How to Play the French Winawer

ECO C15 1,641,764 games Stockfish +0.62

The French Winawer (1.e4 e6 2.d4 d5 3.Nc3 Bb4) is Black's most combative French reply — pinning the c3 knight and threatening to double White's pawns with ...Bxc3+. Play it against the engine below, then see what 1.6 million Lichess games reveal.

Play the French Winawer against the engine

Free, no signup — you play black, the engine adapts to your level.

You just played the French Winawer as Black against the engine. Create a free Chessy account for move-by-move AI coaching that explains the right plan in every pawn structure.

Create a free account →

The idea behind 3...Bb4

By pinning White's c3 knight, Black immediately creates structural tension. The core deal: Black will give up the dark-squared bishop for a knight via ...Bxc3+, saddling White with doubled c-pawns. In return White keeps the bishop pair and a central space advantage. Stockfish rates the position +0.62 — the largest White edge of any opening in this series — but the concrete imbalances (doubled pawns vs. bishop pair, cramped Black vs. dynamic counterplay with ...c5) make this a genuine fight, not a concession.

White's main tries after 3...Bb4

The Advance (e5) is White's critical response:
- e5 (518,053 games, White 51.6%) — the main line and engine's choice; stakes central space and leads to the Winawer mainline where Black plays ...Ne7, ...a6, ...Bxc3+.
- Bd3 (195,557 games, White 50.8%) — solid but Stockfish marks it an inaccuracy (−0.50 vs. e5).
- Ne2 (62,580 games, White 51.3%) — sidesteps the pin; avoids doubled pawns but concedes the Winawer's full tension.
- exd5 (324,184 games, White 48.5%) — exchanges and gets an easy position for Black.
- Bd2 (275,471 games, White 47.6%) — Stockfish flags as inaccuracy; Black scores well.
- a3 (104,074 games, White 47.7%) — forces the bishop decision immediately; still below e5 in scoring.

How to play it as Black

Your blueprint: let White push e5, then play ...Ne7–d7–b6 (or ...f6 to break the chain), exchange on c3 when a3 is played, and open the queenside with ...c5 and ...Qc7. The doubled c-pawns on c3/c4 are targets for the endgame; your main middlegame weapon is queenside pressure backed by a half-open c-file. Against Bd2 or a3 (both under 48% for White), you're already on the right side of the odds. Against e5 (51.6%), precision matters: play ...Ne7 not ...Nf6, and time ...Bxc3+ carefully after White commits to a3.

What 1.6 million games say

Across 1,641,764 Lichess games White scores 49.4%, Black 46.8% — a real White edge but not a clean advantage. Bd2 (White 47.6%) and a3 (47.7%) are the two tries where Black's practical results flip the engine's assessment. The Advance e5 at 51.6% over 518,053 games confirms it's the move Black must prepare hardest for; the pv line e5 Ne7 a3 Bxc3+ is the critical sequence to know.

Results across 1,641,764 Lichess games

49.4%
3.8%
46.8%
■ White 49.4% ■ Draw 3.8% ■ Black 46.8%
Most-played continuationGamesWhite wins
e5518,05351.6%
exd5324,18448.5%
Bd2275,47147.6%
Bd3195,55750.8%
a3104,07447.7%
Ne262,58051.3%

Frequently asked questions

Why does Black give up the bishop for a knight in the Winawer?

The trade saddles White with doubled c-pawns (c3+c4 or c2+c3) that are long-term liabilities. Black gets a concrete structural target in return for the bishop pair deficit.

How should Black handle White's 4.e5 Advance?

Play ...Ne7 (not ...Nf6 which is blocked by e5), then ...a6 and ...Bxc3+ when White plays a3 or b4. Follow with ...c5 and ...Qc7 to attack the e5 chain from the queenside.

Is the Winawer harder to play than other French variations?

Yes — it's the sharpest French, with concrete plans on both sides. But the ideas are forcing: once you know ...Ne7, ...Bxc3+, and ...c5, the structure plays itself. White's trickiest try is e5 at 51.6% over 518,053 games.

What if White avoids the doubled pawns with 4.Ne2?

Ne2 sidesteps the Winawer's main tension. Black still has a fine position: the bishop on b4 remains active, and White scores only 51.3% — below the Advance's 51.6% — so it's not a refutation.