But I should have seen 25.bxc4 26.Qxc4+ Ne6 27.Rxe6! Rxc4 28.Re1 and white will win back the queen and simplify to an advantageous but not won, ending.

So I found a tricky line that requires good defense and doesn't lose outright -- that wins a small bag of peanuts, a penny and a "keep tryin'" sticker, right?

What? No peanuts? *phooey*

-- Looking at the correct solution...Damn it it's not very hard! It strikes me as exactly the sort of combination I need to find. The "one step removed from being obvious" combination.

There's no glaring flaw in White's position that can be taken advantage of on the move, but there's obviously plenty to work with if you can see how to create the problem.

I examined Bxc3 thinking it had potential because it pulled White's rook off the back rank... it never occurred to me that the N was important in the way it guarded e4.

I think (or my rather, my chess software thinks) that 25.fxe4 would have been better for White than the text. Black wins Q for R+N... still seems like a pretty good deal for Black to me. ;)

Material: Down a P. The Black Qb7 is on the same diagonal as the White Kg2. The Black Bg7 and Re8 are mobile, and Nh5 has a check Nf4+. The Nc3 is protected 1x by Rc1 and attacked 1x by Bg7, but Rc8 also attacks it implicitly, shielded by Nc5 and Rc4. The White Qh4 has one escape square, h3.

Candidates (25...): Bxc3

25...Bxc3 26.Rxc3

Candidates (26...): Ne4, Re4

26...Ne4 (forking Rd6 and Rc3)

The Ne4 is immune, because

27.fxe4 Rxe4 (threatening 28...Rxh4 and 28...Rd2++ 29.K any Qg2#)

Black gains the exchange and plays bxc4. He has lots of time to invade with his heavy pieces, because White has no serious counterattack with Ng1 so far out of play.

There's no glaring flaw in White's position that can be taken advantage of on the move, but there's obviously plenty to work with if you can see how to create the problem.>

If it makes you feel better, <JG27Pyth>, I looked at 25.bxc4 also, because little P moves can be good preludes to bigger things. Generally, they are not forcing enough, but sometimes they come up roses. Tal is a master at such moves, e.g., a move I do not yet understand completely, 27...b5 in Spiridonov vs Tal, 1969.

<25...Bxc3> eleminating an e4 guard <26.Rxc3 Ne4> (forking the ♖♖) <27.fxe4 Rxe4> when the threat of a discovered check will win the white ♕.

Other lines ? DAUT !>

27 fxe4 Rxe4 28 Bf4 Re2+ 29 Kf1 (or h3) Qg2#-so black must lose a rook for a knight.

Thanks for pointing that game out! Wow, b5 is either deep as hell or it's a brilliant positional deflection... I actually think it's super-deep super-accurate chess. It does set up the pin at the very end that clinches the combo. I think Tal calculated it all the way there. We are talking about Tal here... it's not just another GM. But that's just my not-very informed opinion. (and I didn't work through all the kibitzing around that move.)

Well, I missed this one. Because I found something else that I liked a lot and stuck to that. Check out this little sequence of pawn moves:

25...h6 and now the dark bishop must move. It can't create any serious threats immediately. In fact 26.Bd2 looks to be the only place where White doesn't lose material. (27.Bf6 g5 28.Qd4 Ne5 is wild but losing I think, 27.Bd8 g5, 27.Bf4 g5, etc.)

26...g5 forces 27.Qh3

Now, if Black can cover h4, then g4 wins the Queen, but the Queen is out of the game for awhile.

Maybe 27...Qe7 28...g4 29.Qh4 Bf6...

I used Toga II 1.3.1 to analyze, because I am still fathoming Smirin vs Beliavsky, 1989 and I am exhausted. I hate to burst your bubble, but here is the refutation.

25...h6 26.Bxh6 Bxh6 27.Rxg6+,

and the White Qh4 picks up whichever Black minor piece does not move (Nh5 or Bh6).