AMM wrote on 12/29/09 at 16:11:14:

Earlier in this thread I strangely forgot that 5...Qg6 was alive and kicking. My 2009 article on ChessCafe had mentioned your main line, it also gave the alternative 11...Kd8!? (instead of 11...Be6) which also seems playable for Black:
"In the endings after 11...Be6 12 Ne5 0-0 13 d4 Nc6 14 Bd3 Bf5 15 0-0 Bxd3 16 cxd3 or 11...Kd8!? 12 Ne5 Re8 13 d4 (13 f4 g5) 13...c5 14 c3 Nc6 15 f4 cxd4 16 cxd4 g5 17 fxg5 Qg7 18 Bb5 (18 0-0-0 Nxe5) 18...Nxd4 19 0-0-0 Qc7+ Black has active pieces and reasonable drawing chances."

From my ChessCafe article: 
"3...Nc6 is sometimes mentioned as an unusual alternative, but there is a clear refutation: 4 Qh5+ g6 5 Nxg6 Nf6 6 Qh3 hxg6 7 Qxh8 Qe7 8 Nc3! Nb4 (Perez Cruz – Acosta, corr. 1994; Massow Memorial) 9 d3 +-."

I agree with almost everything which Alejandro wrote above. It will be interesting to watch whether 5...Qg6 can be repaired. In the Dreibergs (5...Na6) and the 6...d5 Variation (Atars?!), my article wasn't very detailed. 5...Na6 offers interesting chances, if White doesn't find 6.a3!. However, this pawn move was quite frustrating for Black, so it would have been pointless to give all the details for the weaker replies to Na6. I also didn't really like the 6...d5 line.

Conquistador wrote on 12/28/09 at 15:30:33:


In December 2008 I wrote two "Godzilla" articles in a Spanish web ( respectively http://www.ajedreznd.com/visor/leton1.htm and http://www.ajedreznd.com/visor/leton2.htm) trying to save the Latvian on some lines very difficult to play with Black. During this year 2009 I have read with interest in this forum, in articles published by Stefan Buecker in www.chesscafe.com, and especially in USA forums how it has been analyzing very carefully some of the ideas that I myself suggested.

Just really these days I was working in an English translation of the article I wrote in Spanish ( abbreviated, since this one is enormous !, and focusing on the line 9...Bc5, former main line so as we know has been absolutely refuted ).

Looking for new ideas for the second player I think I have found new assessments to 5..Qg6 ( reading an old book by Hagen Tiemann of 1989 ), as well as the idea I already suggested in 5..Qf7 and 6..d5!?. I believe even recognizing Stefan is best player than myself  , he give few analysis of this last option, the ideas even being correct, barely it devotes two lines. As for the "old" 5..Qg6 I cannot explain at this time my ideas, on February 1 starts the Final of 6th World Ch.of the gambit ( by correspondence ), and clearly, I need to keep my secrets ... if not some player could read here. Indeed these players are among the most "updated" on the Gambit Theory nowadays, in addition, some of them are strong postal players ( They don't need to have computers, because they were already strong several years ago ! ).

There is a clear summary lately with the gambit. Theoretically we can try on refuting here in forums, but I can assure you in a real OTB ( wherever I usually play ) all this is very different; White player can to know something of Theory, even lines leading to the game 10-12, but if the Black player knows the lines more concrete ( as is my case ) it will come to the middlegame with a sufficient position, and the most important, he will know those positions. It is sure any player stronger than oneself always will suffer ( also it would happen if we play any other opening ), but I can assure you I have won several OTB players ranked between 2000-2200 ELO FIDE. I believe, up to 2300, problem is not the gambit, ... but the level of play !, although you equalizes after opening, ..you will lose at the middle game, .. or worse, in the Ending ! ( At least, for me, so usually I have been in 2100 ELO FIDE ). 

If someone has an interest, in a few days I will publish a "short" translation of the article I wrote in Spanish. About 5..Qf7; 6..d5!? is not necessary because it was commented extensively on this same forum ( see the link that says "Conquistador" ).

Relating on 3..Nf6, analyzed very extensively by Stefan this past summer ( in the link he says ), I think it's a great article, perhaps some analysis more specific is needed so I have doubts, but at least, in most lines White should have permanent initiative, although slight.

I still believe Black should play 3..Qf6, and also 4.d4 d6 5.Nc4 fxe4 is wherever Black has even greater problems than 4.Nc4

Finally, two ideas barely known: 3.Nxe5 Nc6 4.d4! (John Nunn ) 4..Qe7!? and 4..Qf6!?

Certainly Latvian is dying ( I have experience in official post and OTB games from 1978 ! ), but I doubt some day die ( perhaps only by postal, and revenging hard his death ).

In forums, is already buried ( But for me I am only interested in OTB ! )


.

Yeah, I did my own work on the 5...Na6 variation and it was not sufficient for black.
5...Na6 6.d3 Bb4 7.Bd2 exd3 (7...Qe6 8.dxe4 Nf6 9.Qe2; 7...Ne7 8.dxe4 0-0 9.Qe2 d6 10.a3 Nc6 11.e5 Qg6 12.0-0-0) 8.Bxd3 Ne7 (8...Qe6+ 9.Ne3) 9.0-0 0-0 10.a3
White has a pretty significant advantage here.  I think I remember you came to a similar conclusion.

Yes, I recall that thread (and commented in it).  It seems that much of this Latvian "theory" leads to positions where Black is just a pawn down, but then there is a comment like "if White is better, it cannot be by much" or "game X ended in a draw" or some such.

I dare say White should be satisfied with an extra pawn, a lead in development, a better structure and no weaknesses after 7 moves.