Indian Defense: Anti-Grünfeld, Alekhine Variation, Leko Gambit — How to Play as Black

With 3...e5 you are gambling on dynamic play. White's move 3.f3 is a committal way to support the centre — it weakens the kingside dark squares and takes away the f3-square from the knight. By hitting d4 immediately you are asking White to make a decision: capture or push? The engine's best move is dxe5 (played in 23 of 37 games in the database), which leads to dxe5 Nh5 Nh3 Nc6. That position is still rated +0.80 in White's favour, but the play is unbalanced and sharp. Your knight on h5 is temporarily offside, but the pawn on e5 gives you a space advantage and your light-squared bishop will be powerful on the long diagonal once it develops.

White has several options here, and their quality varies drastically. The move d5 has been played 8 times but is an inaccuracy, losing about 0.8 pawns compared to the best line. White scores 62.5% from d5, but that number is misleading — it's a positional mistake that gives you a clear target. Even worse for White is e4 (3 games, loses ~0.6 pawns) and e3 (2 games, a full mistake losing ~1.5 pawns). The only move that keeps the advantage is dxe5. If White plays anything else, you immediately get counterplay. Notice that Bg5 (1 game, 0% White score) was a disaster for White — a reminder that premature pinning moves backfire when Black has ...e5 in the air.

Across 37 games at this position, the results are remarkably balanced: White wins 48.6%, Black wins 48.6%, and only 2.7% of games end in draws. Despite the engine giving White a clear edge (+0.80), the practical results are even. This tells you two things. First, the position is complex and offers Black plenty of winning chances even if the engine prefers White. Second, White's advantage is difficult to convert at the club level. The draw rate is tiny, meaning most games end decisively — so play aggressively, trust your compensation for the pawn, and look for chances to generate attacks against White's slightly exposed king position.

The database identifies three suboptimal White moves, all of which you should be ready to punish. d5 is the most common mistake and loses about 0.8 pawns in evaluation. With the centre closed, your plan is simple: develop quickly, target the d5-pawn with ...c6, and activate your dark-squared bishop on the long diagonal. e4 loses ~0.6 pawns and is another inaccuracy — White wastes a tempo and weakens the d4-square. After e3 (a full mistake, ~1.5 pawns lost), White's dark-squared bishop is blocked and the centre is fragile. In all these cases, your task is to develop with tempo, complete kingside castling, and exploit the weakened central squares. The drill will train you against each of these responses.