Reading through this thread, I noticed that most of the recommended lines are in agreement with Palliser's recommendations in
Fighting the Anti-Sicilians. I would like to propose what I believe is an underrated line that perhaps is overlooked by many people:
1.e4 c5 2.Nc3 Nc6 3.g3 e6 4.Bg2 Nf6
A few things to note:
1) This line is not restricted to only e6 Sicilian players as the 2...e6 move order is not required and Black can play for a quick ...d5 in all major variations.
2) As Black is aiming for French structures, I think this makes a good companion to Palliser's main recommendation against the Grand Prix (1.e4 c5 2.Nc3 Nc6 3.f4 e6).
3) Although there are similarities with 1.e4 c5 2.Nc3 e6 3.g3 d5, the primary lines appear to lead to independent play often neglected by coverage in books on the Closed Sicilian. Although the 2...e6 lines are recommended by several books and have a reputation for being drawish in a few lines, this move order seems to cut out many of the theoretical recommendations leading to early simplification. In my opinion, these factors give Black better winning chances while still offering play with a solid classical foundation of central development.
Confirming my intuition that Black's play is logical and offers good play, I found games by several strong players who were successful with this approach. I noticed that Sveshnikov and Fier are regular practitioners of this line, usually playing directly for ...d5. I also noticed several grandmasters like Stocek would often delay d5, sometimes allowing the option of transposition into Open Sicilian lines with g3 in the Taimanov, Kan, or Scheveningen. This is certainly an option for people familiar with these Sicilians, although I believe this approach gives White more options. (From an optimistic perspective, one could perhaps say it gives White more opportunity to go wrong as many Closed Sicilian players are not familiar with the Open Sicilian.) People comfortable with e6 Sicilians could also consider 1.e4 c5 2.Nc3 Nc6 3.g3 e6 4.Bg2 a6!? which has also been played by several strong players and seems to have even less coverage in existing literature.
There are two primary approaches by White I considered in database research:
1.e4 c5 2.Nc3 Nc6 3.g3 e6 4.Bg2 Nf6
a) 5.Nge2 d5 6.exd5 exd5 (6...Nxd5!? is another independent option to consider):
7.d3 at first seems likely to transpose to lines traditionally reached via the 2...e6 move order. Although Black is solid in these lines, the enterprising 7...Bg4!? may cause White problems specific to this move order (e.g. after 1.e4 c5 2.Nc3 e6 3.g3 d5 4.exd5 exd5 5.Bg2 Nf6 6.d3 Nc6, I think White would prefer 7.Bg5 rather than 7.Nge2 transposing to the above line). If that is indeed the case, this subvariation would further support the argument for using this line to create greater winning chances as its sister variation in 2...e6 line has a reputation for being drawish.
7.d4 leads to IQP positions that seem to fall outside typical coverage of the traditional 2...e6 lines. I found several examples of 7...cxd4 8.Nxd4 Bg4 9.Qd3 and I think Black has a promising game while retaining good winning chances. Also 7...Bg4!? seems to be interesting and has been played by some strong players as well. My impression is the d5 pawn is very dangerous to take after an immediate dxc5. If that is the case, 7...Bg4 could well be more accurate than 7...cxd4. Currently, I see no problems for Black in this line and I have seen many instances of White going wrong early in the opening.
b) 5.f4 d5 6.e5
This is considered by Palliser in
Starting Out: Closed Sicilian to be the critical approach. If the move order chosen by Black is perceived by White to be inferior to 2...e6 3.g3 d5, then this approach may be the most probable attempt to seek an advantage from the opening. I believe this opportunity for contention creates greater winning chances as well by provoking White into sharper play. When White plays 1.e4 c5 2.Nc3 e6 3.g3, he usually knowlingly enters unambitious positions; however, not all players of 1.e4 c5 2.Nc3 Nc6 3.g3 would willingly accept these positions, perhaps preferring 3.Nf3 against 2...e6. As this is a rarer variation, it has a greater likelihood of being perceived as substandard, which creates a psychological advantage in practical play.
Palliser describes the position:
"White should close the centre with 6.e5 Nd7 7.Nf3 Be7 8.d3 when e5 is much more secure than in a standard King's Indian Attack position. A race situation might well develop, but White appears to be slightly faster while he tends to have the opposition monarch to aim directly at."
He then gives a game reference to V.Golod -
Junior, Givatayim 1998. I think his comparison to the KIA is a bit superficial and I believe there are drawbacks to f4 as well. f4 also creates weaknesses and could also be said to interfere with development when compared to the KIA French. Also central play with ...f6 should not be underestimated. I have noticed that Sveshnikov has used this lever several times in his games. When this break is effective at opening the center, f4 can sometimes be exposed as a weakening move.
When I looked at this line independently, the rerouting maneuver initiated by 6...Ng8!? also looked appealing to me. The knight is often effectively placed on f5 in positions like this and Black appears to have time to execute this maneuver. I was pleasantly surprised to see this move used by Bent Larsen when I did a database search on the position.
I believe this line also offers good winning chances for Black, and its use by several grandmasters as Black further confirms my intuition. It also helps that there may be misconceptions on the value of this variation in existing literature. I think it helps to have some familiarity with the French to properly understand these positions, and quite possibly a lack of experience in the French may be a major factor contributing to this variation's lack of popularity among Sicilian players. However, I believe the French-type positions arising from this variation are favorable for Black, and I would guess that most French players would be comfortable in these lines.
These are my first impressions after examining this variation. It looks very promising to me and I intend to do more independent research. Sometimes a gap in existing literature signals a great opportunity!