The Latvian Gambit rocks for Black. 
It's my preferred opening with Black. I play the Fraser Defense.
Here's an example of a game I played here on Chess.com.
Is the Latvian Gambit sound?
Sort:
Godspawn wrote:
From Sam Shankland:
There are many openings that are approved at top level, and as long as a student is playing solid openings rather than dubious ones, I will let them make their own choices in that regard. However, it would be a cause for concern if I saw a student playing the Latvian Gambit, for example. I also do not approve of goofy systems as white that have no ambition to fight for an advantage, rather just to get a position where all your moves will be easy (the London System and Colle System come to mind).
Welcome to the "there-is-only-one-way-to-play-chess" school. It's silly, really.
I asked my coach if I could play the Latvian Gambit until FIDE 2000. He said: of course you can!
@ cigoL:
Ah, yes, the good ole Fraser variation. Very clever: Black "sacrifices" material for no compensation, but in that way he's guaranteed he will not blunder it later! 
All that 4.Qh5+ stuff is funny, and probably winning for white as well (and 6.Qh4 is much better than 6.Qh3), but rather irrelevant: after the simple 4.d4, Black is as good as lost on move 4.
pfren wrote:
@ cigoL:
Ah, yes, the good ole Fraser variation. Very clever: Black "sacrifices" material for no compensation, but in that way he's guaranteed he will not blunder it later!
All that 4.Qh5+ stuff is funny, and probably winning for white as well (and 6.Qh4 is much better than 6.Qh3), but rather irrelevant: after the simple 4.d4, Black is as good as lost on move 4.
Maybe at your level, but seemingly not at club level or below. I've had a lot of success with the Fraser Defense.
alec40 wrote:
The Latvian Gambit is ultra risky it's anything but safe the Latvian Gambit is the equiviliant of holding a big red stick of dynamite that may blow up in the white players face or the black players face.
It's not for everyone if black likes to play very cautiously and safe doesn't like wild tactics chaos or gambling danger doesn't have the nerve to dance on the edge of a straight razor with a lit stick of dynamite the Latvian isn't for him or her.
After having played it OTB I concluded it it was more like gathering up a huge charge of static electricity and grabbing someone's arm; usually shocking(literally and figuratively) for the other guy, and can be very amusing... provided, of course, that you can stand the charge of electricity shooting through your arm in the first place.
Nov 16, 2011
0
#27
Thank the lord that most people do not carry around opening books on how to defend against an opening :).
With the argumentation some are using most openings should be completely forgotten for black.
No one is saying the Latvian is sound. But it's a nice surprise weapon.
No. It's junk. Learn chess, not tricks.
Risky openings are certainly fun to play. But the Latvian is 100% unfounded, positionally, and white can have a very good game without carrying any book- just by playing regular, sane developing moves. For example, try to bust that one:
1.e4 e5 2.Nf3 f5 3.Be2!?
Now 3...fe4 4.Nxe5 is a reversed KG, 3...Nc6 4.d4 a reversed Vienna, 3...Nf6 4.ef5 not any better than the above, and so it goes. White's innocuous third move is a very handy tempo won over well-known positions in reverse, and Black will be hard pressed to equalize.
BTW I have not found any games registered after 3.Be2, so it maybe is fun to try it - playing white, that is!
The wheel was invented several thousand years ago, no need to reinvent it. The Latvian was invented at the romantic age of chess, and was fun- but today, even a relatively weak chess engine will burst your ...f5 bubble with no trouble at all.
Ah, yes, that magical water of Latvia!
Shirov, one of the most inventive and risky players currently, is Latvian. The combination magician Misha Tal was Latvian, too. But neither Shirov, nor Tal ever played the Latvian Gambit! 
Nov 16, 2011
0
#31
The question is not whether it is unsound, but rather, how unsound is it?
I had to stop playing the Latvian because it was insufficiently unsound. Now I often play the Elephant Gambit 1.e4 e5 2.Nf3 d5 and I win more than half my games with it.
