Move 13, White to play and win…
My initial instinct was the same as Alex’s, as soon as I saw …f6 I knew this was a blunder and g4 should win, but then comes the grinch who stole Christmas – namely, the analysis to back that up!
I did not like 13.g4 Bc2, 14.BxB NxB, 15.QxNc2 fxNe5, but I never looked deep enough to calculate the in-between move 15.Nb5 (duh, I don’t need to take the Nc2 yet), followed by Q-moves, _then_ QxNc2 threatens Nc7+ fork, so ..Rc3, Qb3..fxNe5 Qxe6+ is ++-.
I looked at virtually anything you can think of on this move, but I couldn’t calculate deep enough in all of the lines. Ultimately, I simply missed his …Qb4 when Black has the advantage.
I looked at all kinds of stuff like NxNc6 followed by Nb5xa7 threatening Qxc6+, getting a Nb5 and Nc4 to hit Nd6+, and even playing g4 only after playing Nb5 (which is incorrect, around +.7 but not winning since the knight should come back to c3).
I played Ne3 quickly to the tune of the clock, sensing that my d4 pawn would drop and I could hold the draw as long as I get rid of the bishop pair advantage (I saw his Bd3, but missed that Rd1 holds the d4 pawn as well). I guess I didn’t want to get tied down into only-move territory but I had 6.5 min to his 2.5 min. I blundered the exchange, spending less than half a minute on the move. I kept playing but resigned once he was down to around 17 seconds. I finished with around 5 minutes. So, I tried to keep some clock pressure on him, but that didn’t work either as I was too tactically rusty.
The thing that sucks about this time-control (put aside that one needs to save time for the ending) is that each of these lines I mentions really needs to be seen like 8+ moves deep, but there are so many to look at in so little time that you end up looking at 5 or 6 variations 4 moves deep instead of looking at all of them deeper. The clock time-limit literally cuts the depth in half.