How punish wing gambit to Sicilian: 1.e4 c5 2. Nf3 Nc6 3. b4!?

@Paradise_Pete I noticed you don't play much here.

I didn't know that e5 is considered better than Sicilian today. I too don't play e5 because too much theory to learn. I like Sicilian because there are a lot of sharp lines but relatively manageable theory. But in truth the opponents I play, as well as myself, are not at the level to have a real classical Sicilian game. The classical game has white castling queenside and pawnstorming black, while black counterattacks on the queenside, typically with a rook sac against the knight on c6. In games I play, I never see that! Maybe 3% of white players castle queenside in the first place. I often play a trap in Sicilian bullet that is really unsound, but few know how to refute it. Bad habits....

@Paradise_Pete

A new game came up. Made the right moves (that you taught me), captured the rook, was better than -3.00 out of the opening, somehow let him have the initiative and blundered into mate...

@nayf Unfortunate :-)
He did a good job of exploiting the weakened dark squares. The opening certainly went well for you. I suggest on move 9 that you play bxc6 instead of dxc6. That way when he plays e5, preventing your bishop from retreating, you can try to undermine it by playing d6. If you can preserve that bishop (or at least exchange it for his bishop) it's much easier to defend.

@Paradise_Pete Thanks again. It's funny you suggest bxc6, as that's what I usually do in Sicilian after Nxc6. Normally one wants to take toward the centre in the opening, all else equal. On this occasion I must have thought that with his king looking rather naked after the exchange, opening the d file for my queen might facilitate a quick attack, while a queenside pawn majority could be useful later. It's bullet so no calculation, just intuition and instantaneous judgments.
Another possibility was to pull back my bishop on 9..., giving up a pawn to preserve it. But the fact is, as late as move 26 I had a -6.30 advantage but blundered. Yeah, he seized control of the dark squares. They're hard to defend without the B, because he could have (and should have) shooed away my N with g4. I'll probably try bxc6 next time!

@nayf You keep wanting to attack :-)
You should be consolidating here, not attacking. You've already got a winning advantage. The only thing that could go wrong is white generating a winning attack, so don't help him.

Yes, you still had a huge advantage late in the game, but not all advantages are equal. That one was based on precise defense and difficult-to-find moves. For the engine that doesn't matter, but a human playing bullet it's hard to do. If instead you've consolidated you could make a lot of smaller mistakes and still be winning. There's be a lot of room for error.

If the computer says a position is, say, -6, but there's only one line that does that, otherwise you're lost, that's not so good. It's better to have a half-dozen reasonable moves that are all -3.

@Paradise_Pete Very true, good insights. Actually, despite appearances in this game, I have a general tendency to consolidate when up a piece, but not always in a good way. I have lost many a game by grabbing a lead with a good tactic, then yielding the initiative and getting mated. I've also won plenty of games after blundering and then going on the attack. It's interesting that being down materially sometimes gives one an ironic advantage of feeling there is nothing to lose by launching an all out attack. Often the material difference doesn't matter if the opponent is relatively undeveloped and can't activate their pieces fast enough in the defense, or has just failed to take initiative out of complacency and never capitalise on their material advantage.
But yes, this game required very precise play to defend adequately. In a longer game I likely would have found the solution, but as you said, in bullet it's not easy. I don't even know whether my opponent KNEW he had checkmate with that last move. Probably but not necessarily.