In Live Chess, it's not really that important to debate whether an opening is unsound. Very few people know these unusual openings enough to refute them over the board. In 5 minutes or less, Black's superior knowledge of the opening more than compensates for any deficiency in the opening.
To prove the point, I've been playing 1.e4 f5 2.exf5 Kf7 as Black and enjoying great results with it in 3 min and 1 min.
OK, the Elephant Gambit is most probably unsound, but it is "healthily unsound": Black's second move does contribute to Black's development, and while being one or two pawns down in the critical variations, Black has sane piece play. Actually a good friend of mine, Greek IM P.Pandavos, has played this thing with Black even against moderately stong opposition, with fairly good results. He has also beaten myself, several times (in blitz games).
It does need to be treated with respect, if White wants to get something out of the opening.
The Latvian, on the other hand, is fairly innocuous. 3.Nxe5 is certainly best, and very close to a total refutation, but after every sensible white third move (even after 3.Be2), Black is hard pressed to justify the permanent weaknesses he voluntarily created as early as move two.
pfren wrote:
Risky openings are certainly fun to play. But the Latvian is 100% unfounded, positionally, and white can have a very good game without carrying any book- just by playing regular, sane developing moves. For example, try to bust that one:
1.e4 e5 2.Nf3 f5 3.Be2!?
Now 3...fe4 4.Nxe5 is a reversed KG, 3...Nc6 4.d4 a reversed Vienna, 3...Nf6 4.ef5 not any better than the above, and so it goes. White's innocuous third move is a very handy tempo won over well-known positions in reverse, and Black will be hard pressed to equalize.
BTW I have not found any games registered after 3.Be2, so it maybe is fun to try it - playing white, that is!
The wheel was invented several thousand years ago, no need to reinvent it. The Latvian was invented at the romantic age of chess, and was fun- but today, even a relatively weak chess engine will burst your ...f5 bubble with no trouble at all.
pfren, I have a couple of comments for you.
- Maybe 3. Be2 would work against the Latvian Gambit. But this is irrelevant, since I've never seen this move in any of my games (and I play the Latvian Gambit exclusively as Black in reply to 2. Nf3).
- Maybe there is a reason why there are no games with 2. Nf3 in the database?
- You contradict yourself a little, saying there is no reason to reinvent the wheel, while suggesting a new reply (3. Be2). That's exactly reinventing the wheel.
- Then you say that even a relatively weak engine will burst the Latvian Gambit. This isn't relevant either, for people playing humans, rather than engines. And whatever engine you are talking about would burst the same player, no matter what opening he or she plays. So, following your logic, we simply shouldn't play chess. Maybe the Latvian Gambit isn't sound, but neither is this argument.
cigoL, coincidentally, my little Blackberry chess program plays Be2 against me. It always follows the same line:
1.e4 e5 2.Nf6 f5 3.Nxe5 Nc6 4.Nxc6 dxc6 5.Be2 Qh4 6.Nc3 fxe4 7.O-O and I proceed to bash my head like a seige ram against the castle walls for the rest of the game.
TheDestructivePawn wrote:
You're playing someone in a tournament. The game goes 1. e4 e5 2.Nf3 f5?! You stare perflexed at this phenomenal move. You cautiously analyze each option you have. Is this necessary? This is the question I have for you. Several people have suggested f5 as a dubious move. Being a Latvian gambit player myself, I tend to question this move. According to game explorer, 2... f5 wins more games than 2... Nc6 does for black. Only you can decide for yourself. To assist you are two games I found using the Latvian. I intend to settle this debate for once and for all.
P.S. And whether or not I should use the Latvian Gambit.
TheDestructivePawn wrote:
I think the Latvian Gambit is favorable for Black because in most cases it will put White on the defensive, break White's concentration and tempo allowing Black to gain the momentum.
