Chessy
Latvian Gambit (1.e4 e5 2.Nf3 f5)
Black counter-attacks immediately with 2...f5, offering a pawn to seize the initiative — or so the plan goes. Engines are brutal about the verdict here. But over the board it creates messes that White often fumbles. See how long you can hold the attack in the drill below.
Play the Latvian Gambit against the engine
Free, no signup — you play black, the engine adapts to your level.
You just felt what a +1.45 advantage feels like from both sides of the board. Create a free Chessy account and let AI coaching show you the exact move where Black's attack turns from dangerous to desperate.
Create a free account →The engine is not kind: White is up +1.45
The Latvian is Black's gambit — Black sacrifices the f5-pawn and weakens the kingside to get counterplay. The FEN shows it's White to move, meaning it's White who responds to the challenge. Stockfish at depth 16 gives White +1.45 — nearly one and a half pawns in White's favour. That is a substantial objective advantage, one of the steepest on this list. Black is gambling hard.
How White should punish it
Stockfish's recommendation for White is 3.Nxe5 — the most direct refutation. After 3.Nxe5 Qf6 4.Nc4 fxe4, White returns the knight but keeps the structural advantage. The practical data backs this up: 3.Nxe5 yields White 52.2% across 636k games. Common mistakes by White actually bail Black out:
- 3.exf5 — inaccuracy, 66 cp loss for White vs Nxe5; White only 48.2% (below average!)
- 3.d3 — mistake, 127 cp loss; drops White advantage significantly
- 3.Bc4 — inaccuracy, 87 cp loss
Why Black keeps playing it anyway
Despite the −1.45 deficit, Black scores 47.1% across 3.2 million games — far better than the engine number suggests. White routinely chooses 3.exf5 (the most popular reply, 1.4 million games), which is the inaccuracy that gives Black a fighting game. When your opponent helps you out of your own gambit mistake, the scoreboard flatters the opening. Take 3.Nxe5 away and Black's real win rate craters.
Playing the Latvian as Black: know the trade-off
You are willingly handing White a +1.45 position. In exchange, the game becomes immediate hand-to-hand fighting with no safe quiet moves. At blitz or in rapid online games against opponents who choose 3.exf5 out of habit, the Latvian absolutely creates chaos. At classical time controls against a prepared player who reaches for 3.Nxe5, expect to be fighting uphill for the whole game.
Results across 3,262,727 Lichess games
| Most-played continuation | Games | White wins |
|---|---|---|
| exf5 | 1,424,594 | 48.2% |
| Nxe5 | 636,153 | 52.2% |
| d3 | 410,880 | 49.5% |
| Nc3 | 372,494 | 50.7% |
| Bc4 | 211,007 | 53.0% |
| d4 | 143,697 | 52.4% |
Frequently asked questions
Is the Latvian Gambit sound?
No. Stockfish at depth 16 evaluates the position at +1.45 in White's favour after 2...f5 — that's a serious objective disadvantage for Black. It's an unsound gambit kept alive by White's practical mistakes.
What is the best reply to the Latvian Gambit?
3.Nxe5, which Stockfish recommends and which scores 52.2% for White across 636k Lichess games. The most popular move, 3.exf5, is an inaccuracy — it only scores 48.2% for White, below the overall average.
Why does Black score nearly 47% with the Latvian if it's so bad?
Mostly because White plays 3.exf5 instead of 3.Nxe5. Over 1.4 million Lichess games use 3.exf5, the inaccuracy that lets Black back into the game. When White finds 3.Nxe5 correctly, Black's practical results drop sharply.
Is the Latvian Gambit good in blitz?
It has a higher practical ceiling in blitz because White tends to grab on f5 out of habit. You get double-edged positions where time pressure helps Black. At longer time controls, an opponent who knows 3.Nxe5 will convert the +1.45 advantage routinely.
How many games feature the Latvian Gambit?
Over 3 million Lichess games have reached the Latvian Gambit position. White wins 49.9%, Black wins 47.1%, with 3.1% draws — based on real rated games.