The word "gambit" traces back to the Italian gambetto, the act of tripping an opponent by the leg — no physical contact involved at the chessboard, of course, but the spirit carries over. A gambit is a deliberate sacrifice of material, almost always a pawn, in exchange for a positional edge — faster development, extra central control, or an exposed enemy king — and occasionally for an outright quick checkmate as early as move seven or eight. Gambits can be accepted (the material is taken) or declined (it is left alone), and the two choices often lead to completely different games.
Not every gambit carries the same risk. Some are "gambits" only in name, giving up little real danger to either side. Others are extremely sharp: if the tricks behind them do not work — because the opponent knows how to defend — the gambit player can end up simply down material with nothing to show for it. Because of that risk profile, gambits work best as an occasional weapon rather than the backbone of your entire repertoire — Rozman's own comparison is to dessert: a treat you enjoy now and then, not something your whole chess diet should be built on.
Roughly 90 percent of White gambits begin with 1.e4, since it is the most "open" first move, freeing both the queen and a bishop at once. Against the symmetrical reply 1...e5, one of the oldest and most famous tries is the King's Gambit, 2.f4:
The King's Gambit dominated the 19th-century "Romantic era" of chess, when flashy material sacrifices for fast attacks were prized above caution. Its most famous outing may be the "Immortal Game," played in London in 1851 between Adolf Anderssen and Lionel Kieseritzky — Anderssen sacrificed both rooks, a bishop, and his queen, ending up an almost unthinkable 21 points of material down, yet still delivered checkmate. Engines have since shown that accurate defense equalizes for Black — but at club level, accurate defense is rare, which keeps this gambit dangerous in practice.
A second White try in symmetrical e4-e5 positions attacks from the other direction: the Danish Gambit, offering not one but two pawns for a devastating pair of bishops.
In sharp gambits like the Danish, material is temporary and development is not: every extra pawn Black grabs costs a full move, while White's pieces come out with each recapture skipped. Reported online results put White's win rate here at around 60 percent — the highest of any common White club-level gambit.
Against the Sicilian Defense (1...c5), White has a gambit option too: 2.d4, offering the same kind of rapid development for a pawn. If Black accepts, 3.c3 offers a second pawn to open the position further; if that is also taken, White recaptures with the knight. This is the Smith-Morra Gambit:
Some gambits are genuinely dangerous to accept — the Vienna Gambit from Chapter 10 is one of them, and Black is generally better off declining it rather than grabbing the offered pawn, since White's compensation comes on fast. Not every gambit is that risky, though. The Queen's Gambit, 1.d4 d5 2.c4 — yes, the same one the Netflix series borrowed its title from — is barely a gambit in practice: it is one of the most popular moves in the entire position, and if Black accepts with 2...dxc4, White simply continues with 3.e4, seizing the full center while the light-squared bishop already eyes the pawn on c4.
The Queen's Gambit has survived centuries of scrutiny, including the arrival of chess engines, without being refuted — proof that a "gambit" can also just be a sound, long-term opening choice.
Gambits are riskier to play with Black than with White: moving second is already a small disadvantage, and adding a material deficit on top of it raises the stakes further. Black gambits tend to be all-or-nothing — they either win quickly or leave Black in real trouble. The first example, the Blackburne Shilling Gambit, is a pure trap out of the Italian Game, a position reached over 200 million times online. From the common Italian Game position (1.e4 e5 2.Nf3 Nc6 3.Bc4), Black offers a knight with 3...Nd4 — played only about 500,000 times out of the roughly 9 million games that reach this position (around 5 percent). Rozman compares the whole idea to a poker bluff: a weak hand played as if it were a strong one. White has two natural-looking ways to answer it, and only one of them is actually safe — see if you can work out which before revealing the line.
This trap only works if White walks into every step. The correct response, simply capturing the offered knight on move four, refutes the whole idea and leaves Black clearly worse. Treat it as an occasional surprise weapon at faster time controls, not a sound opening.
According to the Lichess database, this exact mating sequence has worked over 400,000 times — including, in bullet (one-minute) chess, against players rated as high as 2800. At the more ordinary average rating of 1600, roughly 18,000 recorded games have followed this precise script. The bluff clearly does get called in the real world — but as your own rating climbs, expect opponents to start finding 4.Nxd4 far more often.
The Englund Gambit answers White's Queen's Pawn Opening with an immediate 1...e5, offering a pawn from the very first move. White is not obliged to fall into anything by capturing it, but a specific sequence of natural-looking White replies walks straight into a forced mate. After 2.dxe5 Nc6 3.Nf3 Qe7, Black adds attackers to the e5 pawn while White defends it with 4.Bf4; then 4...Qb4+ forks the king, the f4 bishop, and the b2 pawn, and 5.Bd2 Qxb2 grabs the pawn while eyeing the corner rook. On move six, White has two natural-looking tries — see if you can spot which one walks straight into a forced mate before revealing the line.
