Benoni

I LIKE to think this blog is nothing if not eclectic (catholic used to be a widely-understood synonym, but not these days), and in that spirit I present a Benoni from round three of the Fareham open.
Spanton (170) - Paul Northcott  (160)
1.d4 c5 2.d5 e5 3.e4 Nf6 4.Nc3 a6 5.a4 d6 6.Be2 Be7 7.Be3
Normal is 7.Nf3, which Stockfish10 and Komodo10 reckon leaves White the equivalent of at least a half-pawn up. But, to a large extent, it does not matter what White plays - the engines apparently love White's central space advantage.
7...h5?!
A hard-to-fathom novelty, which I thought was perhaps directed against a g4 thrust, which I had been considering. However, three rounds later I was sat next to PN when, playing with the white pieces, his game began something like 1.c3!? d6 2.h4!? I could not help thinking he is one junior who, for better or worse, has escaped the clutches of coaches.
8.h3
The engines agree this is best, one possible point of 7..h5?! being to follow-up with ...Ng4.
8...Nbd7 9.a5
Asking Black how he intends developing his light-square bishop.
9...b5
Stockfish10's top choice, but after …
10.axb6 Qxb6 11.Ra2
... we have a queenside set-up in which it is generally acknowledged that an a pawn is weaker than a b pawn.
11...Nf8 12.Nf3 Ng6 13.Qd2 Bd7 14.0-0 0-0 15.Rfa1 a5 16.Ne1 Ra7

Does the pawn-push b4 work for White?

*****
*****
*****
*****
*****
17.Bg5
I tried to make 17.b4! work but rejected it because of 17...Qxb4 18.Rb1 Nxe4 19.Rxb4 Nxd2, missing that after, say, 20.Rb6, the black knight on d2 is lost.
Nevertheless, White is also much better after the text.
17...h4
This comes to be Stockfish10's choice, but Black is in big trouble whatever he plays.
18.Nf3 Qb4 19.Bd3 a4!?
This is the engines' first or second choice, but it loses a pawn.
20.Nxa4 Rfa8 21.Qxb4 cxb4 22.Nc3! Rxa2 23.Rxa2 Rb8
I expected 23...Rc8, but the text is the engines' choice.
24.Nd1 Bb5 25.Nxh4 Nxh4 26.Bxh4 Bxd3 27.cxd3 Nxd5 28.Bxe7 Nxe7
The smoke has cleared, and White is a good pawn up. But the extra pawn is on the kingside, and there is the added complication of Black having an advanced queenside pawn.
29.Kf1 Rc8 30.Ke1 Nc6 31.Ne3 Nd4 32.Nc4
The engines have generally agreed with the recent sequence of moves for both sides, and White's advantage seems assured as the knight on c4 prevents funny business down the c file.
32...Rc6 33.Ra4 Kf8 34.Ra7
Not 34.Rxb4?? Nc2+.
34...Ke8 35.Rb7 Ra6 36.Kd2 Nb3+ 37.Ke3 Nd4 38.Rb6
Not 38.Rxb4?? Nc2+.
38...Ra1!?
The best practical chance as after 38...Rxb6 39.Nxb6, Black has a bleak defensive task.
39.Nxd6+?!
The engines give 39.f4 Re1+ 40.Kf2 Re2+ 41.Kf1 Rc2 42.Nxd6+ Ke7 43.fxe5 Rxb2 44.Nf5+ Nxf5 45.exf5 with a winning rook-and-pawn ending.
39...Kd7 40.Nxf7?!
The engines are unsure how to proceed, but best seems to be 40.Nc4 Kc7 41.f4 Re1+ 42.Kf2 Re2+ 43.Kf1 Rxb2 44.fxe5, when White is two pawns up but Black's b pawn is a source of counterplay.
40...Re1+?
Much less clear is 40...Kc7, eg 41.Rg6 Re1+ 42.Kd2 Re2+ 43.Kd1 Rxb2 44.Nxe5 b3 45.Rxg7+ Kc8, when White is four pawns up but the engines give White just a slight edge after, for example 46.Nc4 Rxf2 47.h4 b2 48.Nxb2 Rxb2 49.h5. However, this looks a very difficult position for Black to play over the board.
41.Kd2 Re2+ 42.Kd1 Rxb2 43.Nxe5+ Kc7 44.Nc4 Rb1+?
The engines give 44...Rxf2 45.Rxb4 Rxg2 46.Ne3 Rg1+ 47.Kd2 Nc6 48.Rb5, when White is two pawns up and should win.
45.Kd2 b3 46.Rg6 b2 47.Rxg7+ Kc6 48.Nxb2 1-0