Latvian Gambit

by Matt Grinberg, Founder of Alamogordo Chess Club

I was thinking about the Latvian Gambit recently.  I used to play it regularly and had good success against higher rated players.  Unfortunately it didn’t seem to work so well against lower rated players.

I think the problem for higher rated players was that they knew a little about it, but not as much as me.  So they would play the most aggressive moves for a while, but wouldn’t know how to follow up.  On the other hand, lower rated players just played natural moves and got a good position.  Eventually I gave up on it.

But, if you haven’t ever tried it, you will find it a fun opening.  It is worth trying, especially against higher rated players.  But first you have to have some idea of what you are doing.

The Latvian Gambit starts with 1. e4 e5, 2. Nf3 f5!?.  It is questionable whether the Latvian is really sound.  You don’t see Grandmasters playing it, but it is popular among many amateurs with an adventurous spirit, or who just want to take their opponent out of “normal” lines.  It is also popular in correspondence play.  The main strategic idea for Black is to plant pawns on d5 and e4, then use his dominating center and open lines to stifle White’s play and launch an eventual counter attack.  Black will frequently give up his g and h pawns or even his rook on h8 to lure White into wasting valuable time.  White’s strategy is to either prevent or undermine Black’s center and exploit the weakening of Black’s king side.  If White doesn’t get too caught up in trying to punish Black by winning material, he gets the better game.  But if White isn’t careful, he can often find himself blown away before he knows what hit him.

The intent of this primer is to give the non-Latvian player a brief over view.  I have borrowed heavily from five sources:

1. “Latvian Gambit,” Ken Smith and John Jacobs, Chess Digest, 1977

2. “ECO C, Second Edition,” Paul Keres (Aleksander Matanovic, Editor), Chess Informant, 1981

3. “Latvian Gambit,” Kon Grivainis, Thinkers Press, 1985

4. “Chess Assistant 16,” Convekta Ltd, 2018

5. “New Developments in the Latvian Gambit,” Chess Enterprises, 1998

Keep in mind that this is just an over view.  There are oodles (a technical term) of interesting side variations for both sides.  Let your imagination run wild!

1. e4 e5 2. Nf3 f5!? Black aggressively fights for the center, but is it worth weakening the kingside?  This is played about 10 times more frequently in correspondence chess than in over the board chess.  I guess correspondence players are more daring.

<At move 3 White has several choices. 3. Nxe5 This is the main line and White’s only serious attempt at refuting the Latvian.  The defense is unquestionably difficult, but Black has resources.  In spite of extensive analysis over the last century, no refutation is in sight.  But first, let’s look at the alternatives for White.
  1. 3. d3?! This move is safe but insipid.  It causes Black no problems and blocks in White’s king bishop.  Black equalizes easily. 3… Nc6 4. Nc3 Nf6 5. Bg5 Bb4 6. exf5 d5 7. a3 Bxc3 8. bxc3 Black is at worst equal;
  2. 3. d4!? Better than 3. d3 since it doesn’t block in the bishop.  However, with a little care, Black still has no trouble equalizing. 3… fxe4 4. Nxe5 Nf6 5. Bg5 Be7 6. Nc3 d6 7. Nc4 O-O 8. Ne3 Nc6=;
  3. 3. Nc3 A good solid flexible move which gives White a slight pull, but gives Black no great problems so long as he is prepared to play the Schliemann Defense. 3… Nc6 4. d4 (4. Bb5 Transposing to the Schliemann Defense of the Ruy Lopez. 4… fxe4 5. Nxe4 d5 6. Nxe5 dxe4 7. Nxc6 Qg5 Theory says this is slightly better for White.  I’ll take theory’s word for it) 4… fxe4 5. Nxe5 Nf6 6. Bg5 Bb4 White is slightly better;
  4. 3. exf5 This move is tricky, but Black holds his own if he knows what he is doing.  White wins material if Black follows theory, but White may end up regretting the time wasted. 3… e4 This is the main line of this variation. (Black can also play 3… d6 when White transposes to a line of Philidor’s Defense. 4. d4 e4 5. Ng5 Nf6 6. Ne6 Bxe6 7. fxe6 d5 8. c4! It will be difficult for Black to both hold his center together and win back the pawn) 4. Ne5 Nf6 5. Be2 Be7 (The old move is 5… d6 6. Bh5 Ke7 7. Nf7 Qe8 8. Nc3 (Or 8. Nxh8 Qxh5 9. Qxh5 Nxh5 10. g4 Nf6 The outcome hinges on the fate of the knight on h8) 8… Nxh5 9. Nd5 Kxf7 10. Qxh5 g6 11. fxg6 Kg7 12. Nxc7 Qc6 13. Nxa8 Keres claims this is slightly better for White, but I would say it is about as clear as mud) 6. Bh5 Kf8 7. Nf7 Qe8 8. Nxh8 Qxh5 9. Qxh5 Nxh5 Again the outcome hinges on the fate of the knight on h8;
  5. 3. Bc4 This at one time was considered a refutation of the Latvian.  Now theory just considers it unclear.  It’s hard to argue with the idea of just ignoring Black’s silliness and continuing development, but it’s not that simple.  By far the most common move here is 3… fxe4.  Moves such as 3… Nc6 and 3… d6 are playable, but allow White to get a good position with no trouble.  Deserving some attention, however, are 3… b5!? and 3… Nf6!? 3… fxe4
  1. 3… b5 The Strautins Variation – but it has two holes in it. 4. Bb3White continues with the policy of ignoring what Black is doing – hole number 1. (4. Bxb5?! White obligingly takes the bait. 4… fxe4 5. Nxe5 Qg5 6. d4 Qxg2 7. Rf1 Nf6 Black has a fine game; 4. Bxg8This is hole number 2.  White gets a great game with simple, straight forward play. Rxg8 5. d4 fxe4 6. Bg5 Be7 7. Nxe5! g6 (7… Bxg5? 8. Qh5 g6 (8… Ke7? 9. Qxg5 Ke8 10. Qh5 Ke7 11. Nc3 d5 12. Qf7 Kd6 13. Nxb5# 1-0, Tomson, Hans – Padula, Angel, corr. 1971)9. Qxh7 Rf8 10. Qxg6 Ke7 11. Qxg5 White wins) 8. Bxe7 Qxe7 9. Qe2 White regains the pawn with a good position) 4… fxe4 5. d3! Nf6 (5… exf3? 6. Qxf3 wins the rook on a8) 6. Nxe5 Black has a bad position with no counter play;
  2. 3… Nf6 The Morgado Variation.  Black intensifies the struggle for the center. 4. Nxe5 Qe7 5. d4 (Not 5. Nf7? Qxe4 6. Qe2 Qxe2 7. Kxe2 d5!) 5… Nc6 6. Nc3 (6. Nf7? again fails to 6… d5) 6… Nxe5 7. dxe5 Qxe5 8. O-O fxe4 White is better due to Black’s lack of development and the exposed position of the king and queen on the half open e-file)

