Scandinavian Defense: Gubinsky-Melts Defense with 4.Nf3

Your queen on d6 looks awkward, but it's doing important work. From d6, the queen supports ...c6 and ...Bf5, controls key dark squares, and stays safer than it would be on a5. Your knight on f6 attacks e4 and prepares to castle kingside. The central tension is real: White's knight on c3 and pawn on... well, there's no pawn yet — White hasn't played d4. That's the whole story of this position. White wants to push d4, seize space, and cramp you. You want to develop solidly with ...c6 and ...Bf5, creating a compact pawn structure that's hard to break down. If White plays accurately (the engine's suggested line is d4 c6 h3 Bf5), you get the bishop to f5 and a sturdy centre. Your job is to keep the position closed, avoid tactics against your queen, and reach a middlegame where your coordinated minor pieces compensate for White's space advantage.

Stockfish's top pick for White is 5.d4, and across over 218,000 games it scores 49.0% for White — almost exactly average. After 5.d4, you should play 5...c6. This does two things: it gives your queen a retreat square on c7 if needed, and it prepares ...Bf5, developing your light-squared bishop to a nice diagonal before White can play h3. The engine continues with 6.h3 Bf5, and White's h3 is a useful little move that prevents ...Bg4 pinning the knight. Your bishop sits happily on f5, where it pressures White's queenside and frees your kingside for castling. From here you have a solid, if slightly passive, position. Black scores 47.7% from the starting position overall, and after 5.d4 the numbers are similar — the game is far from decided.

White has two common inaccuracies that improve your chances significantly. First is 5.d3, played in over 75,000 games. This loses roughly 0.7 pawns of advantage compared to the best move d4. The problem? White gives up the fight for the centre, and your queen on d6 suddenly looks more comfortable. Black scores 46.4% after d3 — not a huge jump for you, but the engine says White is making your life easier. The second mistake is 5.Bb5+, played over 37,000 times, which loses about 0.5 pawns. White checks your king, but after you block with ...Bd7 or ...Nbd7, White's bishop is misplaced and you're developing with tempo. White's score drops to 45.8% here — your best statistical outcome from the main position. If you see either of these moves, you know the engine disapproves, and you can play with extra confidence.

The most dangerous alternative to d4 is 5.Bc4, played nearly 133,000 times. White's score drops to 46.6% here — good news for you. Your queen on d6 already eyes the a3-f8 diagonal, and White's bishop on c4 doesn't threaten much. Develop naturally with ...c6 and ...Bf5, and White's bishop can become a target later. The sharpest try is 5.Nb5, played over 30,000 games, where White's score jumps to 53.1% — your worst result statistically. This attacks your queen directly, and you need to be precise. Retreat to ...Qd8 is safest, forcing White to lose time moving the knight again. Don't panic — 30,000 games is the least common of the major replies, so you won't face it often. Against all options, your formula is the same: secure the centre with ...c6, develop ...Bf5, castle kingside, and let your solid structure absorb White's early initiative.