Is the Latvian Gambit sound?

I think the only possible explanations for the weird direction this thread has gone is a) masters don't play the LG frequently enough to understand which are main lines b) phren has read some theory book which has him totally gaslined or c) he's gaslining us himself. Who plays 3. ...Qe7? That's preposterous. If you're going to develop your girl on move 3, she's Elektra, not arm candy. Is that the Latvian Gambit corollary to the "mysterious rook move"? Black does not get a wide array of correct choices. 3. ...Qe7 is not on the menu. The kitchen is 86 3. ...Qe7. Could I interest you in some e4 and a side of "please move your knight" instead? I think know what pfren is going to play to that... Ng1 :)

There's a difference between an opening being sound and an opening being playable.

To me, "sound" means that with optimal play on both sides the person choosing the line gets what they want. For black in a gambit line that means that they need to achieve dynamic equality, where htey have full compensation for the pawn. I don't believe this is posible in the Latvian. 

To be "playable" means that you can get exceptable results using that opening that won't be worse than what you'd expect playing any other opening correctly at that player's rating. 

The Latvian is definitely playable. In a practical game the surprise value and lack of knowledge the average opponent will posses makes it rather formidable. But it's almost certainly not sound.

But below GM level, who cares if it's sound. Isn't the point to have fun? 

We may need to define what is meant by the word sound. The lower the grading of the opponent the more sound will become what would be unplayable against an expert player. The latest computer technology has thrown some so called sound sacrifices into question.

I suggest the following new definition of 'sound'.

A move can be said to be sound if it produces an advantage under the circumstances in which it is played.

I know this is wide open, but it is at least a sound definition and will enable all the arguments to be settled.

Those who wish to debate further could of course then take up the cudgle on  sound sound moves or unsound sound moves.

@ wikipedian: 3...Qe7 is not a good move, of course, but your 3...Nc6 is not any better, either.

It's not fun at all to play, too: Say that White obeys to your plan, and plays 3...Nc6 4.Qh5+ g6 5.Nxg6 Nf6 6.Qh4 Rg8, and now, instead of 7.Nxf8 white plays 7.e5. Now, no matter if you reply 7...Rxg6 or 7...Nxe5, you are a pawn down with zero counterplay, and you just pray that white does not know converting this to a full point. Seriously now, you couldn't have as main defence against 1.e4 an opening which might end up by force to this absolutely miserable position- could you?

And no, I have not read any book about the Latvian- I prefer either useful chess books, or Penthouse. I do not play that much lately (my current rating is 2364 FIDE), but still I can tell if a position is total crap, or not. I've not got chess Alzheimer yet.

gee, I dunno phren. I admit that white does have a better continuation from that line than Morgado-Elburg 1978 0-1. Maybe you are on to something, but for now the Frayzh still looks good to me.

I've read that the Latvian Gambit is actually more successful in correspondence games, than OTB games. If this is true, it can't be so horrible. The game I posted earlier was a correspondence game too. 

At the GM level, the Latvian has scored 25% in correspondence games, and 33.3% in OTB games at classical time controls, so I think actually it's the other way around.  Might do better in blitz though.

Where did you get those numbers, riga...? I read it in a book on the Latvian Gambit.

I keep a database of GM games for both correspondence and OTB.  The only black win is Apicella-Ivan Sokolov, European Team Chamionships 1992.  Other than that it's mostly white wins.

Nice! How many games in the database in total? How many games in the database with the Latvian Gambit? 

There are around 250 000 games between GMs up until 2010.  There are 2 correspondence games with the Latvian (both from a thematic tournament in the 1960's with Morgado) and 6 OTB games at classical time controls.

So only 8 Latvian Gambit games, out of 250,000! Not exactly a popular choice.  By any chance you could copy/paste the PGN's of those 8 games here? Thanks! 

pfren, if would be nice if you addressed the remarks I made to you in post #41. 

Nice! Thanks! Just copy/paste the PGN's of the last ones. No need to make fancy diagrams. 

You can snatch the PGN by clicking on PGN at the bottom right.