On a casual glance 8...Bg4 is also possible, though 9.f3 Bc8 10.Nc4 Qd4 11.Ne3 should favour White anyway.

good man Johnny boy! At the very serious and probable risk of being laughed at: 8.b3 goes forwards and 8.Ne1 goes backwards. That being said I know nothing of the theory or Marin's ideas after Ne1, I just liked the look of Timman's play.

This is all very interesting, but what do the most recent posts have to do with Flank Openings?

topandkas wrote on 02/05/12 at 08:29:46:

As far as I know there is no clear way to sufficient compensation after 5...d6 6.Nc3 Bxa6 7.g3 g6 8.Bg2 Bg7 9.Nf3. I don't believe in the unnatural 9...Nfd7 and Black's results after 9...Nbd7 10.Rb1 Nb6 11.b3 always have been discouraging.
If the Benkö will be revived my bet is 9...Nbd7 10.Rb1 O-O 11.O-O Qa5 or ...Qb6, arguing that 10.Rb1 has limited White's choice somewhat.

I think the Benkö is theoretically not at the same good status as Grünfeld, QGD, or Slav but playable enough. Not many grandmasters have been using it too often recently.

Gilchrist is a legend wrote on 02/04/12 at 22:26:57:

Why is that so certain? If there is no refutation no computer will find it, no matter how strong it is. From a practical point, since 15 years or so computers have been used to find new dangerous lines in all kind of gambits as well.
In fact even some titled players are busy to try that with the Benkö, though I agree with little success.

Well, according to my special handmade TopCorrDatabase 2006-2011 (where engines' help prevail  ) Benko is not doing very well - 57.7 % for white (mainline 59.5 % accordingly) ..which is too much in correspondence chess. And I have to admit I never played Benko though I thought about it..   
I happened to know the hard way that computer chess engines are very, very good defenders...and I don't allow myself to underestimate their power. I try to concentrate on some of their weaknesses instead. 
Back to the topic: I'm very happy that because of Marin (and not only) the variation we discuss here attracted the attention of many good chess players again. Now Timman...tomorrow Carlsen or Aronian?.. Who knows? 

Great game by Timman! One of my favourite players of all time. 
As a correspondence chess player I always suspect positional pawn sacs as not healthy for obvious reasons (i.e. computers are way better in defending such positions with a pawn up than an OTB player). Not in this case though. I spent more than 2 hours trying to strengthen the black defence after the pawn sac. And what amazed me was the fact that even after the queens' exchange black was not ok. On the contrary, black's position went down like a cardboard house. Well, I don't say white is winning after the pawn sac, no...not at all. But, I say that this straight b3-idea is good enough to be played even in the correspondence chess at the very highest level.