakower Gambit skills? Jump into the interactive drill and face the engine from the position after 3.Be2. Create a free account to track您的
Create a free account →What Are You Playing For?
With 3.Be2 you decline to recapture the f4-pawn immediately. Instead, you make a useful developing move that keeps Black guessing. The main idea is to meet ...d5 (likely the engine's top move) with 4.exd5, then handle the check 4...Qh4+ by tucking your king with 5.Kf1. That might look scary, but your opponent has spent a tempo moving the queen early, and your king is surprisingly safe. Your long-term assets are a half-open f-file, a pawn on d5 that cramps Black's pieces, and the fact that Black's extra f4-pawn often becomes a liability rather than an asset. You're fighting to prove that the gambit pawn isn't worth the discomfort Black suffers.
How the Engine Answers You
Stockfish's best reply when you play 3.Be2 is 3...d5, which aims to strike in the centre immediately. The engine line runs 4.exd5 Qh4+ 5.Kf1. This is the critical test: Black gives check with the queen, and you walk your king forward. After 5.Kf1, Black has some pressure but no immediate knockout. The engine evaluates this at -0.78, a clear edge for Black. That means you are clearly worse here according to the computer. But that evaluation assumes perfect play from both sides — and the statistics show that in real games, White wins more often than Black. The engine's line shows you the toughest road; everything else gives you better practical chances.
The Statistics: Where Black Goes Wrong
The 7,379-game Lichess database reveals which replies cause Black the most trouble. Black's most popular move is 3...Qh4+ (1,383 games), but White scores only 49.0% against it — it's solid for Black. The second-most popular is 3...Nc6 (1,025 games, White scores 51.0%). Here are the moves where White does best, from best to worst for White's winning chances: - 3...d6 (870 games, White scores 55.6%) — but the engine calls this an inaccuracy that loses about 0.7 pawns compared to the best move. So if your opponent plays 3...d6, they've already slipped. - 3...g5 (770 games, White scores 54.7%) — Black tries to hold the gambit pawn, but 3.Be2 is well placed to undermine it. - 3...Nf6 (649 games, White scores 54.4%) — natural development, but Black's kingside can come under pressure. The worst statistical move for White is 3...Be7 (668 games, White scores just 48.2%). So if you see 3...Be7, you know you're in for a harder fight.
The Mistake to Watch For
The FACTS identify 3...d6 as a concrete inaccuracy, losing roughly 0.7 pawns of advantage compared to the better 3...Qh4+. That's a meaningful slip at this level. Why is 3...d6 a mistake? It's too passive — Black fortifies the e5-square and prepares ...g5 to hold the pawn, but they lose the chance to strike back immediately with ...d5 or to inconvenience you with ...Qh4+. After 3...d6, you can continue with natural development (like 4.Nf3 or 4.d4) and build a strong centre while Black's extra pawn on f4 becomes increasingly isolated. If your opponent plays 3...d6, you have every reason to be confident.
Results across 7,379 Lichess games
| Most-played continuation | Games | White wins |
|---|---|---|
| Qh4+ | 1,383 | 49.0% |
| Nc6 | 1,025 | 51.0% |
| d6 | 870 | 55.6% |
| g5 | 770 | 54.7% |
| Be7 | 668 | 48.2% |
| Nf6 | 649 | 54.4% |
Frequently asked questions
Is the Tartakower Gambit a good opening for beginners?
Yes — it's an excellent practical weapon. The statistics show White scores 51.3% in real games, which is very playable. You avoid heavily analysed main lines, and your opponents often mishandle the position. Just be ready to walk your king to f1 after ...Qh4+; it's safer than it looks.
Why does White play 3.Be2 instead of 3.Nf3?
The Tartakower Gambit (3.Be2) is a flexible developing move. It prepares to castle quickly, avoids the standard King's Gambit lines, and can catch Black off guard. It also keeps the option of meeting ...g5 with d4 and h4 to undermine Black's pawn chain.
What is the engine's best line after 3.Be2?
Stockfish recommends 3...d5 4.exd5 Qh4+ 5.Kf1 as the critical continuation. The engine gives -0.78, meaning Black is clearly better with best play. But in practice, few opponents know how to press that advantage, and White's position is fighting and sound.
Is 3...d6 a bad move for Black?
Yes, it's classified as an inaccuracy. According to the engine, 3...d6 loses about 0.7 pawns of advantage compared to the better move 3...Qh4+. It's too passive — Black should strike in the centre or check immediately rather than quietly defending.