MNb, my PhD in Theoretical Physics and subsequent research has taught me one important lesson: sometimes there's just no point in trying to "understand" statistics...!

My "data" is not statistically significant, but for the record this is where I found it (I prefer to use publically available data!):

www.chesslive.de, position after 1.e4 c5 2. Nf3 d6 3.d4 cxd4 4.Qxd4 Nc6 5.Bb5 Bd7 6. Bxc6 Bxc6 7.c4 f5.

A search on this position within this database with criteria White >=2400 and Black 2400 yields 3 games: Turov - Vaulin 1998, 1-0; Hasangatin - S.Ivanov 2001, 0.5; Belikov - Solovjov 2002 1-0. Widening the search to White >=2300 and Black >=2300 yields 7 games with White scoring 4/7 (57%).

Now, none of this is of any statistical interest to me! The validity of the data, the grades, the search criteria, the size of the database etc. is not the issue, and I'm sure you have access to a bigger and better database. What is interesting is that both Belikov and Dzhumaev both got good results as White. I haven't looked at the games yet in detail to see whether these results had anything to do with the opening. Was Dzhumaev's win the result of a big edge from the opening?

I suspect most of us are guilty of basing opening evaluations on the result of a game. John Emms makes an interesting comment in his book on the Open Games for Black, that often our view of an opening variation is based on a single experience with it. A bad experience can make us lose confidence in an opening. I've had one bad experience in the 7...f5 variation, but perhaps the games we've identified above indicate that White has reasonable prospects. The point I was trying to make in my previous post is that the recent games with this line seem to suggest there is no theoretical reason for White to fear this kine more than other lines in the 4.Qxd4 variation, and no reason to avoid it with 7.Nc3 rather than 7.c4.

Well I'm a mathematician by training so I certainly wasn't intending to demonstrate a statistical edge for White from a three game sample!!! Personally I believe statistics from large databases are over-rated in terms of providing information about the strength of opening variations.  Up to at least 2200+ I think the absolute strength of any particular opening variation is not the most significant control on the result of the game. (I hope that statement doesn't get me banned from ChessPublishing.com!) 

What I was trying to say is that in games between players each graded over 2400, White has done OK. I chose this arbitrary cut-off grading because I think that someone graded >2400 who plays the Sicilian against another 2400+ is likely to be quite well prepared!

I don't know of a particularly good line for white, however both (4.Qxd4 Nc6 5.Bb5 Bd7 6.Bxc6 Bxc6 7.c4 f5) 8. Nc3 fxe4 9.Ng5 and 8. exf5 Qa5 seem to lead to interesting unbalanced positions. Does anyone know of any recent "theoretical" games in either of these lines? If not I'll browse chessbase once more...

Statistics is a dangerous subject. My private database shows a mediocre 45 % after 7.c4 f5.
But I would be glad if Unlucky or someone else could give a good line for White in this variation. Maybe 8.Nc3 fxe4 9.Ng5 is not too bad.

In the Baker/Losetowin line, 13.Nxe5 Rfd8 and 13...h6 14.Bxh6 Ne4 15.Qh5 both resulted in draw, Vasjukov-Tal, Baku 1973 and Gipslis-Tukmakov, URS 1972.

White should definitely meet (1.e4 c5 2.Nf3 d6 3.d4 cxd4 4.Qxd4 Nc6 5.Bb5 Bd7 6.Bxc6) 6...Bxc6 with 7.Nc3!

This approach is also advocated by Murray Chandler in his "Bashing the Sicilian with Bb5" video (he recommends this line for White after either 2...d6 3.Bb5+ Nc6 4.d4 or 2...Nc6 3.Bb5 d6 4.d4).

I am not so sure Black can claim full equality in the line given by MNb. 

Perhaps it's a bit late to add to this discussion...in case it isn't, I've been playing 4.Qxd4 for over 25 years (! is it really that long?), initially as a way to avoid theory and lately as a surprise weapon against "strong" players ("strong" to me is >2000ish). I've exclusively played lines with c4 when I can get this in (as opposed to Chris Baker's recommendation of lines based on quick development with Nc3). Usual Black replies are 4...Nc6, 4...a6, and even 4...e5 (though not many opponents have got away with this flagrant disregard of classical principles!). 

The most usual and probably best is 4...Nc6 5.Bb5 Bd7 6.Bxc6. Now some surprisingly good players (most recently a c.2200) have played 6...bxc6 which imo is ?!, and I think Black is struggling after 7.c4.

Otherwise, 6...Bxc6 7.c4 (Baker prefers 7.Nc3 aiming for quick development and Q-side castling, which is also OK) and Black generally chooses 7...g6 with an Accelerated Dragon style position. I think this line works OK for white because if Black's usually a (say) Nadorf player, he's unlikely to be happy in this type of position (he's probably windering how to launch his usual minority attack with a pawn on c4!). Some less experienced players have tried ...e6 (and even ...e5?!) when again I think white is slightly better due to the "weakness" on d6 and control of b5 and d5. Generally in this type of position white builds up slowly with moves like b3, a4, Qd2, Nd4, f3 etc and waits for Black to self destruct.

The reason I lost enthusiasm for the line was when a c.230 BCF annihilated me with 7...f5!? a few years ago, which apparently is an old recommendation from BCO. If anyone has a good response to this please post it and I may take the variation up once more!

Chris Baker's book 'A Startling Opening Repetoire' recommends this for White. I think that the best moves that he gives are already recommended in this forum, but it provides very good coverage of what to do as White and Black. 

I think better is 4..Bd7 to gain a tempo attacking the queen with 5..Nc6 as white cannot pin the knight anymore.