I don't have enough time to analyze the position, maybe tomorrow I'll continue it. After 9...Qa5. 10.Bxe7 Nxe7. 11.Qd6! Qxc3. 12. Ne5 or 12. Rc1 is good for white. Black's king is cemented at the center.White has a strong play on C and D open files. And with white's knight landing on e5, his initiative will strengthen.

Hm if 9. Ba3 Qa5 10. Qd6 0-0 11. Rab1 Qxc3 12. Rfc1 Qf6 13. e5 Qh6 14. Bd3 can't make that much of it.

Or 9. Ba3 Qa5 10. Bd6 Qxc3 11. Rc1 Qf6 12. Qd2
I don't know, white's still a bit better, but I don't know how long that'll hold.

I still like Qd6 better, it seems more direct and to the point. I'm not sure it matters in the long run, black can probably equalize without too much trouble there too, but white seems a little more coordinated in the lines I've seen.

Taking a closer look at the battery after:

9.Ba3 Qa5
I’m sure you’re aware that Panno met 9.Ba3 with 9...d5 and that seems to be best in my opinion.

10.Qd6.
I don’t think Black can take the P.

10...Qxc3 11.Bb5
Threatening to remove one of the defenders of e7. Black has to find a way to defend e7 and I don't think there's a satisfactory way to do it.

11…Qa5
The idea of this move is to hurry back to d8. Black has lost time grabbing the c3P. He can’t play 11...0–0 because 12.Bxc6 wins a piece and other tries look to fail:

If 11...a6 12.Rac1 Qf6 (To defend e7) 13.e5 Qg6 14.Bxc6 Nxc6 15.Rxc6.
If 11...Qf6 12.e5 Qg6 13.Nd4 Threatening to remove both defenders of e7 13...Nf5 14.Nxf5 exf5 15.e6 keeping the threat alive. 15...fxe6 16.Bxc6

12.Nd4
Threatening to remove both N's from the defense of e7.

12...Qd8
He still can't play 12...0–0 because of 13.Nxc6. Black is behind in development, but he does have a material advantage and unless you can find a way to dig out his K at some point you’ll be left a P down.

I didn’t check any of this with an engine so you might want to do that to see what its evaluation is (or if I've missed anything!) but I suspect different engines will evaluate it differently and engine analysis is unlikely to be conclusive.

That was what I would play as well, but Fromper's line of thought involved Qa5 first.

I saw both possibilities, but I had a harder time coming up with a response after Qa5, which is why I asked about it.

--Fromper

Which is why I went into 9. ... Qa5

I've heard that several times from high level intermediate players (1700-1900 USCF ratings), but I've never heard a master say that. I think the masters understand that you don't necessarily need an obvious attack to have compensation for a pawn. This isn't some wild attacking gambit - it has a sound, classical basis.

There are quite a few masters who use this gambit in slow games. I even saw one master win a game with it on his way to an IM norm last weekend at the Miami Open.

I really think I could learn a lot about chess in general by learning as much as possible about what it means to have compensation for the pawn in this gambit in situations where there's no obvious attack.

--Fromper

I think you should experiment with not worrying about the c3 pawn in lines with Bb4xc3. It's not always "correct" but you can't freak out about it - you have to play actively. And if he wants to waste time winning a second pawn and falling further behind in development, let him.

You'll lose some games this way, no doubt, but you've got to learn when you don't have to worry about things like that. One habit I see sneaking into your comments is that you - down a pawn - start thinking defensively if you might lose another pawn.

Dude, you're down a pawn already. The endgame is lost. (Maybe not ... but practically, yes.) Your best chances to draw or win are to play actively and force concessions. And if he wants to chase pawns ... great! Just increase your attack and increase your pressure, and let him count all his precious pawns after you've mated him.

As the saying goes, "In for a penny, in for a pound."