@Guitam:

How is Nc6 better? You still get exchange, but can not trade the 2nd pair of rooks.

A better transformation into the gambit: 1.e4 c5 2.d4 cd 3.Knf3 e5 4.c3 dc 5.Knxc3 with an additional backward pawn for Black

he just missed the right side of the discovered on 36 if it was nc6+ he could win.

Yeah, the Gashimov game is no brilliancy or anything. I just thought it was worth a look. However, from many of the Smith-Morra games I've seen, it seems black does just as well with the pawn push 3...d3 as with the pawn grab, 3...dxc3. In fact, even 3...Qa5!? looks quite playable for black........

D3 is a weak response!!!I wish someone would play that against me!!!!

@melvinbluestone:

White's play looked a bit unsound. At least from a 5-second analysis of mine :) Gashimov just trashed the guy.

Well, we know his rating NOW :)

3...d3 is fine.

Here's Gashimov playing the positional 3...d3 and getting into a bit of a crazy tactical game anyway: Karakehajov Kalin vs Gashimov Vugar . This is juniors under twelve competition. Note the ratings of the "kids".

I guess it is mostly psychological effect - you play d3 and instead of crazy-tactical game you switch to slow-positional fight. And obviously, positions are more flexible that once which occur in Smith-Morra gambit.

@Karkunt - of course, I've abused the ELO difference, but he put it a good fight at some point, and position with knight on e7 is plain funny

450 points in ELO difference !!!

I think d3 is unsound. teh result has to do with the ELO difference

3...d3 is OK if Black doesn't mind playing a hedgehog type of position.

I realize the push variation of the Smith-Morra has a pretty good reputation, but it certainly defies logic: Giving the pawn back and letting white develop a piece. I guess the idea is that white's c3 pawn is misplaced and disallows Nc3. So maybe white should try 4.c4!? after the pawn push......... but then he's wasted a move with 3.c3!  Hmmm..... maybe this 3...d3 is sort of logical after all!