From's Gambit

From's Gambit, 1. f4 e5, is the most direct way to punish the Bird Opening's loosening of the kingside. It takes its name from the Danish player Martin Severin From (1828–1895). Black offers the e-pawn, aiming to recapture the time with rapid development and with the queen's access to h4. The obvious drawback is that White can sidestep the whole system with 2. e4, reaching the King's Gambit, so a Black player unwilling to face that opening may prefer another first reply. The careless 2. g4?? allows 2...Qh4#, a version of the Fool's Mate.

Taking the pawn with 2. fxe5 is the usual response. The main line continues 2...d6 3. exd6 Bxd6, leaving Black a pawn down but with the bishop aimed at g3 and h2 and the queen ready to join. White is obliged to play 4. Nf3, since 4. Nc3?? Qh4+ 5. g3 Qxg3+ 6. hxg3 Bxg3 is mate. Instead of recapturing on d6, Black can try 3...Nf6, the Langheld Gambit. White, for their part, can return the material with 2. fxe5 d6 3. Nf3 dxe5 4. e4, trading the f-pawn for Black's d-pawn; the half-open f-file may then serve White's attack after castling, though it can also leave the king exposed.

After 4. Nf3 Black chooses between several plans. The Lasker Variation, 4...g5, was introduced by Emanuel Lasker against Bird at Newcastle upon Tyne in 1892 and tries to dislodge the knight. IM Timothy Taylor, in his 2005 book on the Bird, regards it as dubious and recommends either the calm 5. d4 g4 6. Ne5, which leads to an early queen exchange with a small White plus, or the sharper 5. g3 g4 6. Nh4. The alternative 4...Nf6 threatens ...Ng4 followed by ...Nxh2. A further option is 4...Nh6, when 5. d4 is the Lipke Variation, while ...Nc6 offers a more positional treatment.

On move two Black can also play 2...Nc6, a rarely examined try, against which Taylor prefers 3. Nc3 Nxe5 4. d4 with e4 to follow, rather than 3. Nf3, met by 3...g5. White's other second moves, such as 2. Nf3, 2. e3 and 2. Nc3, the Bahr Gambit, leave the f-pawn exposed to capture. The engine gives White a slight edge after 1...e5, with 2. fxe5 its first choice, while Lichess results are almost exactly even.