1.e4 c5. Black declines to copy White and starts a different game: a wing pawn goes for a centre pawn, the c-file opens for a black rook, and the two sides end up attacking on opposite halves of the board.
8
7
2
a
b
c
d
e
f
g
h 1
8
7
2
a
c
d
e
f
h 1 What the trade on d4 buys
Count the pawns after 4.Nxd4. White has one centre pawn, on e4. Black still has the d-pawn and the e-pawn, so in a pure endgame Black is the one with the better structure. That is the long-term half of the deal.
The short-term half belongs to White, who has a knight on d4, a knight coming to c3, and a free hand to castle either side. Castle queenside, throw the g- and h-pawns forward and the game becomes a race: White mates on the kingside or Black gets there first down the c-file. Sicilian games are not quiet.
Why I do not give it to beginners
Not because it is unsound. Because of the sheer number of things your opponent can do.
Meet 1.e4 with 1…e5 and you need answers to the Italian, the Ruy Lopez, the Scotch and a couple of gambits. Meet it with 1…c5 and you need answers to the Open Sicilian, 2.c3, 2.Nc3, 2.d4 as a gambit, 2.Bc4 and the rest, and in the Open Sicilian the theory keeps going past move fifteen. Lose the thread on move eight and your king is in the middle with an open e-file pointed at it.
There is also a habit problem. Sicilian positions are won by counterattack, and counterattacking is a skill you develop after you can defend, not before. Start with 1…e5 and come back.
A setup if you want it anyway
Play the version with the fewest forcing lines: 1.e4 c5 2.Nf3 e6, then …a6, …Qc7, …Nf6, …Bd6 or …Be7, and castle. No early …g6, no early …d5, no Dragon. You will not win many games in fifteen moves, and you will not lose many either, and that is the right trade while you learn what the pawn structure wants.
Then work on the thing that actually decides Sicilian games: calculation. In an opening where both kings are being attacked at once, the player who counts one move further wins, and no amount of memorised theory substitutes for that.