There are different types of gambits. Some of them are quite sound and playable, others are just cheap tricks.
Are Gambits Worth It?

The gambits are played to gain space and development. You are not sacraficing that pawn to win it back immediately. It's all about who will get their pieces out faster. To achieve development advantage people sacrafice material: pawns and pieces. Look at Anand vs Topalov game where topalov played with a piece down.

If I remember correctly GM Roman Dzindzichashvili said gambits work better for white then for black. Some white gambits are playable while almost all black gambits are more or less dubious.

. . .
Bottom line - if you can get something for it, go for it!
That pretty much sums it up. Material is only one type of advantage you can have. If you can trade material for the equivalent of something else then it's fine. When it's not clear how much compensation you get is when you have to think twice. As players get better so does their ability to evaluate and so more and more types of sacrifices (or gambits) are viable/don't work.
BTW I thought there were at least a few sound gambits by black... like the Benko and that line in the Samich out of the KID? I don't play either of those so not 100% sure :D I'm you you guys can name others...

The Averbakh Gambit in the Nimzo-Indian 1.d4 Nf6 2.c4 e6 3.Nc3 Bb4 4.Bg5 h6 5.Bh4 c5 6.d5 b5 and the Marshall/Rubinstein gambit in the Four Knights 1.e4 e5 2.Nf3 Nc6 3.Nc3 Nf6 4.Bb5 Nd4 are other examples of gambits that have done very well for black at the GM level. Dzindzi doesn't like playing gambits, but if they win for you, go for it!

oh..the marshall gambit of course
I totally forgot about this one too :p I play the exchange ruy anyway heh.

Technically the Ruy offers a gambit for black after:
1. e4 e5
2. Nf3 Nc6
3. Bb5 a6
4. Ba4 Nf6
5. O-O
And since the most popular response by far is 5. ... Be7 I would definitely say that gambits can be worth it. The Ruy offers a gambit and it is rarely accepted.

LMAO Riga!!
What a game BTW!
LBGM - "Unlike the Queens Gambit where you can always take back the gambit pawn ..." - that's why it's not really a gambit mate - as has probably been said countless times on this site.
As my handle would suggest I play the Benko Gambit every time I get the opportunity. I also love the Roy/Marshall as Black (or White for that matter). The two defences are almost chalk and cheese. The latter is all tactics, the former highly positional .... and I love them both dearly. The only problem I see from the "dark" side of the board is that White has to play into them - if they decline, quickly return the booty, or launch counter measures ..... that's when you start to worry about your gambit "gift".
I've posted this before in other forums but I think it's worth reposting. I also think the "sentiment" applies to many gambits:
From Neil McDonald's "The Benko Gambit Revealed":
"When you play the Benko Gambit you set White the astonishingly difficult task of exploiting the extra pawn: it is so difficult in fact as to be virtually impossible, and for this reason many players hand back the pawn as quickly as possible in return for a positional, rather than a material, plus. You will need patience, endurance and fighting spirit no matter what form the game takes. It is a great consolation to know that Black is acting from a sound positional base - the endgame is normally good for him and the onus is therefore on White to "do something". This is too much pressure for most players, even Grandmasters, and the thing they "do" is often horrendous."
My emphasis - from GM Neil McDonald's book
Are gambits worth it ?? Have you ever experienced your opponents' hesitation when it suddenly dawns on them that they're woefully behind and perhaps mortally wounded because of your superior development? .... or that perplexed look they sometimes get when they realise they are one crucial tempo behind: You bet your arse they're "worth it"

padman, in the Capablanca-Lasker game, theory has regarded 6...exd4?! as a mistake since around the time of that game. I think Capablanca had the better position through much of that game, and Lasker's choice was probably black's strongest line. The main line of the Open variation runs 6...b5 7.Bb3 d5 allowing white to recapture the pawn. Here is another example of why accepting the gambit is thought not to work:

Still LMAO I'm not sure whether Riga is taking the piss or not (my money is on taking the piss) but those were 69 and 73 move refutations for those not counting ... and beautifully played games by White and Black.
You are taking the piss .... right?

Topalov won the first game and Anand the second and the fourth game of their current title match by gambiting pawns straight out of the opening. Gambits can be deadly if your opponent is unprepared for them.
Benko, I am just trying to summarize the current state of theory in the Ruy Lopez. It strikes me that the results of those games did much to discourage players from trying to hold onto the gambit pawn. I would agree though that Lasker's line looks like it might be vaguely playable if your opponent wasn't too strong. I don't think I'd want to try Trifunovic's approach though. I've won a lot of games on the white side where my opponent has tried to hold onto the pawn.

padman, as I expect you know, it is quite common in GM practice for a single game to bust a variation. GMs all see the game, and then they stop playing that line. If with best play by both sides, a certain move leads to a loss, why would you continue to play it?
If you want to play 6...exd4, then be my guest. I think it should be obvious though that there have been thousands of games where black plays 6...b5 7.Bb3 d5 giving back the pawn or 5...Be7 declining the gambit, and only a handful of brave souls have been willing to try to hold onto the gambit pawn.
5.0-0!? is the gambit move. Black can choose to accept the pawn with 5...Nxe4 6.d4 exd4 or return it by playing 6...b5 7.Bb3 d5 or decline it by playing 5...Be7 instead. It is probably one of the most common gambits. White does end up material behind in the Lasker line. It's just that most GMs regard the gambit as so strong that they choose to decline it, or return the pawn.

Yes, gambits are definitely worth the material deficit. Tal and Petrosian won many games using various pawn sacrifices and gambits. In fact, I am a loyal player of the King's Gambit and another gambit in the Sicilian, which goes 1. e4 c5 2. c3 d5 3. exd5 Qxd5 4. d4 cxd4 5. cxd4 Nc6 6. Nf3 e5! which is rarely accepted. This gambit was adapted from one of Tal's pawn sacrifice games.
If there is any kind of Favourable Omen, go for it! Even a sacrifice to cause confusion (of a pawn) is worth it, because you will win on time!
In short: Gambits ARE worth it!

Tal and Petrosian won many games using various pawn sacrifices and gambits.
Perhaps you mean Bronstein or Spassky instead of Petrosian. The former world champion Tigran V. Petrosian was a very conservative player.

@ Fiveofswords - So a gambit played right out of the opening - with the ensuing initiative prosecuted over a long period of time is the same as a sac in the middle/end game?
I couldn't disagree more. One is a strategic decision made right at the beginning of the game ... the consequences of which are not reversible. The other is not - the mid game sac may be strategic in nature but it's often made to open up a tactical shot.
The opening gambit sets a "tone" for the entire game. The mid/endgame sac is only a "chord" in the game.
... and if you believe that ... can I interest anyone in a used bridge ??
It comes to me that the gambits is are very popular among the chess world currently and they are widely used. But is it really worth it? I mean alot of grand masters and international masters would tell you that it is worth sacrificing a pawn or two for rapid developement, but is it really worth it?
I mean if your not careful, you can lose a tempo and all of your rapid developement would go down the drain. So, is all that stuff worth it? Unlike the Queens Gambit where you can always take back the gambit pawn on c4, the other gambits you sometimes won't get the pawn back at all. So are those gambits good enough to win you games?
This is my question. Does anyone have suggestions?