Grob Opening: Zilbermints Gambit, Schiller Defense — Black's Guide

The Schiller Defense (3…h5) does two things at once: it stops White from kicking your knight with g5 later, and it opens a line for your rook if White ever captures on h5. You're a pawn up and you've already disrupted White's unusual setup. Your main job now is to complete development without letting White stir up cheap threats. The centre is open — your light-squared bishop will be powerful once it gets out, and the h5-pawn gives you a foothold on the kingside. Don't rush to return the pawn; White's compensation is shaky at best.

Stockfish's top recommendation is 4.g5, which continues with Bg4, then after Be2 you play Qd7. The idea is simple: develop your bishop to g4 pinning whatever knight or pawn White puts on f3, bring your queen to d7 to support the bishop and connect your rooks, and keep your extra pawn. White's pawn on g5 looks active but it's actually a target — once your kingside is solid, you can undermine it later with …f6 or just leave it as a weakness. This is the plan that keeps your advantage cleanest. In the 13 games where White played g5, White scored 69.2% — that sounds scary, but remember those games include both colours and the engine says you're better. Play accurately and you'll outperform that statistic.

The most-played move in practice is 4.gxh5, appearing in 13 games. Here White scores only 46.2%, which is below average for White — a clear sign Black is doing well. After …gxh5 you recapture with the rook (…Rxh5) or sometimes the knight (…Nxh5) — either is fine, though the rook keeps your knight flexible. You're still a pawn up, your rook is active on h5, and White's kingside is compromised. Just be careful not to get checked on the h-file; a timely …Be7 or …g6 can shut that down. This is the line where Black scores best in practice, so if you see gxh5, you should be confident.

The engine flags three white moves as clear errors. Nxe4 (6 games, White scores 33.3%) is an inaccuracy — it loses you about 0.7 pawns of advantage if you respond correctly. Don't rush to take on e4; instead, develop naturally and keep the pressure. Bh3 (2 games, White scores 50.0%) is a mistake costing White ~1.1 pawns; your bishop on h3 is awkward but not dangerous if you ignore it and continue development. d3 (1 game, White scored 0.0%) is another inaccuracy (~0.6 pawns lost). Against any of these, just keep your pawn advantage, finish development (…Nc6, …e6, …Be7), and you'll consolidate. The engine's consistent message: White should play g5, and anything else makes your position even better.