32...♖xf3! 33.♖xe8 ♖xg3

If white takes the rook right now, he will be temporarily an exchange up, but he will have to give up his rook for the powerful d pawn: 34.fxg3 d2 35.♖d8 d1=♕.

34.♗f6 g5!

The white king escapes from the mating net, and white will still have to give up his rook for the d pawn.

35.♖d8

If 35.fxg3 d2! 36.♖d8 d1=♕+.

35...d2! 36.♖xd2 ♖d3

A third simple solution in time trouble would be
c)1... d7 2.Rxe8 d8Q+ 3.Re1 Rxf3 4.Rxd1 Rxg3 5.Rd8 Rd3 6.Rb8 Kg7 and black is a piece up.

Not at all! I was pointing out that different engines don't all agree. I have seen many times where Fritz gives move A as best whereas Rybka gives move B as best.

Plus my point was that what Fritz deems best isn't always the best move for a human to play. Fritz or Rybka isn't going to miss a tactic where a human player might! If I have a choice between a move that without question is winning and a move that 'might' be winning IF I've calculated everything correctly...then I'm always going to choose the move that clearly wins.

Kramnik was once asked a question as to why he didn't play the best move? He replied that his move was winning and so he didn't look further. He didn't care if there was a better move. He chose the move that was clearly winning.

In my opinion it's just practical to play the winning move that is clear than to play a move that may be winning but looks more complex. Even though to a computer the complex move may be 'best'.

On that note whenever you are winning, your opponent is the one who wants to complicate the position so as to give you a chance to miscalculate. Conversely, you should be trying to keep the position as simple as possible when you are winning.

So in my opinion, Qxe1 is best. On Qxe1+ Nxe1 d2 and black is winning--simple as that. The pawn cannot be stopped and white has no counterplay. But look at other posts and how much calculation went into Rxf3! Why choose a move that is more complicated?

32...Qxe1+ 33 Nxe1 d2 and the Pawn promotes on either e1 or d1. White has no counter-attack so the "sacrifice" is quite safe. This tactic comes up from time to time so is worth remembering.

Of all the posts, bickering back and forth over which is the correct answer, I like yours the best. Well said!

<Occam's razor> - (in basic terms) The simplest model tends to be the correct one.

Then there's that other saying - "There's more than one way to skin a cat."

Rybka 2.2, confusingly, finds Qxe1+ and suggests that; however, after Qxe1+ Rybka gives the eval at -4.5 whereas after Rxf3 Rybka gives the eval at -5 so it seems to actually favor Rxf3.

<32....?>

Target: 1:25;000
Taken: 0:49;012 (but some more after I quit, around par I think)

Material: Black up, ♙

Candidates: Rxf3, <[Qxe1†]>

-ML-
It's quite easy to spot that in this unbalanced position, Black doesn't just want to move his Queen. He's going to sac her here and now. There are two abundant ways, and at first I wanted to play Rxf3. Although a nice take-out-the-defender move, Qxe1† is more forcing:

<32....Qxe1† 33.Nxe1 d2 > now, with the White Rook off the board, White cannot counterattack at all. Black promotes with either d1=Q or dxe1=Q†. This is unavoidable, leading to a Rook up for Black.

Black to play (32...?) "Easy"

I have seen greater contrasts in pictures but not many.

Jaan Ehlvest looks for the all the world like a mid-level manager at the Christmas Party of a company that sells brand-X. This is not a man who has driven back both Karpov and Ivanchuk; this is a man who wants to tell you why brand-X is beneficial to your wife, your life and your home. Tanned, a little more weight than his prime and wearing a look that says, "it's a life, and I'm satisfied," he's a year older than me, has probably worked a few jobs I would find familiar, and we would have much to discuss.

Tamaz Gelashvili looks looks for all the world like a man who took the wrong meds or just got done visiting his fifth underground Russian opium den. He may or may not have slugged somebody and he may or may not slug you--depends on his particular mood. Somebody shouted, "Hey, Tamaz!," and he looked up to see a flash of light, not thinking the image would be forever immortalized on a heavily visited chess website. I'm not sure what to make of the red cord around his neck--ID badge at tournament? His shirt is rumpled, he is pasty-faced, his visage curious mixture of bored indifference contrasted with eyes that look ready to pounce. One sentence is devoted to his bio on CG. I'm not sure if we'd have much to talk about.

I'm pretty sure Ehlvest was posing at a Christmas party. I'm pretty sure Gelashvili was at a chess tournament. I'm also sure I'm having way too much fun writing about these photos.

32... Qxe1+

(-7.63) 32. ... Rxf3 33. Rxe8 Rxg3 34. Bxg3 d2 35. Rd8 Bd5 36. c4 d1Q+ 37. Kh2 Qd4 ....

For anyone who is up an extra 3 hundredths of a pawn on me, congratulations.>

The moves that jump out are 32 .. Ne4, 32 .. Rxf3 and 32 .. Qxe1+. The last of these lines looks to be the clearest:

32 .. Qxe1+
33 Nxe1 d2
34 Nf3 d1=Q+
A rook up and maintaining the attack

32 .. Rxf3
33 Rxe8 Rxg3
34 Bxg3 d2
35 Rd8 d1=Q+
36 Rxd1 Bxd1
Is up a minor piece with WQ side pawns about to be demolished.

32 .. Ne4
33 Nd2
Is less promising for black.