Blackburne-Shilling Gambit
Black's 3...Nd4 sets a smothered-mate trap. See why 4.Nxe5 still holds, why 5.Nxf7 loses on the spot, and how 4.Nxd4 refutes the gambit outright.
Opening Moves
1. e4 e5 2. Nf3 Nc6 3. Bc4 Nd4
Name the move that actually loses the game for White, and choose between the two legal king moves as Black with a reason rather than a guess.
After 1.e4 e5 2.Nf3 Nc6 3.Bc4 Nd4, Black has moved one knight twice and left the pawn on e5 undefended. Both of those look like beginner mistakes. The second one is the bait, and the line is called the Blackburne-Shilling Gambit, or the Blackburne Shilling Gambit without the hyphen, depending on who is printing it.
Black has just played 3...Nd4. The knight attacks f3 and eyes the c2 fork square, e5 is hanging, and White's bishop is already pointed at f7. — r1bqkbnr/pppp1ppp/8/4p3/2BnP3/5N2/PPPP1PPP/RNBQK2R w KQkq - 4 4
Useful next actions:
The Mate At The End Of The Trap
The pawn on e5 is genuinely free for one move. Take it with 4.Nxe5 and Black answers 4...Qg5, hitting the knight on e5 and the pawn on g2 in the same breath.
Now comes the move the whole line is built to invite. 5.Nxf7 forks the queen on g5 and the rook on h8, so it looks like the point where White wins the argument. Black ignores both.
The finish is 5...Qxg2 6.Rf1 Qxe4+ 7.Be2 Nf3#, a smothered mate in the middle of the opening.
Checkmate. Every square around the white king is taken by a white unit, and the one piece that attacks the knight on f3, the bishop on e2, is pinned by the queen on e4. — r1b1kbnr/pppp1Npp/8/8/4q3/5n2/PPPPBP1P/RNBQKR2 w Qkq - 2 8
Two details in that position are worth more than the mate itself.
The bishop on e2 attacks f3 and cannot capture, because the black queen on e4 pins it to the king. That pin is why 6...Qxe4+ is a check worth giving before the knight comes in.
And 7.Be2 was not forced. After 6...Qxe4+ White has exactly two legal moves. The other one, 7.Qe2, avoids the mate and loses anyway. The mate is the memorable ending, not the reason White is lost.
Taking The Pawn Is Not The Mistake
Most pages written about this gambit treat 4.Nxe5 as a blunder. It is not one. It is an inaccuracy, and the difference matters, because the move that actually loses the game comes one move later.
After 4.Nxe5 Qg5 the queen hits the knight and the g2-pawn at once. White is worse here by about half a pawn, which is a price, not a catastrophe. — r1b1kbnr/pppp1ppp/8/4N1q1/2BnP3/8/PPPP1PPP/RNBQK2R w KQkq - 1 5
From that diagram, White has two sensible moves. 5.O-O Qxe5 6.c3 Ne6 7.d4 lets the knight go and takes the centre in exchange, and keeps the price at the half pawn 4.Nxe5 already paid. White is a piece down for a pawn, with pawns on d4 and e4, a lead in development, and a black queen that the d4-pawn hits the moment it arrives.
5.Bxf7+ costs a little more and is the move Steinitz recommended in 1895, writing that the check followed by castling was White's best chance, with two pawns and the attack for the piece.
Against those, 5.Nxf7 simply loses. That is the shape of the whole line: the pawn grab on move four is affordable, and the fork on move five is not.
Even after 5.Nxf7 Qxg2, White is not obliged to be mated. 6.d3 walks the king out of the box, and 6...Qxh1+ 7.Kd2 Qxd1+ 8.Kxd1 leaves Black a rook up for a pawn. That is the honest way to describe the trap: 5.Nxf7 loses the game, and 6.Rf1 decides whether it also ends in a picture.
One Square Decides The Rest Of The Game
Say White has taken the pawn, met 4...Qg5, and found 5.Bxf7+. Black cannot take the bishop, because the knight on e5 guards f7. There are exactly two legal moves on the board: 5...Ke7 and 5...Kd8.
They point in opposite directions.
After 5...Kd8 6.O-O Qxe5 7.c3 Ne6, Black gave back nothing, the king is inconvenienced rather than exposed, and the extra piece starts to tell.
Now the version where the king stepped to e7 instead.
5...Ke7 6.O-O Qxe5 7.Bxg8 Rxg8 8.c3 Ne6. The balance is the one from the Kd8 line, two pawns for a piece, with a pair of minor pieces traded off. This time it is White who stands better. — r1b2br1/ppppk1pp/4n3/4q3/4P3/2P5/PP1P1PPP/RNBQ1RK1 w - - 1 9
Two pawns for a piece in both lines. Opposite verdict. The king on e7 blocks its own bishop, which is what allows 7.Bxg8, and it stands on the file its own queen is using, so White's d4 and f4 pushes arrive with a target to gain time against. Black usually ends up walking the king across to d8 several moves later anyway, by which point White's centre pawns are already rolling.
