Queen's Gambit Declined: Marshall Defense (3.Nc3) — Playing Black

The core idea of 2...Nf6 is aggressive: you attack the c4-pawn immediately and dare White to play 3.cxd5, when after 3...Nxd5 you reach a Scandinavian-style structure with easy development and no space disadvantage. When White avoids that and instead plays 3.Nc3, you respond with 3...e6, entering a Queen's Gambit Declined where White's knight is committed to c3 before developing the kingside. In theory, this matters because White sometimes prefers to keep that knight on b1 for a setup with Nf3 and Bg5, then later bring the knight to d2 or e2. By forcing 3.Nc3, you limit White's flexibility. That said, the statistics show White still scores well here: across nearly 17 million games, White wins 51.4%, draws 4.3%, and Black wins 44.4%. The position is playable for you, but you start with a small gap to close.

By far the most popular response at this point is 4.Bg5, appearing in over 5.4 million games. White pins the knight on f6, threatening to double your pawns or simply increase the pressure on the d5-pawn after an eventual cxd5. White scores 51.5% from this line — in line with the overall average. The engine's top choice is also 4.Bg5, followed by the natural developing moves 4...Be7 5.Nf3 O-O. That sequence reaches a standard Queen's Gambit Declined position where Black has castled and is ready to challenge the centre with ...c5 or ...b6. When you face 4.Bg5, aim to play ...Be7 next, unpinning the knight and preparing to castle quickly. Do not panic and play 4...h6 unless you know the specific lines — the simple ...Be7 is solid and scores well in practice.

White has several other reasonable options you should be ready for. The second-most popular is 4.Nf3 (over 4.6 million games, White scores 52.0%), which is a quieter approach. White simply develops and keeps the tension in the centre. In that case you can continue with the same idea: ...Be7, ...O-O, and prepare to break with ...c5 or ...dxc4. The move 4.cxd5 (2 million games, 51.8% for White) exchanges pawns immediately, leading to 4...exd5 and a symmetrical pawn centre. That position is perfectly fine for Black — you have easy development and no weaknesses. The quiet 4.e3 (1.9 million games, 50.4% for White) and 4.Bf4 (1.1 million games, 51.6% for White) are less testing. Against either, develop naturally with ...Be7, ...O-O, and ...c5 at a suitable moment. Notice that White's scoring drops a bit with the slower moves — that reflects the fact that you have more time to organise your counterplay.

The numbers tell a consistent story. Across White's five most popular moves — Bg5, Nf3, cxd5, e3, and Bf4 — the White win rate hovers between 50.4% and 52.0%. That is a narrow band, which means none of White's options crush you outright. Your task is practical: develop your pieces onto good squares, castle early, and choose a sensible central plan. The engine's +0.32 evaluation is a small edge for White, so you are slightly worse but in no real danger. In the drill below, you will face 4.Bg5 — the engine's best and the most common choice. Try the natural 4...Be7 and see how the game develops. Pay attention to moments where you can break with ...c5, which is Black's main source of counterplay in QGD structures. If White plays something slower like 4.e3, you have good chances to equalise comfortably and even outplay a careless opponent.