What went wrong here?

11. Nd4 would've won material on the spot. The only two temporarily save moves the queen can make that still guard the fork on c7 are Qb6 and Qc5. Qc5 Nxe6 discovering an attack on the queen, Q moves Nxg7+ If Qb6, same principle but winning one less pawn by Nf5 instead. Also possibility (I'm too tired to calculate) of outright trapping the queen or winning additional material in lieu of it.

I also like 19. Nf5. The knight can not be captured. 19. exf5?? Bxb7!! So, his king can subsequently be deprived of his right to castle, again with that capture on g7 with check. Or, if he defends that, say with Rg8 or Kf8, you've still kept him from castling kingside. If he castles on the spot, 20. Ne7+ Kh8 21. Ne5! dxe5 22. Bxh7! e4 23. Rxe4 or some similar sort of line could give his king a very hard time. Or for that matter, the simple but elegant 20. Nxg7...

Painful as it is, 29. h3 was sadly mandatory, to be followed by Kh1 and Bg1 and f4.

The game was plainly drawn from ... gawd. Pretty much any time both sides have more minor pieces than pawns and no major pieces and even material/within 1 pawn is gonna draw. So at least from move 36 on. His best chance would've been 44. ... a2. Since you can sac the piece for the pawn anyway, he might as well force it. Then maybe he could've corralled or traded off your knight... although, still, with his wrong-colored bishop, it still would've drawn most likely unless you blundered.

Originally posted by paultopia
[b]11. Nd4 would've won material on the spot. The only two temporarily save moves the queen can make that still guard the fork on c7 are Qb6 and Qc5. Qc5 Nxe6 discovering an attack on the queen, Q moves Nxg7+ If Qb6, same principle but winning one less pawn by Nf5 instead. Also possibility (I'm too tired to calculate) of outright trapping the queen or winning additional material in lieu of it.