Markovich wrote on 10/14/10 at 13:08:11:
In my database of 5 million plus games, White scores slightly over 60% from 4.e4 b5 5.a4 c6 6.axb5 cxb5 7.b3. I don't know if statistics would worry me very much here, however, since White's game just looks good to me.
60% may be right, but I was looking only at recent games (2004 onwards) and lately White has been struggling.
Anyway moves are more important than statistics. Critical is
7.b3 Bb7 8.bxc4 Bxe4 9.cxb5 Nf6 10.Be2 Be7 11.0-0 and Black got a good game in Vallejo Pons-Kasparov, 2005 with
11...0-0 12.Nc3 Bb7 13.Bf4 Bb4 14.Na4 Nbd7 15.Qb3 Nd5.
The older
13.Ne5 from Anand-Karpov, 1996 may be safer; Flear in DW: the Queen's Gambit then suggests
13...Nd5 (originally a Psakhis suggestion apparently) following Tregubov-Reinderman, Bundesliga 2000 to claim equality. 13.Ne5 does seem playable in practice, even if Black can equalize with accurate play, at least White isn't risking much.
Flear also mentions an earlier alternative for Black;
11...Nbd7!? which looks very drawish if tournament practice is anything to go by.
(Note to self: I should look up the position after 7.b3 in a corr database)