Square One
    Opening Walkthrough
    8
    7
    6
    5
    4
    3
    2
    1a
    b
    c
    d
    e
    f
    g
    h
    Opening Coach

    Danish Gambit

    C21You play white
    1/9
    Your move

    Stake the center and open the diagonal your light bishop will live on — c4 to f7 is the whole plan.

    Your line9 to go

    Play the moves — your line builds here as you go.

    The Danish Gambit

    Two pawns, gone by move five, for two bishops aimed at his castled king before he has one. The engine says you are worse. The scoreboard says 58.3%.

    Most gambits buy a tempo. This one buys a diagonal — two of them. You hand over the d- and b-pawns so a bishop lands on c4 pointing at f7 and a bishop lands on b2 pointing at g7, and neither of them ever has to move again. Here is the honest scorecard, because you deserve it before you play this: Stockfish has White at -0.62 in the position after 5.Bxb2, and -1.21 once Black plays the best defence. It is not sound. Now the other number — across 3.54 million games at 1000-1600, White scores 58.3% from that same losing position, against 53.0% for the cautious version that keeps the b-pawn. That gap is the entire product. The Danish loses to precision and eats natural moves alive, and every branch below is a natural move.

    The line, move by move

    1. e4 e5 2. d4 exd4 3. c3 dxc3 4. Bc4 cxb2 5. Bxb2 Bb4+ 6. Nc3 Bxc3+ 7. Bxc3

    1. 1. e4 (you) Standard. Nothing to see yet.

    2. 1... e5 (opponent) 51% mirror you, and that is the move that lets the Danish exist. From here it is three moves away and not one of them looks like an attack.

    3. 2. d4 (you) d4! Offer the centre pawn immediately. This is rare — 62% play Nf3 — so most opponents are out of book on move two rather than move twelve.

    4. 2... exd4 (opponent) 63% take it. Nobody declines a free centre pawn at this rating, and you want them to take it.

    5. 3. c3 (you) c3!! Not Qxd4 — 49% play that and it is a duller opening with your queen out early. Instead you offer another pawn, and the point is what opens up behind it: c1 to b2, and the whole a1-h8 diagonal.

    6. 3... dxc3 (opponent) 74% take. The 12% who play …Nc6 instead let you answer cxd4 with a perfect two-pawn centre — the engine has that at +0.89, so their caution costs them more than their greed would have.

    7. 4. Bc4 (you) Bc4! 79% of Danish players find this. You do NOT recapture on c3 yet. The bishop takes aim at f7 first and dares him to take on b2. Every declining move — …Nc6, …Nf6, …d6 — measures between -0.2 and 0.0 after you simply play Nxc3: a normal game, one pawn down, miles ahead in development.

    8. 4... cxb2 (opponent) Half of them take. His pawn is now on b2 attacking your rook, which is exactly where you wanted it — it has walked onto the square your bishop wants.

    9. 5. Bxb2 (you) Bxb2. 99% play this and it is the whole point of the opening: c4 stares at f7, b2 stares at g7, and Black has developed nothing. Be honest with yourself about the price — the engine has you at -0.62 here. You have bought a position, not an advantage.

    10. 5... Bb4+ (opponent) …Bb4+ — the most popular move and the best one. The check gains him a tempo and, more importantly, it stops Bxg7 dead: you have to answer the check first.

    11. 6. Nc3 (you) Nc3. 56% block here: the knight develops while it blocks, and it covers d5. (Nd2 is 35% and measures within a tenth of a pawn; Kf1 at 6% is a shade better still on the engine and costs you castling forever — not a trade worth making at this level.) Now the honest fork in his road.

    12. 6... Bxc3+ (opponent) …Bxc3+ — 21% cash in immediately. This is the mistake, and it is the natural-looking one: he trades a developed bishop to break up your pawns and hands back one of the two bishops you paid two pawns for. The 30% who play …Nf6 instead keep everything and the engine has you at -1.21 — that number is real, and it is why this branch is a defence, not a trap.

    13. 7. Bxc3 (you) Bxc3! Take with the BISHOP, not the b-pawn (100% of the board agrees). The dark bishop simply re-routes from b2 to c3 and keeps the same diagonal. He has spent his best piece to achieve nothing.

    Every branch cashes out

    7 coached branches in this family. Engine-checked finishes:

    • cxb2 / Bb4+Sound line

      +0.5 — you are two pawns down and better, because he traded off the defender of his dark squares and still has a knight on b8. The scoreboard number for this whole structure is 58.3%. It is not a trap; it is the reason the trap keeps getting the chance to happen.

    • cxb2 / Nf6Sound line

      +0.8 — you gave a bishop for a pawn and got a knight back, and his king is sitting on f7 with a White pawn on f6. The other half walk the king instead: …Ke8 (24%) leaves him at -0.6, …Kg8 (19%) at -0.9. Half of them block. That coin gets flipped every single game.

    • cxb2 / d6 / dxe5+5.5 by force

      +5.5 on move nine, and every move he played was the natural one: develop the bishop, recapture the sacrifice, take the checking knight. Three natural moves, one queen.

    • cxb2 / d6 / Ke8Sound line

      +1.6 — a clean pawn up, his king stuck on e8 with no castling rights forever, and your b2 bishop still pointing at g7 through an empty board. The good half of his choice still costs him the game plan.

    • cxb2 / Bc5 / Qh4+8.1 by force

      +8.1 on move eight — the Two-Bishop Reckoning. He developed a bishop, brought his queen out with a mate threat, and developed a knight with tempo. Three good-looking moves. The two bishops you paid two pawns for ate all three.

    • cxb2 / Bc5 / Bxf2++8.3 by force

      +8.3 for the price of walking your king to f2. The other queen sorties — …Qg5 at 13% and …Qe7 at 8% — end the same way one move slower: you take on h8 and he has nothing.

