If you play the Jerome Gambit (1.e4 e5 2.Nf3 Nc6 3.Bc4 Bc5 4.Bxf7+ ) long enough, you will eventually play the 6.d4 variation (4...Kxf7 5.Nxe5+ Nxe5 6.d4) and one day you will come face-to-face with the best in-your-face variation for Black since J.H. Blackburne offered his Rook...
This move (6...Qh4) was first seen in Sorensen - Anonymous, Denmark 1888 (see "Jerome Gambit Tournament: Chapter X").
I have 193 games with 6...Qh4 in my database, with only 10 wins for Black, but those numbers are deeply flawed, as most of the games are computer vs computer, and the great majority of those come from a source that only provided White wins (see "Bright Ideas from Silicon"). Of the 8 person vs person 6...Qh4 games in my database, White has won 4 and lost 4, which still seems overly optimistic to me.
Here's a quick game to try to sort out the variation.
Danivarl - Bb35
1.e4 e5 2.Nf3 Nc6 3.Bc4 Bc5 4.Bxf7+ Kxf7
5.Nxe5+ Nxe5 6.d4 Qh4
This move is simply wrong. If White wants the piece, he should castle first, then go after it: 7.0-0 Qxe4 8.dxc5
Would it be helpful to mention that in their games Deep Shredder 10, Deep Sjeng 1.5, Gambit Fruit1, Hiarcs 11.1 and Shredder Paderborn all chose 7.0-0 instead of 7.dxc5?
Also leading to an advantage for Black (but not as big a one) is 7...Ng4, from a game earlier this year where the second player was finally able to snatch defeat from the jaws of victory: 8.Qd5+ Kf8 9.g3 Qf6 10.Be3 Ne7 11.Qd2 Qxb2 12.0-0 Qxa1 13.Nc3 Qb2 14.f3 Nxe3 15.Qxe3 Qxc2 16.Ne2 Qxa2 17.f4 d6 18.f5 Nc6 19.f6 Bh3 20.fxg7+ Kxg7 21.Qg5 checkmate, yorgos - mofte, FICS, 2009.
A bit better (but still very good for Black) was 8.Kf1, but after 8...Nf6 White has to be scolding himself: I sacrificed a piece for this position??
Black went after the other Knight pawn with 8...Qb4+ in AlgozBr - nikolasa, FICS 2009, a wild game where the second player maintained his advantage for a score of moves and then swapped blunders with White at the end. To be fair, the first player waited to get a winning position before resigning (or losing on time): 9.Nc3 Qxb2 10.Qd5+ Kf8 11.0-0 Qxc3 12.Bd4 Qxc2 13.Qxe5 Qg6 14.Qf4+ Nf6 15.Qxc7 Nd5 16.Qd8+ Qe8 17.Bxg7+ Kxg7 18.Qg5+ Qg6 19.Qxd5 Rg8 20.Rae1 Kh8 21.Re7 d6 22.cxd6 Bh3 23.Rfe1 Bxg2 0-1
9.Rf1 Nf3+ 10.Ke2 Nf6
Just to show you that the Jerome Gambit belongs to the Twilight Zone of chess openings, here is a game where Black chose 10...Qg4 instead, 11.Qd5+ Ke7 12.Qxf3 Qxf3+ 13.Kxf3 d5 14.Bd4 Nf6 15.Nc3 Bg4+ 16.Kf4 h5 17.Rae1+ Kd7 18.f3 Bh3 19.Rg1 Rae8 20.Rxg7+ Kc8 21.Reg1 Nd7 22.R7g3 Rhf8+ 23.Kg5 Be6 24.Kxh5 Rh8+ 25.Kg5 Rxh2 26.Kf4 Rxc2 27.Nb5 c6 28.Nd6+ Kd8 29.Nxe8 Kxe8 30.b4 Rc4 31.Ke3 Rxb4 32.Rh1 Kf7 33.Rg7+ Ke8 34.Rh8+ Nf8 35.Rg6 Bf7 36.Rg5 Rb1 37.Re5+ Kd7 38.Rxf8 Re1+ 39.Kf4 Rxe5 40.Bxe5 Be6 41.Kg3 a5 42.f4 Ke7 43.Rf6 a4 44.Rh6 Kf7 45.Rh7+ Kg6 46.Rxb7 Kf5 47.Rb6 Ke4 48.Rxc6 Bf5 49.Ra6 Kd3 50.Rxa4 Kc2 51.c6 d4 52.c7 d3 53.Rd4 Kd1 54.c8Q Black resigned, yorgos - ErrareHumanumEst, FICS, 2009
Yes, the name of the second player translates from the Latin as "to err is human."
You know you're doomed when you can play a move like 11.Nd2 instead and Black can safely ignore it, as after 11...d5 12.Nxf3 Bg4 it is still a sad position for White.
11...Ne5 12.Qb3+ d5 13.Nd2 Bg4+
White's downfall is his unsafe King: the irony!
14.Ke1 Rhe8 15.Qxb7
Leading to an explosive conclusion.