click for larger view

with PP for N and some chance of pinning Black's Q.

It takes some sitting-on-hands to even consider <Black's> threat, which is the pin on 2. So of course the solution is to make a <quiet move first>. Go figure.

White King and Queen being on the same row, Black will surely want to take advantage of that and put his Rook on b2 backed by the Knight. White can take advantage of that:

34.Bg2 Nc4
35.Bxd5! Rb2
36.Bxc4 Rxe2+
37.Bxe2
White Bishop captured a Bishop+Knight+Rook and black Queen is left with 5 pawns against a Rook+Knight+Bishop and 5 pawns. Not easy to win for Black

<Gilmoy> Yes, <38.Rh4> was a nice touch; so was <45.Bxf5> due to 45...Bxf5 46.Rd6+

Threatening 35.hxg6 fxg6 36.♘xe6 ♗xe6 37.♖xg6+.

click for larger view

Obviously, the black king should not step into the "h" file. However, the obvious 37...♔f7 also loses instantly to 38.♕h5, and white has simultaenous threats against black's king, queen and bishop!

click for larger view

By elimination, the only way to resist a little longer would be 37...♔f8 38.♕h5 ♗xa2 (or 38...♗f7) 39.♕h8+ ♔e7 40.♕f6+ ♔e8 41.♖h6 ♗g8.

click for larger view

After 42...♕f7 white's attack would come to a halt. However, white can now play 42.♗e6!. Now, 42...♗xe6 loses to 43.♖h8+ ♔d7 44.♖h7+ ♔c6 45.♕xe6+ ♔b7 46.♖xc7+ ♔xc7.

click for larger view

And 42...♕e7 loses to 43.♖h8 ♕xf6 44.exf6 ♔d8 45.♖xg8+ ♔c7 46.♖xb8 ♔xb8 47.f7.

click for larger view

Now, all of the above was just a threat, not an actual variation!

<34...♘c4>

Black's minor pieces are too far away from g6 to contribute to its defense. So his only way to contest white's threats is to make a threat of his own: 35...♖b2.

<35.hxg6>

But it is not enough to stop white's attack.

<35...♖b2 36.gxf7+ ♔f8 37.♕xb2 ♘xb2 38.♖h4!>

I spent a long time trying to make a knight sacrifice work. In an alternate dimension 34. Nxf7 or 34. Nxe6 might be a Tuesday/ Wednesday solution. But I just could not get them to shine, even when I tried sacking the rook as well. I soon had that nagging doubt that I had wandered down the wrong path.

That's when the "simple" (ha!) 34. h5 seemed to make more sense. The knight and rook are doing such useful attacking jobs that it doesn't seem right to give them away. So bring another piece into the attack.

I didn't consider the Nc4/Rb2 defence in part because I didn't think that black had time for it. He surely has to do something about the kingside attack and so doesn't have time to go annoying the queen - arguably white's least useful attacking piece at the moment.

Fritzie doesn't get too excited about the position. After a few minutes of whirring he finds 34. h5 and gives it a slightly sniffy +1.3. Nice to have but not exactly decisive.

He "thinks" that 34...Qb7 is black's best reply, keeping the evals around the +1.3 mark. He doesn't care much for the plan of 34...Nc4/ 35...Rb2. His little red tut-tut light comes on and he starts muttering about evals over +2.5.

So all told, this is a good Sunday bone to chew on, but not a great puzzle per se. The winning move (34. h5) is not all that difficult to find. Unless someone finds a better line than my steam-driven Fritz it's not particularly decisive either.

This game as a whole is interesting because Najer used the closed variation versus the Sicilian that is not seen much at the top level anymore. I still remember well the Spassky vs. Geller Match 1968 where Boris used it very effectively against Geller's Sicilian.

At the higher level Short tried it against Nakamura in 2013, but lost:

Black is probably considering 34... Rb1.

The obvious move is 34.h5:

A) 34... gxh5 35.Nxe6+ Kh8 (35... hxg4 36.Qxg4+ Kh8(7) 37.Qg7#; 35... Kh7 36.Rg7+ Kh6 37.Rh7+ Kxh7 38.Qxh5+ and mate in two) 36.Nxc7 wins.

B) 34... Rb1 35.hxg6

B.1) 35... Rh1+ 36.Kg3

B.1.a) 36... Rg1+ 37.Bg2

B.1.a.i) 37... Rxg2+ 38.Qxg2 Bxg2 39.gxf7+ Qxf7 (39... Kf8(g7) 40.Nxe6+ Kxf7 41.Nxc7 + - [R+2P vs B]; 39... Kh8 40.f8=Q#) 40.Nxf7 ends up an exchange and a pawn ahead.

B.1.a.ii) 37... fxg6 38.Kh2

B.1.a.ii.*) 38... Rxg2+ 39.Rxg2 Bxg2 40.Qxg2 and White looks better due to the threats Nxe6, Qa8+ and Ne4.

B.1.a.ii.**) 38... Ra1 39.Bxd5 exd5 40.e6 fxe6 41.Qxe6+ seems to win.

B.1.a.iii) 37... Bxg2 38.gxf7+ looks similar to B.1.a.i.

B.1.b) 36... fxg6 37.Bg2 either transposes or is similar to previous lines.

B.2) 35... fxg6 36.Bg2 either transposes or is similar to previous lines.

Am I wrong?

34.h5, Nc4;

(If ...PxP/h5; then Rh4 and QxP/h5, it looks like a wipe out. If Black plays ...f5; <to prevent mate> then PxP/f6 e.p. looks terminal for Black.)

35.hxg6, Rb2;
What else can Black do?

36.g6xf7+, Kf8; 37.QxR/b2!, NxQ/b2; 38.Rh4! The threat is Rh8+ and promotes, other than give-away, I don't see a way to stop it, the best Black can do is an exchange down ending.

You may want to upgrade your analysis software/hardware . Stockfish 64bit and a newish cpu find h5 nearly instantly -- and the eval holds around +3.0 thru 30 ply.

I think finding the Nc4 defense was the critical element. Qb7 may be objectively best... but it's passive-ish. It makes no one nervous. Nc4 does. One needs to calculate that White can weather the -- Nc4, Rb2, -- counterattack before playing h5. (And, though your intuition would have been validated in this position, I don't think it's enough to say 'black doesn't have time' in that position. The Q sac needed to be calculated for full credit, imho.)

I did some playing around with engine moves after the Qb7 defense -- and to me it was just annoying. The inhuman calculating strengths of engines is never more obvious than in tactically complex positions with active queens. Black defended with non-obvious moves I'd never consider and White effectively countered them and built it's advantage with attacking moves vastly deeper and more accurate than I can manage.

click for larger view

gives White an easy win. For example, here (position above) it's mate-two after 35...hxg4? 36. Qxg4+ Kh7(or 36...Kh8) 37. Qg7#.

Where it got deep for me was after 34...Nc4 (threatening the Queen-winning pin 35...Rb2) when White played the deeply calculated 35. hxg6!! (position below).

click for larger view

Here, White is essentially saying go ahead and pin and win my Queen, because I have a plan to force you to give up your Queen in return (also leaving you the exchange down for your Queen-winning pin) in order to avoid the promotion of my pawn.

Computer analysis on Fritz 12 confirms White's follow-up play after 35...Rb2 with 36.gxf7+! Kf8 37. Qxb2 Nxb2 38. Rh4! Qxf7 39. Nxf7 was practically perfect.

Your analysis is a bunch of junk, you moron.