NeverAgain: <tpstar: If Black omits 9 ... b5 (9 ... Be7 or 9 ... Be6) then White should get in 10. Nc4 with a good game.>
Generally, yes, but it doesn't seem to be appropriate to lump 9...Be6 together with 9...Be7. In Megabase 2012 there are 30,000+ games with 9...b5, with White scoring 54%. There are less than 2,000 with 9...Be6 10.Nc4 (57% for White), while 9...Be6 10.Nc4 occurred in only 251 game, with White scoring an overwhelming 80%.
Looks like Korchnoi was unwittingly railroaded into a variation he was not familiar with, something closer to Pelikan/Chelyabinsk than his customary Taimanov, and he went wrong early with 9...Be7 - the first mistake.
14...Rxc8 was the second mistake. This routine Sicilian move left the b-pawn without support. 14...Qxc8 was indicated. By this point the engines (Komodo and Stockfish) start wising up (they pass 9...Be7 without a comment) and give White a .
16...0-0? was just reckless. Black is down a pawn, White has a passed pawn on the queenside and an 'eternal' Knight on d5. It was not too late for 16...Rb8.
<tpstar: Here Black was a little too cavalier about regaining the Pawn with ... Nxb5/cxb5 Rxb5>
Well, your own Fritz 10 doesn't indicate a good moment to regain it, neither does SF6 - <23 ... Nxb5 24. cxb5 Rxb5 25. b4> is [3.08/41]. White can double the Rooks on the seventh rank and gang up on the King and/or the d6 pawn at his leisure.
After 16...0-0 Black is strategically lost, IMO, and the crooked plan ...h5, ...Bh6 and ...h4 did nothing to change that. He could have offered more resistance by challenging the a7 Rook, at least. As it is, after 26...h4 White's advantage is worth a whole piece.
<tpstar: 28. R7a6 seems even better>
... and 28.Bxe6 fxe6 29.Nc7 is better still. Because of the threat of 30.Nxe6, forking the black Queen and the f8 Rook, Black has no time to grab the b4 pawn.
Actually, it's a draw, as the black King cannot escape checks without allowing the b-pawn to queen:
So 41...e3 was an ingenious trap, a last-ditch effort, kudos to Averbakh for not falling for it.
The position after 9.Na3 occurred only once more in Korchnoi's tournament practice - in O Renet vs Korchnoi, 1988 at the Lugano Open he played 9...b5 and drew. A lesson well learned?