    • d6Sound line

      +0.5 while still a pawn down — you gave two pawns, took one back, and turned the last one into a hole in front of his king. Declining the Danish does not make it go away; it just makes the damage structural instead of tactical.

    Why it works

    The Danish does not buy a tempo, it buys two diagonals. Two pawns go on the table so one bishop lands on c4 pointing at f7 and another lands on b2 pointing at g7, and neither ever has to move again. The honest ledger comes first: Stockfish has White at -0.62 after 5.Bxb2 and -1.21 against the best defence. You have bought a position, not an advantage. The scoreboard disagrees: across 3.54 million games at 1000-1600, White scores 58.3 percent from that objectively losing position, against 53.0 percent for the cautious version that keeps the b-pawn. That five-point gap is the entire argument for playing it.

    The gap exists because the Danish eats natural moves alive. ...Bc5 is the single most natural developing move on the board, and it un-guards g7: Bxg7 measures plus 5.5, White scores 72 percent from that exact position, and only 3 percent of opponents see it coming. ...Bg4 in the d6 line parks a bishop on a square with no defender and meets Bxf7+ followed by Ne5+, a knight check that also unmasks the d-file. The 43 percent who take the knight discover that the d6 pawn was the only thing between the queens, and Qxd8 lands on move nine. Three natural moves, one queen.

    Even Black's counterplay is mapped. When Bxg7 lands and Black answers with ...Qh4 threatening mate on f2, the 1 percent of White players who grab the h8 rook get mated on the spot. The tree's answer is Qf3, defending f2 while the rook stays on the menu, and the branch ends plus 8.1: the Two-Bishop Reckoning, where the two diagonals you paid two pawns for eat three good-looking moves in a row.

    How to punish the Danish Gambit

    The refutation sits right on the first fork and it is not exotic: 5...Bb4+. The check gains a tempo and, more importantly, stops Bxg7 dead, because White has to answer the check before the bishop can eat anything. Then keep the bishop: 6...Nf6 develops, holds both extra pawns, and leaves White at -1.21 with nothing but open lines to show for them. The natural-looking 6...Bxc3+ is the mistake, played by 21 percent: it trades your best piece to break up pawns White is happy to repair with the bishop recapture, and he comes out at plus 0.5 despite being two pawns down.

    One tripwire worth carrying in the other direction: if the g7 reckoning does land, ...Qh4 is a real counter-threat. A White player who grabs the rook with Bxh8 gets checkmated by Qxf2, supported by the c5 bishop, and 98 percent of the 57 humans who reached that position found the mate. It does not survive the correct Qf3, but it punishes the autopilot rook grab every single time.

    How to avoid falling into it

    Do not conclude that declining is the safe road, because the tree's numbers say the opposite. Meeting c3 with ...Nc6 instead of taking lets White recapture into a perfect two-pawn centre at plus 0.89, worse than what greed would have cost. Declining the second pawn with ...d6 concedes a normal game where White is a pawn down but miles ahead in development, and the tempting ...Be6 trade ends with a permanent hole on e6 in front of your king. Accepting is objectively fine at -0.62. It is what happens afterwards that decides everything.

    The rules that keep you alive: check on b4 before developing anything else, follow with ...Nf6, and keep the dark-squared bishop instead of cashing it on c3. Never put a piece where the c4 or b2 bishop turns a capture into a king hunt: the squares to fear are f7 and g7, and the habit to break is collecting material while your king is still standing on e8.

    Questions people actually ask

    Is the Danish Gambit sound?
    No. Stockfish depth 18 has White at -0.62 the moment both bishops are posted, and -1.21 against the best defence. It is a practical weapon, not a correct one: across 3.54 million club games the accepted Danish scores 58.3 percent for White, because the losing side of the eval is also the side whose natural moves keep walking into the two bishops.
    What is the refutation of the Danish Gambit?
    Check first, then develop: 5...Bb4+ forces White to deal with the check before Bxg7 exists, and 6...Nf6 keeps both extra pawns with the engine at -1.21. The key is refusing the natural 6...Bxc3+, which hands back the dark-squared bishop and turns a winning defence into a worse position.
    What if Black declines the Danish Gambit?
    Declining scores worse than accepting in this data. Refusing the first pawn with ...Nc6 lets White build a two-pawn centre measured at plus 0.89. Declining the second with ...d6 gets a normal game where White recaptures on c3 and stands even at 0.00 with a big development lead. The gambit does not go away when refused; the damage just becomes structural instead of tactical.
    Does the famous Bxg7 Qf6 trap line actually happen?
    No, and this tree exists partly because it does not. After 6.Bxg7 the real board plays 6...Qh4, chosen 39 percent of the time; ...Qf6 does not appear in the top nine replies. Worse, the folklore follow-up of grabbing the rook is checkmate for Black: 7.Bxh8 Qxf2 is mate, found by 98 percent of the players who reached it. The move that works is 7.Qf3.
    At what rating does the Danish Gambit work?
    The 58.3 percent score is measured at 1000-1600 blitz and rapid, where half of all opponents take the b2 pawn and the punished branches fire constantly. The gambit loses to precision, and precision is exactly what climbs with rating: the higher you go, the more often you meet the Bb4+ and Nf6 defence and play a real game from -1.21.

    Drill the Danish Gambit against a real opponent

    The trainer above plays every branch with a coach explaining each move. When you are ready, take it into a live game, or paste your Chess.com or Lichess username and see which openings actually cost you rating.

    Play it for realGet your free scout report

    More opening weapons

    New to openings? Learn openings without memorizing explains the approach behind every walkthrough here, and the free AI chess coach guide covers what the rest of the app does without a card.