Move 34 discussion: Everyone vs stockfish

[Brian-E]

I'm a little concerned now about 34.Rf1 Rb2. Black may be able to hold the b pawn and later win our a pawn which gives it winning prospects.

If that is the case, I think 34.Rxb6 might be advisable. As mentioned before, I think endgames where Black has h and f pawns and we have a g pawn are drawn in general.

Can anyone convince me otherwise? I am holding off my vote for the moment.

[WMHalsdorf]

Originally Posted by Brian-E

I'm a little concerned now about 34.Rf1 Rb2. Black may be able to hold the b pawn and later win our a pawn which gives it winning prospects.

If that is the case, I think 34.Rxb6 might be advisable. As mentioned before, I think endgames where Black has h and f pawns and we have a g pawn are drawn in general.

Can anyone convince me otherwise? I am holding off my vote for the moment.

35.a4 h5 36.Ng7 Ra4 37.Nxh5 Raa2 38. Rb8+ Kh7 39.Nxf6+ Kg7 40.Nh5+ Kg6 41.Nf4+

There may be some improvement for us in this line. Basically a drawn position.

[gd_barnes]

Originally Posted by WMHalsdorf

35.a4 h5 36.Ng7 Ra4 37.Nxh5 Raa2 38. Rb8+ Kh7 39.Nxf6+ Kg7 40.Nh5+ Kg6 41.Nf4+

There may be some improvement for us in this line. Basically a drawn position.

This is a good line for us. Nice find with 36. Ng7! It's a strange looking move but a strong one because black cannot protect the h-pawn.

We are a pawn up and there appears to be no improvement for SF here. We have two connected king side pawns vs. black's isolated queen side pawn. The position is still likely drawn but I see no reason why SF would choose a line where he loses a pawn when he can choose many others and have even material. For that reason, I do not fear 34...Rb2.

[Brian-E]

But after 34.Rf1 Rb2 35.a4 h5 36.Ng7 Black has 36...h4 37.gxh4 (otherwise ...h3) Rxa4. Now not 38.h5? or 38.Rxf6? both answered by 38...Rg4+. If 38.Nf5 we have a position with doubled h pawn against b and f pawns which may well be winning for Black.

[gd_barnes]

Originally Posted by Brian-E

But after 34.Rf1 Rb2 35.a4 h5 36.Ng7 Black has 36...h4 37.gxh4 (otherwise ...h3) Rxa4. Now not 38.h5? or 38.Rxf6? both answered by 38...Rg4+. If 38.Nf5 we have a position with doubled h pawn against b and f pawns which may well be winning for Black.

I had considered the advance of 36...h4 and it would be a mistake by black. Although likely not a fatal one he must play exactly to obtain the draw. I will demonstrate below. The reason is that the response 37. gxh4?? by us is a huge mistake and loses big material immediately, a much worse situation than just having doubled pawns. Instead 37. Nh5! and we have the advantage. Black's only response must be 37...hxg3. Almost any other move allows 38. Nxf6 with a big advantage to white and possibly a mating attack in some variations.

To demonstrate, the best line might go:
36. Ng7 h4?!
37. Nh5! hxg3 (forced)
38. hxg3 Ng4 (forced; f6 must be protected to prevent an overwhelming attack by us)
39. Nxf6 Nxf6
40. Rxf6
We are up a pawn and all over black's king. He must be careful.

Also note that black is losing and we have a possible mating attack after:
37. Nh5 h3??
38. Nxf6

Here is why we lose almost immediately with 37. gxh4:
36...h4
37. gxh4?? Rg8!

We drop the knight and it's curtains from there.

[Brian-E]

Thank you Gary. I agree with all that. The counter-attacking manoeuvre Nf5-g7-h5 threatening Nxf6, as originally posted by WMH and endorsed and expanded by you, is indeed nifty.

I'm a bit torn really. To me 34.Rxb6 is a clear strategic draw, even if we lose both our h pawn and our a pawn, provided that Black has no unforseen tactical tricks. But I'm also keen to know what SF has worked out after 34.Rf1. If we had a human opponent, I would want to play 34.Rf1 anyway because it seems to give winning chances as you demonstrate after the substandard 36...h4?! But such an outcome is well nigh impossible against Stockfish. We must play for the draw.

I'll shut my trap now, hold off for a while to see if anyone else has anything else to add, and vote before the poll's deadline.