4. Nxe5 d5! The Svendenborg Variation (I’m not kidding).  This is the ultimate Latvian line.  Both sides follow their basic strategic ideas without regard to what the other is doing. (4… Qg5!? There has been reams of analysis on this line without any conclusion!  In practice White has done well in this line, but who knows?  I’ll just give the main line and let you draw your own conclusions. 5. d4 Qxg2 6. Qh5 g6 7. Bf7 Kd8 8. Bxg6 Qxh1 9. Ke2 Black can either try to find a safe haven for his king with 9… c6, or throw caution to the wind with 9… Qxc1.  In both cases the position is unclear) 5. Qh5 g6 6. Nxg6 hxg6 White now has a choice of which thing to take. (Also good is 6… Nf6 7. Qe5 Be7 8. Bb5 c6 9. Nxe7 Qxe7 10. Qxe7 Kxe7 11. Be2 Black has more space, but White has an extra pawn and the bishop pair) 7. Qxg6 White eschews the rook, but keeps the initiative. (7. Qxh8 As so often is the case in the Latvian, White grabs the rook at the expense of ceding the initiative. 7… Kf7 Black threatens to entomb the White queen with 8… Bg7. 8. Qd4 Be6 9. Bb3 Nc6 10. Qe3 Bh6 11. f4 Nf6 Black’s strong center and better development give him compensation for the exchange and pawn) 7… Kd7 (7… Ke7 is bad because of 8. d4!) 8. Bxd5 Nf6 9. Nc3 Qe7 10. d3 exd3 11. Be3 We have another one of those “clear as mud” positions

3… Qf6 Certainly unorthodox, but it serves well both defensively (preventing 4. Qh5+) and offensively (attacks the knight).  The down side is it takes away the best square for the king knight and the Queen may be exposed to attack.  There are two main continuations.

[A dubious sideline is 3… Nc6?! This is simply an attempt to bait White into playing 4. Qh5+, but if White sidesteps it, he has no problems. 4. Qh5 White “falls” for the trap, but gets the better game anyway. (4. Nxc6 White prefers to keep things simple. 4… dxc6 5. d4 Qh4 6. exf5 Bxf5 7. Bd3 White is a pawn up with a good position) 4… g6 5. Nxg6 Nf6 6. Qh3 (Or 6. Qh4 Rg8 7. Nxf8 Rg4 8. Qh6 Rxe4 9. Be2 Qe7 10. Nc3 Rxe2 11. Nxe2 Nd4 12. O-O Nxe2 13. Kh1 d5 14. Nxh7 Nxh7 15. Qh5 Kf8 16. d3 Nxc1 17. Raxc1 White has a rook and two pawns for a bishop and knight, and his king is safe (unlike Black’s)) 6… fxe4 7. Nxh8 d5 In spite of the extra rook and pawn, White faces severe problems in terms of space and development.  The position favors White]

4. Nc4 White retains the option of playing the d-pawn to d3.  This is the line that tests Black most severely.

[4. d4 Seemingly obvious, but not best.  The problem is that White needs this pawn on d3 to challenge e4 after the inevitable … fxe4. 4… d6 (The immediate 4… fxe4 is unplayable because of 5. Bc4) 5. Nc4 fxe4 6. Nc3 The usual move.  White has two alternatives.

