Black launches a pawn storm before castling in the Italian

I don't like it when White plays 4.d3 against the two knights defense in the Italian. Too positional for my taste, so I felt like being aggressive in the king side before even castling.

Turns out 5...g5 has never been played, yet the computer says it's only +0.4 for White.

PS. My bad, it was played and it's +0.9. It's 6.c3? that was never played as it loses advantage and goes to +0.4

It has been known for hundred years that after h3 and h6 whoever castles first will be attacked by the enemy g-pawn and rook along the g-file. It counts even as blunder, i.e. castling when weakened. Tarrasch‘s „Angriffmarke,“ a token inviting an attack. Without this token it is probably not that dangerous, but who knows...

Botvnik once said that only an idiot is analyzing blitz.You are analyzing bullet

@grondilu it's not true that this has never been played - in fact it was played recently by Ian Nepomniatchi in a classical tournament (I can't remember which, but it may have been the Sinquefield Cup). I can't find this game however, despite the fact that it was broadcast on Lichess.

EDIT: @tpr also found a game, and there appears to be a third one which isn't in the Lichess database.

@newagear #4 It took Botvinnik years to analyze certain positions. But now the analysis takes half a minute. So do not spare the time spent analyzing the bullet.

But unlike you , Botvinnik understood what he was analyzing.

@newagear #7
I don't understand you? If I came with my computer in the 40-50s, he, Botvinnik, would ask me how to play this or that position, and not me)))