There is a practical point in that too. 6.O-O is close to forced for White after 5...Ke7: every other sixth move hands the better game straight back to Black. Both sides have exactly one move here that matters, and both are easy to get wrong at speed.
The Move That Ends The Argument
None of the above is necessary. White can decline the whole conversation on move four.
After 4.Nxd4 exd4 5.O-O, the knight that did all of Black's work is gone and the recapture has left doubled d-pawns. White is clearly better. — r1bqkbnr/pppp1ppp/8/8/2BpP3/8/PPPP1PPP/RNBQ1RK1 b kq - 1 5
4.Nxd4 exd4 5.O-O is the largest edge available anywhere in this line. The trade is the reason: Black's whole idea lives in that one knight, so trading it off removes the gambit, and the recapture leaves Black with doubled pawns on d7 and d4. The front one is what c3 and Re1 work against. The Estonian grandmaster Paul Keres, one of the strongest players never to become world champion, reached the same structure a move later with 4.O-O d6 5.Nxd4 exd4 6.c3.
Here is what the other fourth moves are worth from the diagram at the top of the page.
| White's fourth move | Value | What it means in practice |
|---|---|---|
4.Nxd4 | +1.2 | Best. Trade the knight, leave the weak pawn. |
4.O-O | +1.0 | Steinitz's other recommendation. Safe and good. |
4.Nc3 | +0.6 | Ordinary development, ordinary edge. |
4.c3 | +0.5 | Playable, but 4...Nxf3+ 5.Qxf3 gives back most of the advantage. |
4.d3 | +0.5 | Solid, but 4...Nxf3+ 5.Qxf3 trades the knight off without leaving Black the doubled pawn. |
4.Nxe5 | -0.5 | The pawn grab. Costs half a pawn and starts the argument above. |
The gap between 4.Nxd4 and the quiet developing moves is about three quarters of a pawn, which is a lot to give away for moves that also look sensible.
Books usually give a different route. Silman recommends 4.Nxd4 exd4 5.c3 d5 6.exd5 Qe7+ 7.Kf1, and it does keep a clear edge.
The cost is the king move on move seven. If you would rather not spend the opening with a rook on h1 and no way to connect it, 5.O-O keeps more of the edge and asks nothing of you. Both are good; only one of them is comfortable.
For where 3.Bc4 fits into a real opening rather than a trap, the Italian Game guide covers the plans this line skips.
Where The Name Comes From
Blackburne, a shilling, and a story nobody can verify
Steinitz gave the line its first known mention in 1895, in the addenda to the second part of his Modern Chess Instructor, where he recommended castling or trading the knight and assessed the bishop check as White's practical try.
The name arrived later and by legend. Joseph Henry Blackburne was one of the strongest English players of the nineteenth century, and reference works repeat the story that he used this trap to win a shilling a game from the opponents he met in cafes. Bill Wall has pointed out the obvious problem: there are no recorded games of Blackburne playing it. The name commemorates an anecdote rather than a score.
The line has two other names. It is sometimes called the Kostic Gambit, after the Serbian grandmaster Borislav Kostic, who is reported to have played it in the early twentieth century, though the one game usually cited for that is dated to two different years in the same reference. Graham Burgess records it as the "Oh my God!" trap, from the practice of exclaiming in fake horror after playing the third move, so that the opponent believes the pawn was hung by accident. Burgess adds that this is unethical, which it is, and that the trap leaves Black clearly worse once White avoids it, which it does.
At The Board
| You are | The move to get right | What to actually do |
|---|---|---|
| White | Move four | Play 4.Nxd4. The trade removes the one knight Black's idea depends on, and the recapture leaves doubled d-pawns for c3 and Re1 to work against. 4.O-O is the calm alternative: castle, ignore the knight, and take on d4 a move later. Both leave the pawn on e5 alone. |
| Black | Move five, after Bxf7+ | You have two legal moves and a real choice between them. ...Kd8 keeps the extra piece and the advantage. ...Ke7 puts your king on your queen's file, invites Bxg8 and the d4 and f4 pushes, and hands the better game to White. |
| Goal | Go here |
|---|---|
| Rehearse the trap from both sides | Practice the Blackburne-Shilling Gambit |
| Compare White's fourth moves | Open the explorer |
Build a real opening around Bc4 | Italian Game |
| Study the other attack on f7 | Fried Liver Attack |
| Make the pattern stick | How to practice chess openings |
Keep studying this opening
- Practice Blackburne-Shilling Gambit: Train the position against realistic replies.
- Explore the Blackburne-Shilling Gambit position: Review the move tree from the opening tabiya.