Improvements for Black, anyone?

Yes, sorry, <21. Rab1> is what I should have written.

Phony Benoni: <Helios727> My suspiciion is yes, by playing a well-timed ...d5. For instance, after 34...Ke7 35.Kc3 d5 36.exd5 Kd6 37.Kb4 Kxd5 38.Kxb5 we have this:

click for larger view

Now Black is a pawn ahead with a passed pawn and no structural weaknesses. Also, Black's king in threatening to invade profitably on the kingside while White's pieces are stranded on the queenside.

This isn't forced, of course, but it doesn't seem that White will be able to stop ...d5 in time. If the knight were on b4, for example, things might be a whole lot different.

But Fischer was one of the first to show that exceptions to this rule were possible. When the knight is a long way from the d5-square, control of that square does not mean anything. Moreover, this flexible position harbors considerable dangers for White. Black has practically no weaknesses (the d6-pawn is securely defended), whereas he has counterplay on the queenside, pressure on the weak e4-pawn and the possibility of an attack on the white king.> (From Kasparov's OMGP 4)

15.Na5 looks quite good for Black after 15…Bxe2 16.Qxe2 d5; but 15.Bxb6 indeed doesn't seem like a good move either. Earlier that year, in the Curacao Candidates tournament, Tal had already tried it against Fischer and got into a losing position - though he managed to escape with a draw (Tal vs Fischer, 1962). It was Geller in his game against Fischer, later in the tournament, who found a better idea for White – the subtle 15.Kh1!, and after 15...Rfc8 16.Bxb6 Qxb6 17.Bxc4 Rxc4 18.Qe2 Rb4 19.Ra2 Qb7 20.Na5 Qc7 21.Nd5 Nxd5 22.exd5 White gained an enduring pressure on the Q-side (Geller vs Fischer, 1962), though Fischer also managed to escape with a draw. In MSMG Fischer suggests an improvement for Black in this line which is supposed to equalize – that's probably what he was planning to play against Unzicker in such a case – and later some further refinements for both sides were discovered. But at any rate, it's clear that Unzicker failed to prepare seriously for this game when choosing this line, and that his understanding of its subtleties wasn't even remotely close to Fischer's.

http://www.imdb.com/title/tt2310458/

Then after, 26. Kf1 He writes in MSMG " What else? On 26. Rxa2 Qxa2 27. Re2? Rxc3! and in his 2007 notes he interposes, so White needs 27. Qxb5 and now 27...Qxb2+ 28. Re2 is holding for the moment. But Black has 28...Qxc3 which picks up a pawn, and soon a second one after 29. Ra2 Qe3 30. Re2 Rc2! 31. Rxc2 Qxe4+ 32. Kf1 Qxc2.

RandomVisitor: As RookFile, Kasparov and Huebner note, 21.g3 had good prospects to equalize for white:

click for larger view

Rybka 4.1 x64:

-0.10/22 21...b4 22.c4 Ra2 23.Qd3 Rc8 24.Rxa2 Qxa2 25.Qc2 Qa6 26.Nd2 h6 27.Rd1 Qb6 28.Kg2 Qd4 29.b3 Qc3 30.Rc1 Bg5 31.Nb1 Qe3 32.Rd1 Qb6 33.Nd2 Ra8 34.Nf3 Be3 35.Rd5 Ra1

-0.09/20 21...Rc8 22.Kg2 Qb6 23.Nd2 Qa7 24.Nb3 Qa6 25.Rfd1 Qb6 26.Rdb1 h6 27.h4 Rb8 28.Nd2 Qa6 29.Rxa4 bxa4 30.Qc4 Qa7 31.Qd3 Qc5 32.Nc4 Qb5

RandomVisitor: After 10...Bc4

click for larger view

Rybka 4.1 x64:

[+0.26] d=24 12.a5 0-0 13.Bxc4 Qxc4 14.Bg5 h6 15.Bh4 Rfc8 16.Ra4 Qc6 17.Qd3 Bd8 18.Kh1 b5 19.axb6 Qxb6 20.Rd1 Ng4 21.Bg3

There's a cute continuity mistake too. The actor who records the moves with a marker pen on the glass board in the detectives" office writes Kg2 as white's penultimate move, but in the next camera shot we see it as Kg3. He then writes Kf1 as the final move. When we see the moves from the correct side of the glass, the Kg3 mistake has been corrected. The conclusion is that the actor initially made a mistake, which someone noticed - well done that person! They got him to rewrite it but forgot to edit out the second incorrect camera shot!