However, 1. e4 c5 2. d4 d5 is a reversed Albin-Counter Gambit. I think it's good for white, but can certainly throw him off if he's not prepared, I'd imagine.

There's certainlly this one:

Deming vs Cornell, 1980

From now on, Black scores 60% or more in every sub-variation, in every database>

E.g.:

<1 e4 c5 2 d4 cxd4 3 c3 d3 4 Nf3> It seems to me to be profitable to White NOT to capture this pawn immediately. There is no real way for Black to support it, and neither Bxd3 nor Qxd3 seems really great - neither of these pieces wants to be at d3. Therefore, White holds off on taking Black's 're-gambit' pawn.

<4 ... Nc6 5 Be3 d6 6 Nbd2> Here we see the major positive effect of 4 ... d3. It forces White to make a choice - either he gives back tempo, or must place his pieces on less strong positions, giving Black an easy defensive game. If White instead goes in for the Maroczy formation with 6 c4 g6, again things are more comfortable for Black than in the normal Maroczy Bind, because White's Queen or KB will have to lose time or position due still to that need to recapture at d3.

As a result, IMO, 3 ... d3 most likely refutes the Smith-Morra, not in the sense of that the gambit is unsound but in the sense that it appears that any advantageous formation White can secure, could have been secured under better circumstances in some other variation.

<6 ... g6 7 Bxd3 Bg7 8 O-O Nf6 9 c4 Nb4!> Here we see the trouble with having to recapture on d3. White will have to either retreat this Bishop or allow it to be exchanged. If 10 Bb1, the QR is shut in for at least a few more moves and White will have to lose yet another move with the B in order to develop it. 10 Be2 may be strongest here, leading into probably a more traditional Maroczy situation; instead White decided to develop another piece to try to take advantage of Black's loss of time with the N.

In any case, White's Queen's wing would definitely have to suffer significant softening in order to prevent this intrusion by the Knight. a3 would be necessary, but then Black can make definite counter plans on the Q wing beginning with ... a5. After this ... a4 threatens, and White will have to make some form of Pawn concession on the Queen's wing however he addresses this, which gives Black the characteristic Sicilian counterplay.

<10 Qb3 Nxd3 11 Qxd3 0-0 12 Bd4!> A good idea, essentially pinning Black's Nf6 for a lengthy time, as Black can hardly afford to allow the Bg7 to be traded in the Dragon formation, as well his QN has been exchanged and ... e5 would definitely be faulty, leaving the d6 pawn a permanent weakness.

<12 ... Qc7 13 Rfd1 b6 14 e5?!> An attempt to capitalize on space in the center and superior development, but it also seems that this move weakens White's tough grip on d5, and disposes of Black's need to worry of a weakness at d6 should the e-pawn need to advance.

<14 ... dxe5 15 Bxe5 Qd7 16 Qxd7 Bxd7 17 a4 a5 18 b3 Bc6 19 Re1 e6 20 Rad1 h6> To at least restrict some of the mobility of White's Knights. Despite the two Bishops and the White Pawn majority being somewhat weakened by 17 a4, Black has difficulty coming up with a strong plan here; in particular he is hampered by the fact that moving the Knight results in Bxg7 and numerous King's wing square weaknesses.

<21 Nd4 Bb7 22 Nb5 Bc6 23 Bc7 Nd7 24 Ne4> In addition, Black has no good way to shut out White's Knights on the Q wing.

<24 ... Rac8 25 Rxd7?!> This leads to a favorable pawn formation on the Q wing in exchange for the Exchange. It seems strong but later apparently looks like it's only good for a drawn position. In any case however the position is mainly even and White has only a slight advantage at best.

<25 ... Bxd7 26 Bxb6 Bxb5 27 axb5 a4!> The Q side pawns look to be strong, but they are not strong enough. This move actually makes White's passed pawns even further outside, but it also removes the support of the rear b-pawn and leaves the remaining two pawns open to attack, it will turn out White can't consolidate them enough for a win.

<28 bxa4 Rxc4 29 a5 Rb4! 30 Nd6 Be5! 31 Bc5 Rb3 32 b6 Bxd6 33 Bxd6 Ra8!> An additional nicety - White has never had a chance to make luft, and this assures the draw, as one of the passed pawns must be lost; the other will be neutralized by returning the Exchange.

<34 Bc7>
If 34 Ra1?? Rxa5! or Rxb6! and Black even wins!

Good questions, I´ll try a very general approach:

<I want to play the Sicilian more regularly in OTB play, but I don't know very much of the theory about the Smith-Morra gambit.>

The best video I´ve seen of the S-M (and I have seen like 4 or 5) is an old Daniel Martin one, maybe you can track it down; if I remember correctly he made an approach from white, but if you just write down the lines and check them with computer you will find the refutations for black. About books I can´t really tell you since I don´t use them that much, but maybe you will be able to find for free an article Albert Hoogendoorn wrote about the gambit: very general but very clear, so good for begginers.

<Are there any lines that I should know about?>

Hmm, there is the Bxf7+ famous trap with Bc4 if the d file is open and Rh8 is not yet connected with Q when K takes the B.

I.e., here (or with a similar setup):

click for larger view

The sharpest early deviation for white is to give a 2nd pawn with 4.Bc4:

click for larger view

...but will turn into rubbish if black knows how to defend.

Other than that... well, just study all the lines, is the only way of learning an opening. I would reccomendt you to get a good .ctg opening book (Rybka 4´s or Fritz 12´s one or something similar), and, once you know the general lines, then start studying famous games and analyzing with computer (the best engines nowadays are free: Firebird, Robbolito...) Remember that is equally important to be prepared against "good" moves than against "bad" ones.

<is there anything I should know about the 2.f4 Sicilian, except that 2...d5 equalizes quite easily?>

Sicilian, Gran Prix Attack
1.e4 c5 2.f4

click for larger view

Another good way is 1.e4 c5 2.d4 cxd4 3.c3 dxc3 4.Nxc3 Nc6 5.Nf3 d6 6.Bc4 e6 (not Nf6? e5!) 7.0-0 Nf6 8.Qe2 Be7 9.Rd1 e5 10.h3 (almost forced) 0-0 11.Be3 Be6.

But have a look here (DcGentle chessforum), and you might discover why. This line is the core of my private Smith-Morra analysis, which lasted several years, yes, right, I dedicated my analysis time to nothing else. For I wanted to know whether the common notion is true, that this gambit is unsound and only for amateurs. Good news: It's not. But current theory is right, with current theory Black can win with best play.

When I have finished the development of my new paradigm engine, then I'll write a book on the Smith-Morra.

And by the way, declining the gambit doesn't help Black.