20...Kg6 really isn't such a bad move, it's much more realistic to honor White with his brilliant response to it!

But something told me White's position was too good for there not to be anything better than the line above. It looked to me like the best plan had to involve pinning the e-pawn with Qc3 and building up the pressure on the f-file by doubling rooks. So I ran some possibilities through Fritz (4.01, an oldie but as far as I can tell not sigificantly weaker than more recent Fritzes). It came up with the following killer line that was based on that plan: 18. h4 g4 19. Rf2 f6 20. Rdf1 b5 21. Qc3 b4 22. Bxe5!! fxe5 23. Qc4 Rxf2 24. Rxf2, and with his Rook and Bishop out of play, all Black can do to stop the enemy piece invasion is give the piece back with Nd5. Black winds up a pawn down with no compensation.

20...Be6 fails to 21. Rxf6! Rxf6 22. Rxf6 Kxf6 23. Qxh6+ Kf7 24. Qh5+ Ng6 25. Qh7+ followed by h5.

The other reply to 18. h4 g4 19. Rf2 that I had Fritz look at was 19...f5. Here's what came up: 20. Qc3 Nc6 21. Re1 Qd4 22. Qxd4 cxd4 23. exf5 Rxf5 24. Rxf5 Bxf5 25. Bxe5 and, again, White is a pawn up with no compensation for Black.

<goldenbear> got it right ... I am very, very, very weary of personal attacks. (I have been criticized here for religion, politics, shoe color, you name it. And I certainly will not respond to threats, insults, harassment, etc. When someone intentionally messes up my handle, - an an attempt to cause ridicule - then demands information from me; it makes me wonder if such an individual is a few rocks shy of a truckload.)

Trying out the "pure" version of Fritz 9, (under the ChessBase interface); apparently Fritz already considers this a COMPLTELY WON POSITION (just before the sack on f7) FOR WHITE. ("plus-slash-minus") Secondly it likes (AGGRESSIVE) moves like h4, intsead of the sack on f7.

Thirdly, when analyzing these type of problems, I use a MODIFIED version of Fritz ... especially set to be hyper-agressive. (sac's) Modifying an engine under the ChessBase interface is very simple, simply right-click on the engine part of the window and select properties. You can modify dozens of attributes, and if you wish to go under the program itself, you can make many more modifications to any aspect of the engine that you like. (Just don't save it under the main engine file, otherwise you have to do like I did and re-install the program to get the "standard" settings back.) When analyzing with this aggressive engine, taking on f7 was the second choice of the program, h4 was the first, (I believe).

Fourth point - ...Kg6 may have been an error, and I was NOT the first to point this out.

Fifth, I don't have this game on the CD-ROM version of the informant, but I have many Informants for that period. (# 63-71 or 72) As soon as I finish up a few other projects, I may copy parts of that analysis here, as to shed some light onto how the players viewed this game.

As Ermenkov - Kovacevic already amply proved, even GM's can miss key moves at crucial points in the game.

I'm including it in my "demolition of pawn structure" collection, even though it appears to be a genuine sacrfice with an uncertain outcome. However, it is certainly a pratical OTB try, as it offers White winning chances with apparently no worse than equality with strong play.

I would have to say that this is a pretty accurate assessment.

As to whose computer is better, we're looking at vintage <Life Baiter AJ> in motion, with threats and vituperation aplenty.