  1. First is 6. Ne3 This is an attempt to control d5, but moving this knight four times in the first six moves just can’t be right.  Black gets a satisfactory position by continuing his development 6… Nc6 7. Bb5 Bd7;
  2. Second is 6. Be2 preventing the normal …Qg6, but Black does well with either 6…Qd8 or 6… Nc6)

6… Qg6 7. f3 exf3 8. Qxf3 Nc6 9. Bd3 Qg4 Keres rates this position as equal]

4… fxe4 5. Nc3 This is the key position in the current theory of the Latvian.  There are four commonly played answers by Black. 5… Qf7 This is currently considered to be Black’s best try, but it is still not quite sufficient.

5… Qg6 This used to be considered the “only” move. 6. d3! Bb4 7. Bd2 Bxc3 8. Bxc3 Nf6 9. Bxf6 gxf6 10. dxe4 Qxe4 11. Ne3 Qb4 12. c3 Qxb2 13. Rc1 Black has an extra pawn, but his lack of development and exposed king give White the better game;

  1. 5… Qe6 Black attempts to prevent d3, but it works for only one move. 6. Ne3 c6 7. d3 d5 8. dxe4 dxe4 9. Bc4 Black is behind in development, his e-pawn is weak and his king and queen are exposed to attack;
  2. 5… c6 This seldom played move is better than the above two queen moves because it immediately forces the strategically important …d5. 6. Nxe4 White plays for an advantage. (6. Ne3?! White plays it safe. 6… d5 7. d3 exd3 8. Bxd3 Bd6 9. O-O Ne7 10. Ne2 O-O=) 6… Qe6 7. Qh5 Best. (7. Qe2?! d5 8. Ncd6 (8. Ned6!? Kd8 9. Nxb7 Kc7 10. Nba5 Qxe2 11. Bxe2 dxc4 12. Nxc4 In this unbalanced position, White has three pawns for a knight and better pawn structure, but Black has a freer position.  The chances are about equal) 8… Kd8 9. Ng5 Qxe2 10. Bxe2 Bxd6 11. Nf7 Kc7 12. Nxh8 Be6 Black will eventually win the knight on h8, but the time wasted will give White the advantage) 7… g6 8. Qe5 Qxe5 9. Nxe5 d5 10. Ng5 Nf6 11. d4 White has an extra pawn and Black has no compensation.

6. Ne3 c6 7. Nxe4 White simply takes the pawn and defies Black to find compensation.

[7. d3 This is another move that was once thought to be a refutation. 7… exd3 8. Bxd3 d5 9. O-O Bd6 (Black cannot play 9… d4?! because of 10. Bc4 Qd7 11. Qh5 Kd8(Or 11… g6 12. Qe5) 12. Rd1 White is better) 10. Re1 Ne7 Now if Black can castle, he has no problems.  So White plays one of two sacrifices. 11. Nc4 (11. Nexd5 cxd5 12. Nb5 O-O! By returning the piece, Black gets a good position. 13. Nxd6 Qxf2 14. Kh1 Bg4 15. Qd2 Qh4 16. Qe3 Nbc6 17. Bd2 Rf6 18. Qg5 Qxg5 19. Bxg5 Rxd6 20. Bxe7 Re6 21. Rxe6 Bxe6 22. Bd6 1/2-1/2, Stevens – Bianchi, Guillaume (FRA) 2113, France 1995) 11… dxc4 12. Bxc4 Bxh2 Black returns the piece for equality. 13. Kxh2 Qxc4 14. Bg5 O-O 15. Bxe7=]

7… d5 8. Ng5 Qf6 9. Nf3 Bd6 10. d4 Ne7 11. Bd3 Be6 12. O-O Nd7 Black has a freer position, but not enough for the pawn.

<image002.gif>

I note that there is no record of the Latvian Gambit being played by a World Champion.  However, there are three games known where it was played against a future World Champion – two against Jose Raul Capablanca and one against Bobby Fischer.  Capablanca split his two games and Fischer lost.

 If you want to give the Latvian Gambit a try, I’ll see you in the proverbial trenches!

Alamogordo Chess Club:  Plateau Espresso, 2724 Scenic Dr. Alamogordo; Monday from 4pm to whenever you feel like leaving.

Matt Grinberg, matt.grinberg@erols.com, (575) 415-3628

Leave a Reply

Your email address will not be published. Required fields are marked *