I do not play this because of: 1.e4 c5 2.b4 cxb4
a) 3.a3 d5 4.exd5 (4.e5 Nc6 5.d4 Bf5) Qxd5 5.Nf3 e6
b) 3.d4 d5 4.exd5 Nf6 5.Bb5+ (5.a3 Qxd5; 5.c4 bxc3 6.Nxc3 Nxd5) Bd7 6.Bc4 Bg4 7.f3 Bf5
in both lines White's compensation is vague.

After 3.c4 I think 3. ... e5 is strong as you already stated in another thread. You said that it is more difficult to counter in the center than usually in the wing gambit but the opening of Hector-Kudrin,1989 looked convincing to me. As white has wasted a tempo with 3.c4, there is no direct need for quick counterplay.

Some weeks ago I discovered some new ideas in the wing gambit and since then I have played some games with it online.
The outcomes were mixed, but my overall score is not worse than with the open sicilian.

I discovered an old post from MNb where he wrote about the gambit:

MNb wrote on 01/27/06 at 20:39:32:



I think he is right when saying that white gets decent compensation with 3.a3 d5 4.ed5: Qd5: 5.Nf3 e5 6.c4!, this is very much fun for white, at least in blitz!

I found a game by Ruslan Sherbakov where he beat 5. ... e6 :

[Event "Pune op"]
[Site "Pune"]
[Date "2004.09.03"]
[Round "1"]
[White "Sherbakov,Ruslan"]
[Black "Das,Arghyadip"]
[Result "1-0"]
[Eco "B20"]
1.e4 c5 2.b4 cxb4 3.a3 d5 4.exd5 Qxd5 5.Nf3 e6 6.Bb2 Nf6 7.Bd3 Nc6 8.0-0 Be7
9.Re1 0-0 10.Ne5 Bd7 11.axb4 Nxb4 12.Nc3 Qd4 13.Be4 Qb6 14.Nxd7 Nxd7 15.Bxh7+ Kxh7 16.Qh5+ Kg8 17.Na4 Qd6 18.Bxg7 Kxg7 19.Qg4+ Kf6 20.Nc3 Nc5 21.Re3 Qc6 22.Qf4+ Kg6 23.Rg3+ Kh7 24.Qg4 
1-0

An excellent attacking game!
BTW, does Ruslan visit the forum from time to time? Perhaps he could share some thoughts about that game!
To me that looks as if white gets quite some pressure for the pawn with this Bb2,Bd3,Re1,Ne5- plan.

The variation I found least comfortable for white is 3.a3 d5 4.ed5: Nf6!?
Now white can restore material equality by 5.ab4: Nd5: 6.c3, but somehow that is ugly.
MNb, how should white continue after 5.d4 Nb4: ?

I tried your suggestion 1.e4 c5 2.b4 cb4: 3.a3 d5 4.ed5: Nf6 5.ab4: Nd5: 6.d4
on the internet but there was an opponent who gave me a hard time with 6. ... e5!? and now 7.de5: Bb4:+ 8.c3 Nc3: 9.Qd8:+ Kd8: 10.Bd2 Nd5 =+ or 7.c3 ed4: 8.Qd4: Nc6 9.Bb5 Be7 10.Nf3 0-0 11.Bc6: bc6: and the bishop pair gives black a slight advantage.

I already adressed that variation in my post, see above. 

What happened to the line 2. Nf3 d6 3. b4? A few years ago I read that this was definitely thought to be an improvement on 2. b4, but now I can't find anything on the line. Did it die out, or am I just not looking hard enough (which is likely)?

In the new Kaissiber 29 there is an interesting analysis of 1.e4 c5 2.b4 cb4: 3.a3 d5 4.ed5: Qd5: 5.Nf3 e5 6.Bb2 Nc6 7.c4 and 6.c4!?

Perhaps white's chances are good even objectively, but in practise they are enormous. There are games by IM Anatoli Zajarnyi who beat several strong GMs in that line (in tournament games), and some strong young Ukrainian players also tried the line.

In the meantime, are there any new ideas about 5. ... e6 or 4. ... Nd5 (see posts above)?

I also wonder what is the best against 3. ... ba3: , 4.d4, 4.Na3: or even 4.Ba3:

@MNb: do you have something to say about 3. ... ba3: ?

I started playing the gambit again in internet games and found out that some people like to decline with
2. ... b6.
2 good ways to counter that:

- 3.bc5: bc5: 4.Nc3 Nc6 5.Rb1 g6 6.g3! with a closed sicilian setup where the b-file (and the unprotected Nc6) are useful for white. In the right moment Nb5 might be annoying.

- 3-5 as above 6.Bc4!? with a kind of GrandPrix-attack.