gambits where accepting is better

...or declining is worse (but the former sounds more upbeat).

Since most people offered a free piece get suspicious and end up declining, what are examples of gambits where declining is worse.  Evan's gambit comes to mind.

The Englund gambit for one, at least initially. So long as white avoids the obvious traps, black gets himself in all sorts of knots trying to hold position

I'm sure a lot of it comes down to whether you know the opening well and what type of player you are.

Odd question, the only gambit I can think of where accepting is probably slightly WORSE is the Queen's Gambit...  Accepting is almost always better!

broze:  I meant better for the acceptor, for example cases  where  maybe its slightly more poison to not accept the pawn.  If someone hasn't seen a gambit and are offered a pawn, chances are they'll turn it down.

Eberulf I meant better for the acceptor,

If a center (d- or e-) pawn is gambited, it's almost always better to accept.

So did I.

Eberulf> If someone hasn't seen a gambit and are offered a pawn, chances are...

I would take it.

Former World Chess Champion Jose Capablanca was waiting in a train station in New York one day, with his coffee, newspaper, and chess set, when a man approached him.  Gesturing at the chess set, he asked Capablanca if he would like to play a game.  Always delighted to play, Capablanca set up the board and removed his Queen from the board to make the game more competitive.  Annoyed, the man said, "Why did you do that?  You don't know me, I might beat you".  Capablanca replied,  "Sir, if you could beat me, I would know you".

I understand in the very early stages of the game a gained piece is advantageous regardless of what your opponent has up his sleeve.  But the creative lines for a gambit seem generally set up to exploit the opponent's acceptance of a piece.  Perhaps I didn't mean two GM's playing, with perfect knowledge of gambits.  But anyway, point taken.

Most gambits are better to accept.

Queen's "gambit" (1.d4 d5 2.c4) - accepted is perfectly solid. About the same as declining it. (But well, this isn't really a gambit)

King's gambit (1.e4 e5 2.f4) - declining it is a bit passive, theoretically best lines are in the accepted with ...g5.

Evans Gambit (1.e4 e5 2.Nf3 Nc6 3.Bc4 Bc5 4.b4) - better to accept.

Both Marshall Gambits - better to accept (1.e4 e5 2.Nf3 Nc6 3.Bb5 a6 4.Ba4 Nf6 5.0-0 Be7 6.Re1 b5 7.Bb3 0-0 8.c3 d5 and 1.d4 d5 2.c4 c6 3.Nc3 e6 4.e4).

Cochrane Gambit (1.e4 e5 2.Nf3 Nf6 3.Nxe5 d6 4.Nxf7!?)- better to accept.

Blackmar/Diemer Gambit (1.d4 d5 2.Nc3 Nf6 3.e4 dxe4 4.f3) - better to accept.

Morra Gambit (1.e4 c5 2.d4 cxd4 3.c3) - better to accept.

Staunton Gambit (1.d4 f5 2.e4) - better to accept.

From Gambit (1.f4 e5) - better to accept.

Tal Gambit (1.e4 c5 2.f4 d5) - better to accept.

Benko Gambit (1.d4 Nf6 2.c4 c5 3.d5 b5) - the ways to decline are (supposedly, in theory) eventually innocuous, best is to accept.

The Poisoned Pawn in the Najdorf is special - of course the best is to take it if White offers the chance, otherwise Black shouldn't have played Qb6.

Quite a few of these are actually so good to accept that the gambits are dubious... at top level you see both Marshall Gambits, the Poisoned Pawn, the Benko, the QG, the Evans as a surprise weapon - and that's about it.

et cetera.

The question the other way is more interesting. Which gambits are so dangerous when accepted, that the opponent had better decline them?

The Scotch Opening is a gambit, because after 1.e4 e5 2.Nf3 Nc6 3.d4 exd4 4.Nxd4 Black can more or less win a pawn with 4...Qh4. That's considered too dangerous so it's not often played. Bit far-fetched.

Perhaps one of the gambits in the Two Knights / Guioco Piano / 2.Bc4 complex, I don't know them really. Or a gambit in the English opening?

There is a discussion today on the Halloween Gambit for example.  The 10 move checkmate starts with Black accepting the free knight.  With the budapest gambit also, the checkmate line begins with a pawn being accepted.  Gambits are set up to exploit unknowledgeable opponents. 

As a rule, I always accept a gambit offering. I try to look out for traps but, when in doubt, as in I'm not familiar with a particular opening, my basic position is to make the person playing the gambit to prove to me why my taking the offered piece is wrong.

Eberulf> Perhaps I didn't mean two GM's playing, with perfect knowledge of gambits.  But anyway, point taken.

My point was somewhat different. I've played many games and have some knowledge of the theory behind my opening lines. If an opponent gambits a pawn in the opening in a way I've never seen before, likely it's not a very good move.

Have to disagree with you, Eberulf...

Gambits are not just "set up to exploit unknowledgeable opponents. "  At least not more so than any other opening...

Gambits are a way to trade one element, material, for others: space, development, piece activity, king safety, etc...

Sometimes this trade is desisive, for good or ill.  Other times it is not.  

Masters would not have played Evans Gambit and King's Gambit for so many years if they were relying on thier peers being n00bs.

Often people will play gambits without knowing all the theory that follows the acceptance of the gambit, and usually (at my level anyway) do not fully understand how to get the compensation for the gambit.  For this reason, I virtually always accept gambits (with the exception of the queen's gambit, where I don't like to abandon the center so easily-my other reason for declining the QG is that I like the exciting play that follows the Tarrasch Defense, and would probably feel like I was missing out on that if I didn't play that defense).

My brother always plays the Center Counter game when he plays chess, and so I sometimes can fool him into playing the Englund Gambit when he doesn't realize that I played d4:)

There's an old chess addage that says the only way to beat a gambit is to accept it.

ALWAYS take the first pawn.  NEVER take the second pawn.

The same rule applies to rooks.

In the Smith-Morra you should accept the first pawn and if he offers you a second definitely accept that too.  Same with the Danish gambit.  If you decline out of fear you need to practice some more!  Double-pawn early opening gambits are almost never sound.

Someone just played an unexpected gambit against me, 4....g5:

As I said, I usually accept such gambits. My thought process goes like so:

1. We're in a common theoretical position where I've seen all sorts of moves but never g7-g5, which suggests this gambit can't be very good.
2. What's the idea behind it? g2-g4 is a common ploy in other openings to attack on the g-file with gain of time, eg 5.Nxg8 Rg8 6.N (retreats). But it has to be worse with my pawn on g3, and worse still for Black (a move down) to take such a gamble.
   

I took the pawn and enjoyed a hearty advantage.

Actually even the computer will tell you that accepting the second pawn is a mistake.  My personal favourite (not because it's best but because it throws white off) in the Danish gambit is to accept the first pawn and follow it up with c2!? ... you do sac the pawn back because of Qc2 but your opponent is usually 'out of book' (while you're not) and you are not behind in development (or at least not much).

 I enjoyed that short story.