Related to Fischer and the Alekhine defense, many viewed it as his back-up play to 1.e4, he clearly preferred it to closed defenses but I'm not sure he would have used it much had he continued to play. My opinion was that he used it more as a surprise than any real conviction in its merits. In the first WC72 instance that strategy seemed to work.

Related to the Alekhine's, there was a guy back in the 70's that used it as a primary defense with variable success. This game shows some of its drawbacks.

15 f6!! wins immediately.

If 15...exd4 16. Bxd4! is devastating.

a) 16...gxf6 17. gxf6

b) 16...Bxf6 17. gxf6 g6 (forced 17...gxf6 18. Rxf6! , 17...Nxf6 18. Nxf6 ) 18. Bb6 Qe8 (or 18...Qd7) 19. Nc7 Qxe4 20. Bd5

c) 16...Nxf6 17. gxf6 (17...Bxf6 18. Rxf6! ) gxf6 18. Nxe7

Fischer was not Tal, because Tal would surely have played 15.f6!!, as he played in this masterpiece Tal vs Ftacnik, 1985
See Tal's 15th move 15. f6!! with the same idea to sac Nd4 for the attack

<Phony Benoni>

Fischer was able to play brilliant games even before 1957. He was just thirteen in 1956, when he played the most brilliant (the game of the century) game of his life D Byrne vs Fischer, 1956

However, Fischer's style was many times refraining from sacrificing, unless it was "necessary", and that's the reason why it is very difficult to find long-term sacs as those with D. Byrne, in Fischer's games.

But it is almost impossible to say WHEN a sac is "necessary", mostly it is a question of taste. Long-term sacs can be "necessary" for a player, and "un-necessary" for another. And yet, Fischer several times blamed himself for having disregarded tactical opportunities and brilliant moves (see in "My 60 memorable games").

Please, read better what I wrote...

I didn't suggest 16. gxf6 after 15...gxf6, I suggested 15...exd4 and then 16. Bxd4! after the sac of Knight. Now you're analysing a totally different variation than mine, where Black doesn't take the Nd4.

It is not true that Fischer didn't miss a "crushing move".

For instance 16. Nxe7? (instead of 16. Qh6!)
Fischer vs J Sofrevski, 1967

or 34. Be5?, instead of 34. Rxf5! gxf5 35. Rg7 Qf8 36. Qxe6!! Rxe6 37. f7!, in Fischer vs Ulf Andersson, 1970

or 28. Nd2! (winning immediately as Black's Queen cannot defend anymore h7 and avoid mate) instead of 28. Rxg7? Fischer vs Reshevsky, 1961

as suggested by an amateur...

We could go on, by finding MANY games where Fischer missed winning moves... Sorry, but Fischer was not Tal in tactis.

Your line is OK, and you don't miss anything.

However, after 20. Qg4 Kg7 Black's f6 + g6 pawns don't seem to be a good shelter...

21. Rf5! (with the idea 22. Kh1 and 23. Rg1) g5 22. Nxf6!! Bxf6 (the best) 23. Bxg5 Bxg5 24. Rxg5+ and Black's position crumbles.

15. f6 gxf6 16. Nf5 Bxf5 17. exf5 Nd7 18. g6 fxg6 19. fxg6 hxg6 20. Qg4 Kg7 21. Rf5 <with the idea 22. Kh1 and 23. Rg1>

Your suggestion 21...Nc5 doesn't work

22. Kh1 Nxb3 23. Rg1 g5 (ok) and now not 24. Nxf6? but 24. Bxg5! fxg5 25. Nxe7 Rxf5 (25...Qxe7 26. Rxg5+ ; 25...Nxe7 26. Qxg5+ ) 26. Nxf5+ Kg6 27. axb3 is clearly won for White, as Black King has no shelter against Rook + Queen.

For instance 27...Qf6 28. Nxd6! White threatens also h4!

A wonderful line arises from 21. Rf5! g5 22. Nxf6!! Ndxf6 23. Rxg5+ Kh8 24. Rh5+! Nh7 (24…Nxh5 25. Qxh5+ Kg7 26. Kh1! ) 25. Rxh7+! Kxh7 26. Qh5+ Kg7 27. Kh1! Bg5 28. Rg1 Rf5 29. Bxg5 Kf8 30. Bxd8! Rxh5 31. Rxg8#

This could have been the “never-played” game between Tal (in place of Fischer) and Witte.

Let’s write down the full text…

“Tal” – Witte

Anyway...it is easy to understand why Witte's position was already lost after 15 moves. He played Kh8 (?), Ng8 (?) and he (as Black) had lost also a tempo by playing Qc7 and then again Qd8.

click for larger view

Let's suppose the line is correct. (I didn't find an improvement, but we are dealing with a 15-move long line.)

Now Black can jettison another pawn with 28...Qf4 29. h4 Qxg4 30. Rxg4 Kh5 31. Rxg5+ Kxh4 32. Rxe5.

The actual game lasted 44 moves just because Witte played very badly.

For instance, he could simplify by playing 33...Be6! and after 34. Bxe6 fxe6 he could build "a wall" defending his King, and the game could have lasted 60-70 moves or more, an endgame in which a player is an Exchange down can last a lot.

The line you proposed seems correct, after 32. Rxe5 White can exchange the Rooks and win another pawn from the couple a6-b7, and then 2 connected and passed Pawns make a "no-story" ending.

I agree with you that White is not winning "immediately" as I was thinking at the beginning.

However, probably 21. Rf5 is not so good as it seemed...

For instance, after 20...Kg7 I was also thinking about 21. Qe6!, maybe it is stronger than 21. Rf5

After 21. Qe6! Black cannot play 21...Rf7 for 22. Nc7!

if 21...f5 22. Kh1 and Rg1

If 21...Nc5 22. Bxc5 dxc5 23. Rad1 Qd6 24. Qh3 with a mortal bind.

Yes, 23...Qc8 seems the best for Black, although it hardly can do something to unbind it.