Chessy
QGD: Exchange Variation (exd5) — How to Play It as White
After 1.d4 d5 2.c4 e6 3.Nc3 Nf6, the Exchange Variation with 4.cxd5 exd5 5.Bg5 gives you a quiet but lasting advantage. Stockfish evaluates this position at +0.28, a small edge for White, meaning you are slightly better. With over 1.26 million games in the Lichess database, White scores 51.5% wins – a solid practical result. The key now is knowing which Black replies are harmless, which one is a real mistake, and how to follow up without overpressing. This page walks you through the data and the best plan.
Play the QGD: Exchange Variation: exd5 against the engine
Free, no signup — you play white, the engine adapts to your level.
Practise the QGD Exchange Variation in the interactive drill above. Focus on the correct response to 5...h6 and the natural development after 5...Be7.
Create a free account →What You're Fighting For
The Exchange pawn structure (d4 and d5 facing each other) is famous for its slow, manoeuvring character. Black's light-squared bishop is blocked behind its own pawns, while your plan is straightforward: develop naturally with e3, Bd3, Nf3, short castle, and later look for kingside pressure or a minority attack on the queenside. You have no immediate tactical knockout, but the position is comfortable and risk-free. The statistics back this up — across the six most-played Black replies, your winning percentage never drops below 48.9%, and against several replies it climbs above 53%.
The Engine's Favourite and the Most-Played Replies
Black's best move is 5...Be7, played in 580,136 games (White scores 51.4%). The engine recommends the natural follow-up 6.e3, then 7.O-O, 8.Bd3. You should be happy to see this — it's a classical developing sequence with no tricks. The other common replies are 5...c6 (204,729 games, White 48.9%), 5...Bb4 (193,470 games, White 53.0%), 5...Be6 (86,392 games, White 51.6%), and 5...Nbd7 (65,308 games, White 49.0%). Only one of these moves is a clear mistake.
The One Mistake to Punish
The move 5...h6 has been played 54,205 times, but it is a clear inaccuracy. It loses roughly 0.8 pawns in evaluation; the engine says Be7 was much better. White scores a massive 55.5% after h6, well above the overall average. Why is h6 bad? It wastes time and weakens Black's kingside without solving any problem. Your bishop on g5 is not really under threat — Black can't win a piece because after 5...h6? 6.Bxf6 Qxf6 you have simply traded your bishop for a knight, and you will follow up with e3, Bd3, Nf3, and enjoy a lead in development and a safer king. In the interactive drill, practise the correct response to h6.
Results across 1,264,704 Lichess games
| Most-played continuation | Games | White wins |
|---|---|---|
| Be7 | 580,136 | 51.4% |
| c6 | 204,729 | 48.9% |
| Bb4 | 193,470 | 53.0% |
| Be6 | 86,392 | 51.6% |
| Nbd7 | 65,308 | 49.0% |
| h6 | 54,205 | 55.5% |
Frequently asked questions
Is the QGD Exchange Variation good for White?
It gives a small but stable edge. Stockfish evaluates +0.28, and in practice White wins 51.5% of games. It's a solid choice if you want a risk-free position with long-term pressure.
What is Black's best move after 5.Bg5?
The engine recommends 5...Be7, followed by natural development with e3, O-O, Bd3. It's the most popular reply (played over 580,000 times) and leads to a standard positional game.
Why is 5...h6 a mistake?
5...h6 loses about 0.8 pawns in evaluation. White can simply play 6.Bxf6 Qxf6, trading bishop for knight, and follow up with e3, Bd3, Nf3. White gains a lead in development and a safer king, leading to a 55.5% win rate.
How should I follow up after the most common replies?
After 5...Be7, play e3, then O-O and Bd3. Against 5...Bb4 (White scores 53.0%), standard development is still good. Against 5...c6 or 5...Nbd7, continue with e3, Nf3, Bd3, and castle. The plans are flexible but straightforward.