The first idea that comes to mind is 32.e6 Qb5 (else loses the bishop) 33.Ra8 Bg8 (33... g5 34.exf7 gxf4 35.f8=Q+ wins) 34.Qe3:

A) 34... Rb8 35.Qe5+ Kh7 36.Rxb8 + - [R].

B) 34... Nc6 35.Nxd5 with an extra pawn and threatening e7.

C) 34... g5 35.Qe5+ Kh7 36.Nxd5 Nxd5 37.Qxf5+ Kg7 38.Qe5+ Kg6 (38... Kh7 39.Be4#) 39.Rxg8+ and mate soon.

-----

Another option is 32.Nh5+ gxh5 (32... Kf8 33.Qxh6+ Ke8 34.Nf6+ wins the queen; 32... Kg8(h7) 33.Nf6+ wins the queen; 32... Kh8 33.Qxh6+ Kg8 34.Nf6#) 33.Qxh6+ Kg8 34.e6:

A) 34... Qb5 35.exf7+ Kxf7 36.Qxh5+ with a bishop and a pawn for a knight and attack.

B) 34... Qc8 35.exf7+ Kxf7 36.Rf6+ Ke8 (36... Kg8 37.Qg5+ Ng6 38.Rxg6+ with an extra piece and a winning atack) 37.Rf8+ Kd7 38.Rxc8, etc.

C) 34... Qd8 35.exf7+ looks similar to B.

D) 34... Qe8 35.exf7+ Kxf7 (35... Qxf7 36.Ra8+ wins) 36.Rf6+ Kg8 37.Qg5+ is similar to the subline in B.

-----

First, 32. Nh5+ screams to be examined because it gives black so few choices in reply. Any move which restricts our opponent like this deserves at least a cursory look through. The point is that the black king can't run away without falling into a royal fork or allowing white to play Qxh6 followed by a royal fork.

Then we have the analysis of the final position, or rather a couple of moves on from it. This is what we get after 34...Bxe6 35. Rxe6

click for larger view

Material is level but we own the sixth rank and the black king is mighty windy. The only slight problem is that it is black to move and try to set up a fortress. In other words, it's a GOOT position. Does black have a counterattacking move or a sound defensive move at his disposal?

First we need to identify white's threats. And the threat that appeals most to me is Rf6. Let's say that we get to it after a pass move like 35...Qc7 36. Rf6

click for larger view

That seems pretty terminal. White threatens Rf8# and Qg5+/Wh6#. Black can't defend against both.

So does black have a move that prevents Rf6 and all this carnage? 35...Qe8 allows 36. Bxd5 which can't be good for black. Everything else seems to be delaying the inevitable.

Fritzie finds 35...f4 (which I hadn't spotted in honest mode) but it doesn't help after 36. Qg5+ followed by Qf6+ and gxf4.

What about...

36 Rxe7 Qxe7
37 Bxd5+ ...

32 Nh5+ gxh5
33 Qxh6+ Kg8
34 e6 (Q moves somewhere generic)
35 exf7+ Kxf7 (forced)
36 Rf6+ Kg8/Ke8
37 Rf8#

The main alternatives I see for Black (other than those covered in my previous post) start with:

34 ... Qe8
35 exf7+ Qxf7
36 Rf6 Qg7/Qh7
37 Bxd5+ Nxd5 (diversion)
38 Rg6+

White wins the queen, and is up a couple of pawns a well.

32. e6 looks attractive. Also interesting is the f6 square which is defended only by the black K.

Looks to me that clearing the c1/h6 diagonal to allow the white Q to attack the black K is key. So, best is

A) 32. Nh5+ gxh5 33. Qh6+ Kg8 34. e6

However I tried both solutions on my computer the result was:

32 e6 mate by move 50.
32 Nh5 mate by move 57.

I had chosen the faster solution to checkmate.
White controls the center &
Black immediately loses a Bishop & a Queen.
Then a white Bishop can join in with the fun.
e6 was better.

White will win because he is queen for rook ahead. The nimbler queen will dance rings around the rook and hoover pawns. It's not an instant kill (it's hard to give mate when you've only got a queen in the attack), but white would have to do something very silly to lose (or allow a draw) from here.

36. Rf6 by contrast is mate in 5 (according to my silicon buddy). I actually thought it mated faster than that, but Herr Fritz prolongs the agony by throwing pieces away.

The other drawback of the 36. Rxe7 threat is that black can counter it, say with 35...Qb5. Then white will need to look for another threat. And that might lead him back to ... 36. Rf6.

The point behind Rf6 is that it creates unanswerable mate threats. Which means that you probably get to the pub faster than the Rxe7 line which ends up with a material plus but a bit of work to do before you can nail the coffin lid shut.

32.Nh5+ is an interesting idea but I don't think it works. 32...gxh5 33.Qxh6+ Kg8 34.Ra8+ Bc8.

But I'm wondering why not the simple 32.e6? Doesn't that win a piece for nothing? 32...Qb5 33.Ra8

I'll just go with 32.e6.

-------------------

32 e6 mate by move 50.
32 Nh5 mate by move 57.

Its all too horrible! >

click for larger view

A Mallahi vs A Al-Bahadly, 2001

White to move and play his 32nd move.

I analyzed this for over an hour this morning ... but never got around to posting my thoughts.

I saw 32.Nh5+!?, but rejected it as dubious, as there was no iron-clad mating line. (I also had already become convinced that e6 was winning, and never went back for a second, really critical look ... at Nh5+.)

Instead, I came up with the following line ... which I checked with several chess engines:

<[32.e6! Qb5[]; (Forced?)
Otherwise, Black drops a piece.

33.Rd6!, Hits d5.

< (RR33.Ra8!? Rb8;

And Black tries to pare down.

(I don't like to trade pieces when I am making an attack.) ) <<<>>> >

33...Bg8; 34.Qe3! Kh7; 35.Qe5 Qe8▢;

Black must try to retreat, (to defend); otherwise his King will come under a vicious <unstoppable> attack.

< (Or 35...Qb1; 36.Nxd5,

In some lines, White will have several connected Passed Pawns ... running right down the middle of the board, and Black will be powerless to stop it.) <<<>>> >

36.Nxd5 Nxd5; What else?

37.Bxd5 Rb2+; 38.Kh3!,

click for larger view

and Black looks helpless to prevent the coming Rd7+, when White will either mate, or win Black's Queen. ] >

There are a few small (computer) improvements to my analysis ... and Fritz DOES choose 32.Nh5+. However, both lines win, (according to the chess engines of Fritz 13, Houdini 3.0 and the latest version of Deep Shredder).