xrt999: In 2008 retrospect, 28.Qxe4 is simply not a move Geller would make; the avoidance in his play is nebulous and detached. The question I ask myself now is did Fischer play 27...e4 aware that Geller would avoid 28.Qxe4, or did he think that Geller would in fact accept the pawn and concede the b file? My gut feeling is that Fischer intended 28.Qxe4 Rxd2; the year is 1962 after all. Geller denies Fischer and plays 28.Rf4
The tension is at b4 (Fischer is aware that at any point b4 by Geller will force Rxb4); Fischer has ample time to play ...Rb7 to allow the escape of the queen yet play continues on with ...e3. Fischer's intention is take the b file and give Geller the e file.
To promote this line with 28...e3, I feel that Fischer is intending that Geller play 29.Re4 to attack the pawn and concede the b file, but again, Geller denies Fischer for the second time. Geller allows the pawn to remain on e3, and instead forces the exchange with 29.b4. So, instead of even pawns and Fischer controlling the b file and Geller controlling the e file, Geller allows the pawn to remain on e3 and forces the rook-knight exchange at b4 with 29.b4. Fischer's game is now totally different; he has a passed pawn on e3 and the bishop in echange for the rook.
Geller, forcing Fischer to transform his play to that of a passed pawn in exchange for the rook, rather than allowing Fischer to take control of the b file in exchange for the e file, takes control of the game not by force, but by denying his opponent's desired line.