BxH6 was a blunder, but it won me the game?

The game-swinging move was Bxh6 by me which undoubtedly won me the game, but after seeing the stockfish analysis, I was shocked to see that it was a blunder. If you have any suggestions for white/black during the course of the game, do leave it in the comments.

The reason I was shocked is because my idea after Bxh6 was 14. gxh6 15. Nxh4 (in an attempt to remove the defender of the f6 square), and then when black responds with some move, look to play Nf6+, forking the queen.

Unfortunately, black attempted to play f5, trying to defend f6 by letting the black rook on f8 watch the square so that the knight could not jump in, but missed the double check, which guarantees the queen capture to follow.

It was easy to see that white has an easy conversion after black resigns, but I can't seem to find any significant moves that either side could have played/avoided to swing the game in their favour.

It shouldnt have though.

After you took the bishop with your knight he should of done...well anything other than what he did.

Your knight check shouldnt of taken queen; or anything, he had your knight with his castle.

Castle takes knight; then your queen is still in jeopardy; castle Queen H7 after you move your queen out of the way; puts king in check jeopardy. King moves G1 its checkmate so King moves G2 Queen goes H5 knights in trouble. Checkmate soon after.
One way it could of gone anyway.

@breakreign The reason black's rook on f8 could not capture my knight checking the king on move 16. Nf6+, is because black missed the double check (by both the knight (f6) and bishop (c4)) which requires the black king to move.

This, in turn, guarantees my capture of the black queen, Nxd7

Hah! I too missed the double check.

Well played. Lets see if someone else has a answer for you then :)

Okay what about if instead of moving his pawn forward he instead moved his Queen back to D8.

This covers your knights check; and also puts your H4 Knight in jeopardy.

@breakreign

After Qd8, his allows for white to play Qxg4+ , landing a check on the king while simultaneously defending the knight that the black queen is targeting...

Also, if black continues play with Qg5, then I'd gladly trade as white as I end up with a 5 point material advantage which I feel quite confident about in converting for the win. Also, the black king looks quite vulnerable in the given position after the queen trades.

If you take a look at the sample continuation that stockfish is offering you can see it's a very sharp line after Bxh6. There is in fact only one move black can play that will capitalize on your "mistake", which is Nd4. After Nd4 your knight on f3 is pinned to your queen, and the computer claims that the best move is in fact Nxd4, taking the knight and giving up your queen. Anything else and after black plays Bxf3 it is, at best, a gigantic advantage of about an entire queen and a pawn, and at worst it is a forced mate in 5.

The biggest mistake black plays is taking the bishop, allowing the queen/king fork and the double check. But in fact there is no other move that will allow black to capitalize aside from Nd4. It is razor sharp and the move is only techncially a blunder because there is a refutation- in human play, especially at amateur level, it is a complication that is hard to refute. remember that stockfish knows how to refute any human move that has any kind of weakness, but just because a computer can refute a move doesn't make it a bad one.

@no_seigen That is quite the insightful response, I have to thank you for your analysis on my game!

It's quite interesting how after Nd5, white's best response is actually giving it's queen to black. The sequence of moves that follow after white plays Nxd5 is really quite fascinating. I agree that if white doesn't play Nxd5 then it is quite the huge advantage given to black.

I really do love your point about how computer engines know how to play against any sort of moves but playing human vs. human games allow us to get away with these sort of moves.

I really do love that point as it does make me think that Bxh6 could still be a good move as many human players around my rating would never see the only countermove Nd4, especially under time pressure. I'll keep that in mind!

Nd4 really did not look obvious (at least to me), as it seems like black is only winning a knight after double attacking it while it's in a pin.

Again, thanks for your sharp analysis, I learned a lot.

Black did not give 14...gxh6?? any thought and captured instantly. If he had thought, he might have found that 16 Nf6+ wins for white. He then might have looked for other moves than 14...gxh6 and found the centralising 14...Nd4! which is often played in that kind of position.
Both players played the opening way too fast. 5 Ng5? and 10 g4? result from lack of thought. This was a 10+0 game. White used only 3 min and black only 6.

@tpr Thanks for the insight. I definitely agree, in hindsight, that 5. Ng5 and 10. g4 were inaccurate moves. I should definitely spend more time on the opening as white.

I also agree that if black spent more time in thinking about gxh6 not being a good move, then I do believe that he would have perhaps seen Nd4.

My guess was that black did not see the sequence of moves that allowed the fork on f6, and decided to capture on h6 with the pawn, seeing no threat.

Again, thanks for the advice on spending more time in the opening to avoid any sort of inaccuracies like 5. Ng5?, it can improve my game a lot.