French Defense: King's Indian Attack with ...b6 — How to Play as White

The move 3.d4 seizes space in the centre before Black's light-squared bishop has a chance to become active on b7. By advancing the d-pawn immediately, you stake a claim to d5 and e5, making it harder for Black to challenge your centre with ...d5 later. Black's second move ...b6 was an attempt to fianchetto without contesting the centre first — and 3.d4 exploits that slight inaccuracy. From here, the engine recommends Bb7 for Black (the best reply), after which White continues naturally with Bd3, then meets ...c5 with c3, building a solid pawn centre and preparing kingside development. The position remains strategically rich: you have more central space, easier development, and clear plans for a kingside attack.

Over 8,631 Lichess games from this exact position, the results are striking: White wins 42.6%, draws 4.0%, and Black wins 53.5%. At first glance those winning percentages favour Black, but that's partly because of how the KIA setup is played — many White players mishandle the position. The engine's +0.48 evaluation tells a clearer story: you are slightly better when you know what you're doing. The statistics also reveal which Black replies are genuinely dangerous. The most popular move Bb7 (7,200 games) gives White a 42.1% score, right around the average. But when Black plays c5 (136 games), White's winning percentage jumps to 51.5% — a sign that Black can easily go wrong. The lesson is clear: this is a position where sound play rewards White, and Black's alternatives often backfire.

The engine identifies several Black moves that worsen their position significantly. Here are the three you should watch for: Qh4 is classified as a mistake, losing about 1.0 pawns of advantage. Black brings the queen out early and invites you to gain time with tempo-gaining moves like Nf3. The best move was Bb7. g6 is rated an inaccuracy (losing ~0.6 pawns). Black tries for a King's Indian setup but falls behind in development. Seize the centre and develop quickly. Ba6 is another inaccuracy (losing ~0.8 pawns). Black tries to trade off their bad light-squared bishop, but this costs time you can exploit. In all three cases, the engine says Black should have played Bb7 instead. If Black avoids these errors and plays the principled Bb7, you follow the engine's recommended line: Bb7 Bd3 c5 c3, reaching a flexible, slightly better middlegame where your space advantage and smoother development do the work.

Once you reach the position after Bb7 Bd3 c5 c3, you have a clear long-term plan. Your pawn centre on d4/e4 is solid, your light-squared bishop points menacingly at Black's kingside, and your knight on g1 can develop to f3 or e2 depending on how the game unfolds. Common ideas include castling kingside, pushing e5 when ready to cramp Black further, or preparing a kingside attack with moves like Qe2, Nbd2, and possibly g4 if Black commits to queenside play. Meanwhile, Black's ...b6 setup has left the c8-bishop already committed to b7, and Black may struggle to find a good square for the b8-knight. Your goal is to maintain the tension while improving your pieces — the engine's +0.48 edge may be small, but it's durable and easy to play with good positional instincts.