Even world-class players have tried the Englund at the board: in 2019, Grandmaster Aman Hambleton — reportedly the highest-rated player ever to try it in serious competition — played it against fellow GM Steven Zierk at a major tournament. Hambleton lost, but the game dragged on for 114 moves, showing the resulting position was far from hopeless. Like the Blackburne Shilling Gambit, this line depends on White choosing the natural-looking moves rather than the objectively best ones (here, 6.Nc3 instead of 6.Bc3 avoids the whole trap).
Objectively, the Englund Gambit is unsound — it needs a whole string of natural-but-wrong White moves (accepting the pawn, defending with Bf4, walking into Bc3 instead of Nc3) to reach the final mate, and any single accurate move ends the idea. Yet the same databases show that exact checkmate has been reached more than 100,000 times online. Unsound in theory and effective in practice are two different things — worth knowing as both Black and White.
Both Black gambits in this chapter share the same lesson: they can knock out an unprepared opponent in seven or eight moves, but a well-prepared White player who finds the accurate response early enough refutes them and leaves Black in a difficult position. Use them sparingly and know the refutation lines yourself, since you may face these same tricks as White. If you want to keep exploring Black gambits beyond these two, the Portuguese and the Budapest are worth a look — both can leave Black with a reasonable game even when the sharpest tricks don't land.
A gambit trades material, usually a pawn, for development, central control, or attacking chances, and can be accepted or declined by the opponent — so a gambit player still needs a plan for a "normal" game if the offer is refused. With White, gambits such as the King's, Danish, and Smith-Morra accelerate development and central control, usually earning the pawn back or more through a lasting initiative; the Queen's Gambit is barely a gambit at all and stands on its own as a sound, long-standing main opening. With Black, the risk-to-reward balance is steeper: tricks like the Blackburne Shilling and Englund Gambits can deliver forced mates in seven or eight moves against natural-looking play, but fail outright against an opponent who finds the best defense. Sprinkle gambits into your repertoire for variety and surprise value, not as your only way of playing chess.
A gambit is a deliberate sacrifice of material, usually a pawn, in exchange for a positional or attacking advantage. The opponent can accept it (take the material) or decline it (leave it alone), and each choice leads to a different kind of game.
An open f-file and an open diagonal toward Black's king, plus faster development — for example, 3.Nf3 both develops a piece and stops the annoying ...Qh4+ check.
Every pawn Black spends time capturing is a move Black is not using to develop. White uses those extra tempi to bring out both bishops actively, only recapturing (with Bxb2) once the second bishop is actually needed on that square.
It arises after 1.e4 c5 2.d4 cxd4 3.c3 dxc3 4.Nxc3 against the Sicilian Defense. White gives up a pawn for open lines, faster development, and a lead in space, typically reaching a comfortable position with an easy lead in piece activity.
The gambit move, 2.c4, is simply the most popular and natural continuation in that position rather than a genuine risk. If Black accepts with 2...dxc4, White is not in danger — 3.e4 claims the full center and Black's extra pawn on c4 is more of a liability than a prize.
Nxe5 grabs a pawn but walks into a forced sequence that ends in checkmate on Black's seventh move. The correct response to 3...Nd4 is simply 4.Nxd4, capturing the offered knight, which leaves White clearly better with no risk.
The bishop on e2 is pinned along the e-file: it sits between the Black queen on e4 and the White king on e1. Moving it to capture the knight would expose the king to the queen's check, which is illegal.
6.Bc3, attacking the Black queen, allows 6...Bb4 pinning the bishop and leads to mate a few moves later. The safer move is 6.Nc3, which defends without walking into the pin.
It comes from the Italian gambetto, the act of tripping someone by the leg — a fitting image for a move designed to make an opponent stumble, even though nothing physical happens over the board.
Adolf Anderssen sacrificed both rooks, a bishop, and his queen against Lionel Kieseritzky — ending up about 21 points of material down — yet still delivered checkmate. It began with a King's Gambit and remains one of the most famous illustrations of the Romantic era's love of flashy sacrifice over material.
No — he calls it objectively unsound, since it only works if White plays a whole series of natural-looking but incorrect moves, and any single accurate move refutes it. Even so, the exact final mating position has occurred more than 100,000 times in online games, which is why it's still worth knowing.
Over 400,000 times according to Lichess, including wins against players rated as high as 2800 in bullet, and roughly 18,000 times at an average rating of 1600 — evidence the "bad bluff" gets called far less often than the evaluation suggests.