Oh Markovich, gg, I think your only real error was 32..Re4 that apparently stopped a4 at first sight (but not really, like the game showed) and something like 32..f6 or 32..f5 were better, but I would have played a4 anyway, but not that powerful here. But I am really really surprised of your premature resigning!! Yes 37..f5 should lose according to my analysis , but even here it's premature to resign. Here is a little possible continuation after 37..f5 that could have explained the resigning : 38.Nc5+ Kc8 39.Qa6+ Kc7 40.Rc1 Rg4 41.Qb7+ Kd8 42.h3! Rf4 43.Qb6+ Ke7 44.Nd3 +- But really, after 37.Rf1 and after my analysis, I can objectively say that you have drawing chances! Yes I like my position, I have nice winning chances, but that's not so desesperate for you. Here is my point : 37..Nf4! the knight come back into the game 20 moves later

38.Nc5+ Kc8 39.Kc2 Qb8! (39..Qe7 40.Re1 Re4 41.Qa8+ Kc7 42.Qb7+ Kd6 43.Nxe4+ dxe4 44.Qb8+ Qc7 45.Qxc7+ Kxc7 46.Rf1 Ne6 47.Rxf7+ Kd6 48.b4 Nd4+ 49.Kd2 Nf5 50.Kc3 etc, and it should not be easy to win because I need to end up with R+p against N to win, that's not an easy task but it doesn't seem impossible.) 40.Re1 Re4 41.Nxe4 dxe4 42.Rxe4 Ne6 43.Qd2 c5 44.Kb2 Qb7 (this line is maybe Black's best chance to draw and I am not really sure that I can convert my advantage into a decisive advantage)
I am interested of your toughts about 37..Nf4! and the lines I gived, to see if you agree with the moves and assesment.
About the opening : 9..Nh5 really deserve respects or even a (!) (other moves than 9..Nh5 is somewhat better
for White) But we have seen one of the drawback of this move : the Knight can stay here for a long time like in this game where he didn't moved one single time after move 9. And my response : 10.Be3 is mostly a positional and solid move, if you compare with my other alternative 10.Nxg6!? hxf6 11.Bd3. I agree with you that you reached equality in the opening, but even if you seemed to control the defense with "easy moves", White's compensation really existed and it's a danger to underestimate that. From a theorical point of view, other players played the same position a few time until move 15, where White always played 15.Rdf1, so my 15.Rhf1 is a novelty, and maybe an improvement (White shouldn't be too affraid to lose the h pawn at that moment, and keep the 2 rooks active)
In conclusion, 9..Nh5 is good, 10.Be3 is solid, positionnal and the position is roughly even, but not very pleasant to play against Black's fortress. Black have a pawn up, a position solid as a rock, but we can't win a game just on staying too solid, they need to liberate their position to show that the pawn up is an advantage, but they liberate White's position also which have dynamic compensations. In the whole game, both side maintened an equality (or a very little advantage for Black.) The position opened, White's dynamical compensation really began to play at this moment(like a delayed compensation, that show that the BDG have somes strategics compensations, mostly the f-file semi-open) but White had just an equality (or very close to..), with nothing more. If Black had not blundered with 32..Re4?, the game would surely have been a draw (32..f6 or 32..f5 were good). So I believe that after 9..Nh5 and 10.Be3 White can draw, but nothing more against the best play, and Black can hope a bit more

but I am not too happy with this idea, even if a draw with the BDG against a strong opponent is not that bad.