My opponent and I were playing a "normal" blitz Jerome Gambit game until we each started to make our moves too quickly. Things degenerated quickly into a state where "the winner is the one who makes the next-to-last blunder". In this case, it was me.
perrypawnpusher - Gryllsy
blitz, FICS, 2014
1.e4 e5 2.Nf3 Nc6 3.Bc4 Bc5 4.Bxf7+
According to The Database, Gryllsy - zagothal, blitz, FICS, 2013 continued 5.d4 Bxd4 6.Ng5+ Ke8 7.Qf3 Nf6 8.c3 Bb6 9.Be3 d6 10.Nd2 Bg4 11.Qg3 Bxe3 12.fxe3 h5 13.h3 Bd7 14.0-0 Ke7 15.Ndf3 Rf8 16.Nh4 Be6 17.Ng6+ Kd7 18.Nxf8+ Qxf8 19.Nxe6 Kxe6 20.Rf5 Qf7 21.Raf1 Rf8 22.Qf2 Ne7 23.g4 Neg8 24.g5 Nxe4 25.Rxf7 Nxf2 26.Rxf8 Ne7 27.R1xf2 Black resigned
5...Nxe5 6.Qh5+ g6
I have a theory about this move. Some defenders push the g-pawn because is part of a defense - Blackburne's, Whistler's - that they are familiar with and are ready to play. Others do so, though, almost as a reflex, to punish White for his early Queen attack - and they figure that they will work out the rest of the defense later.
I checked The Database and found 411 games with the position after 6...g6. Of those games, 139 continued, after 7.Qxe5, with the Blackburne Defense, 7...d6. Another 52 games saw Whistler's Defense, 7...Qe7. That means that in over half of the games where 6...g6 was played, Black was either committed to an inferior defense, or to "figuring something out" - which amounted to the same thing.
"I'll take Door Number Three, Monty."
Instead, Black played 8...Nxe4 in perrypawnpusher - LibertasProVita, blitz, FICS, 2009 (1-0, 45) and perrypawnpusher - ibnoe, blitz, FICS, 2012 (1-0, 16).
Also seen was 8...Qe7 in perrypawnpusher - marbleschess, blitz, FICS, 2009 (1-0, 48); and 8...d6 in perrypawnpusher - MsD, blitz, FICS, 2007 (0-1, 27), perrypawnpusher - brain50, JG3 thematic, ChessWorld.net, 2008 (1-0, 24), and perrypawnpusher - tiagorom, blitz, FICS, 2009 (1-0, 41).
9.d3 d6 10.Qe3 Ng4
Also played: 10...d5 in perrypawnpusher - andrecoenen, blitz, FICS, 2010 (1-0, 15) and 10...Kg7 in perrypawnpusher - Alternative, blitz, FICS, 2005 (1-0, 63).
For historical purposes, let me point out that 11...Kg7 was Black's response in Vazquez,A - Carrington,W, Mexico, 2nd match, 1876 (1-0, 39).
The game has lost its attack and counterattack, but White is ahead two pawns.
13.0-0 Kg7 14.Nc3 a6 15.Bg5 Ng4 16.h3
The mistakes start to creep in, small ones at first. A bit better was 16.Nd5 c6 17.Nc7 Be6 18.Nxa8 Rxa8.
16...Ne5 17.f4 Nf7 18.Bh4
Better still was 18.Nd5 Nxg5 19.Nxc7 Nxh3+ 20.Kh2 Rf8 21.Nxa8 Nxf4 but at this point I wasn't looking that deeply into the position.
18...b5 19.Nd5 Ra7 20.Bf6+ Kg8 21.Ne7+
Missing 21.Bd4 c5 22.Nf6+ Kf8 23.Nxe8 cxd4 24.Nf6 Kg7.
A gift. I had expected simply 21...Kf8 22.Nxc8 Rxc8.
22.Bxe7 c5 23.Rae1?
Returning the favor. I learned to drive in New Jersey, where the two controls on the car are the gas pedal and the horn...
The game is now roughly even, with White having an Rook and two pawns vs two pieces.
24.e5 dxe5 25.fxe5 Rxe5?
25...Nxe5 was the proper recapture, even with the risk of leaving the Knight pinned to an undefended Rook, because of a tactical shot that my opponent and I both missed.
26.Rxe5 Nxe5 27.Re1 Nc6
Black's best here was 27...Nxd3, although he is worse after 28.Re8+ Kg7 29.Rxc8 Nxb2 30.Rxc5 Kf6.
My opponent and I both missed that 28.Re8+ would fork King and Bishop.
28...Bf5 29.Re3 Kf7 30.g4 Be6 31.a3 Kf6 32.Kf2 Kg5?
One last slip, to seal the game.
33.Rxe6 Black resigned