Latvian Gambit
Quick reference
1.e4 e5 2.Nf3 f5 counterattacks the knight on f3 rather than defending the e5 pawn, and it's one of the sharpest ways to meet 1.e4 e5 2.Nf3 that anyone still plays with any regularity. The name comes from correspondence analysis carried out in Latvia in the early twentieth century, and the line has kept a small but devoted following ever since, mostly among players who like forcing, tactical positions over quiet development.
Why it's considered dubious
The straightforward objection is that 2...f5 weakens the e6 and g6 squares and the long diagonal in front of Black's own king before a single piece has been developed, in exchange for attacking a knight that can simply move. White's strongest and most tested response, 3.Nxe5, just takes the offered pawn — and years of correspondence and engine analysis haven't found full compensation for Black in the resulting positions. It survives as a serious try mainly because the follow-up theory is sharp enough that an unprepared White player can go wrong quickly.
The line that actually punishes it
After 3.Nxe5 Qf6, attacking the knight and eyeing f2, the accurate continuation is 4.d4, defending the knight with a pawn rather than moving it. Black's most natural try, 4...d6, kicks the knight again, and 5.Nc4 retreats with a gain of tempo since the black queen or a future ...fxe4 push both cost Black additional time to untangle — the point being that White's development stays ahead throughout, even while temporarily fielding fewer developed pieces than the sharpness of the position suggests.
Defend the knight with a pawn, not a retreat
Don't retreat the knight passively after ...Qf6 — 4.d4 is the move that keeps the extra pawn defensible. A knight retreat instead (like 4.Nf3) gives back the tempo Black is trying to win with the queen sortie and hands back most of the practical difficulty for nothing.
For the standard, fully respected way to meet 2.Nf3 in this same opening family, see the Ruy Lopez, reached instead through the normal 2...Nc6 3.Bb5.
Starting position. Latvian Gambit — 3.Nxe5 and White's development lead.
3.Nxe5 Qf6 4.d4 d6 5.Nc4 and White's development lead.
Frequently Asked Questions
Is the Latvian Gambit ever seen in serious tournament chess?
It appears occasionally at the club and correspondence level and very rarely in stronger tournaments, usually from players who have specifically dedicated years to its theory. It has never become part of a top grandmaster's regular repertoire against 1.e4 e5 2.Nf3.
What if White plays something other than 3.Nxe5?
3.exf5 and 3.Bc4 are the other common tries; both are considered less testing than simply taking the pawn on e5, since they let Black recapture or develop with less immediate cost than facing 3.Nxe5 and the accurate follow-up.
Why do the resulting positions favor White so consistently?
Black's whole compensation is based on quick, forcing piece activity before White's extra pawn and safer king start to matter — once White navigates the first several forcing moves accurately, as with 4.d4, the position settles into a normal material advantage without much remaining danger.