Playing Smith-Morra Gambit: Is it just for avoiding the Sicilian?

It is not sure which side would enjoy advantage in any given variation of the Sicilian after 1.e4c5 but looking at the huge number of matches in various chess database and so many games played at very top level, it seems that the Sicilian can be very tricky and always heated irrespective of colour. Gambit ideas in this line can go very deep even up to 9 or 10 moves and indeed several of them are quite spectacular. I find Smith-Morra a very decent one without involving too much risk for White and yet providing an exciting array of tactical and strategical opportunity. Surprisingly, it's occurance in top level chess is not as frequent as the Sicilian. Does that mean that White has better options playing the core Sicilian than playing this Gambit?

Here it might be mentioned that interesting positions can appear in not so familier or conventional moves in this line. For example:

The game was played between two quite low rated players, but there are certain questions involved in this. For example, what is the best way to decline? What should be Black's motive and plans after declining? 

Actually yes. The ways to avoid theory in the sicilian are usually Nc3 and c3, but these moves have a lot of theory by themself, so the best way to avoid theory is d4. Of course it's not the best move.

23 Re7, drops the rook without any compensation...

Here are some interesting games, but I could observe that examples with top players is very limited:

and here is a recent game:

and this one:

So I guess there is a 'Best Move' in theory to avoid the Sicilian after 1...c5?

2...b6 lacks any inner logic. The reason you have played 1...c5 is to take on d4 whenever white can't retake with pawn, to create a structure imbalance and a majority of pawns at the center.

So, 1.e4 c5 2.d3 cd4 is the way to go, and now on 3.c3 safest and quite likely best is declining politely with 3...Nf6!

This will also save you time and effort, since you can also answer 2.c3 with 2...Nf6!

Play is equal in the ...Nf6 lines, but black has good chances to play for a win.

For the record, white has nothing in the game Esserman- Van Wely you have posted if Black refuses to take the d5 knight. Esserman in his book gives 11...Rb8 12.Rc1 a5 13.a3 a4 14.Ba2 and continues with not-so-relevant moves, concluding that white has a clear advantage, but he fails (does he?) to notice that after 14...b4 white has absolutely no advantage.

Gambits. True gambits are rarely played at the higher levels since so many lead to a very drawish result after the fireworks are over. Even more so with modern computer programs that never miss tactical mistakes. High level players prefer positional chess that let's the opponent "drift" into a bad game.

Black does best by declining this dangerous gambit with 3. ...Nf6. Currently there is a Morra gambit tourny. I've played against all the so-called refutations. See my results

I was looking at a chess database and it appeared that almost all the top level games where White has not lost in a declined variation did not involve 3...Nf6.

3...Nf6!, not 3...Nc6?!

If anyone has any game pgn played between any GM involving 3...Nf6 which the Black side did not win, please upload it here.

pfren: Thanks, just corrected myself by editing!

Virtually impossible- there are a few THOUSANDS of them, some 98% of them coming via the 1.e4 c5 2.c3 Nf6 3.e5 Nd5 4.d4 cd4 move order.

Here is one with a black win of reasonable skill. There are a ton of these on the databases.

Main line in 2nd game demands c1 Bishop to choose a square. 10. Bb5 simply gives away t

he initiative. 10. h3 is maybe best