You're playing someone in a tournament. The game goes 1. e4 e5 2.Nf3 f5?! You stare perflexed at this phenomenal move. You cautiously analyze each option you have. Is this necessary? This is the question I have for you. Several people have suggested f5 as a dubious move. Being a Latvian gambit player myself, I tend to question this move. According to game explorer, 2... f5 wins more games than 2... Nc6 does for black. Only you can decide for yourself. To assist you are two games I found using the Latvian. I intend to settle this debate for once and for all.
P.S. And whether or not I should use the Latvian Gambit.
Well let's see...
3.Be2 is of course not the best move, but playable, since it does not land white in a bad position. 3.Nxe5 is the best move for white, but if you can have fun with no risk, then why not...
Actually I've been playing chess since 1973, and I've met the Latvian just once in a game with regular time controls. Since I was bored, and the opponent was rather weak, I tried 1.e4 e5 2.Nf3 f5 3.ef5?! e4 4.Ng1!? which looks ridiculous, but it does have a point (actually it was a suggestion by the Caissiber chief, FM Stefan Buecker, who loves bizarre openings). Now Black plays the white part in a king's gambit accepted, but his pawn on e4 gives white a target, e.g. 3...Qg5 4.d3 Qxf5 5.de4 Qxe4+ 6.Be2 and then Nf3, 0-0 etc (the pawn on g2 is taboo if Black needs his Queen to stay on the board).
The opponent played badly and I had an easy (and boring) win, although objectively speaking after 4.Ng1 black should be able to equalize.
Ok, lets get a bit more serious.
@ wikipedian: I think that your concept is good, but you move the wrong knight at move three. Mr. Fraser would probably agree, as Black's game after 3...Nc6 4.d4! is simply horrible.
The only Latvian variation in the main line which is not a total disaster for Black, is 1.e4 e5 2.Nf3 f5? 3.Nxe5 Nf6!
White can still claim some advantage by just playing simple chess, e.g. 4.ef5 Qe7 5.Qe2 d6 5.Nc4 (or 5.Nf3) and although Black has a nice advantage in development, white has a pawn with no real weaknesses. Black is not lost, but he will be hard pressed to justify the pawn investment- and there are no direct fireworks.You can find a flawed, but quite thorough analysis in a Buecker article in chesspub: http://www.chesscafe.com/text/kaiss45.pdf
And... no, I will not care to tell you about the analysis holes- you can work it out yourself!
Or 4.d4 (the choice of the late GM Bela Perenyi, who was a very aggressive player- not very good choice here, IMO) 4...d6 (4...fe4!? is sane, and probably better) 5.d5!? (Perenyi played 6.Nc3 and won, but not because he had any opening advantage) when black is not equal yet. 5.Nc3 and 5.Be2 are not really frightening for Black, and finally, the "official" refutation:
5.Bc4
The move is certainly dangerous, but not a "refutation".
If interested, I can be a bit more helpful.
Here's a way to find out empirically: Latvian Gambit Tournament. Let's put this thing to the test!
I don't think Black is completely equal after 3...Nc6 as well. He is equal in the Jaenisch variation (1.e4 e5 2.Nf3 Nc6 3.Bb5 f5 4.d3 fe4 5.de4 Nf6 6.0-0 d6) although white can probe him with tricky moves like 7.Qd3/Qe2, but with his bishop on e2 white can try the ugly looking, but quite reasonable 7.Be3!? Bxe3 8.fe3 0-0 9.Ng5 Kh8 10.Nd5! with some initiative.
Black can only equalize after 3...Nc6 if white plays too ambitiously (4.ef5?! d5) or too routinely (4.Nc3?! Bb4!) or just for fun (4.d4!? is a Vienna Game main line, colors reversed). White has to opt for speedy development, and that is done with 4.g3 or 4.Be2. The former is not bad, but after an eventual ...fe4 by black the bishop on g2 would rather like to use some other diagonal. 4.Be2 is the most logical move, and Black has a little work to do to prove he's equal.