I Need to Keep Learning About the Jerome Gambit

The following game is my third Jerome Gambit in the Chess.com Giuoco Piano tournament. With two wins and a draw, I am happy with "my" opening - but, as the following game shows, my middlegame play needs improvement! 

perrypawnpusher - dzetto00
Giuoco Piano tournament, Chess.com, 2016

1.e4 e5 2.Nf3 Nc6 3.Bc4 Bc5 4.Bxf7+ 

4...Kxf7 5. Nxe5+ Nxe5 6. Qh5+ Kf8

Here we have the Jerome Defense to the Jerome Gambit, dating back to a couple of correspondence games, Daniel Jaeger - Alonzo Wheeler Jerome, 1880 (Black won both).

7.Qxe5 Qe7 8.Qf4+ Ke8 9.O-O

Instead, 9.Nc3 Nf6 10.O-O Bd6 was how perrypawnpusher - Chesssafety, Italian Game, Chess.com, 2012 (1-0, 25) continued. Of Black's 10th move I noted "an interesting idea that sets up tactical chances on the Kingside" although I criticized White's next move, 11.Qe3 - "Carrying on as if everything is "normal", while Black actually can play 11...Ng4 now, with serious threats."

9...Bd6

Less aggressive was 9...d6 in Petasluk - ilanel, blitz, FICS, 2014 (0-1, 49)

10.Qf3 

Trying to follow the lesson of  perrypawnpusher - Chesssafety, Chess.com, 2012. For once I did not want Black's Knight to "harass" my Queen (after 10.Qe3) with a move like 10...Ng4.

10...c5 11.d3 h6 12.Nc3 Nf6 

Black has defended against Bc1-g5, as well as d2-d4.

White's strategy should be to quickly get his Rooks working on the e- and f-files - especially since Black's King and Queen are precariously placed. Unfortunately, that will mean developing his dark-squared Bishop to either a meaningless square (d2) or exchanging it off and easing Black's "traffic jam" (i.e. the Bishop at d6 blocks the pawn at d7 which blocks the Bishop at c8 which blocks the Rook at a8).

13.Bf4 Bxf4 14.Qxf4 d6 15.Rae1 Nh5 16.Qd2 Be6 17.f4

Finally White has a Jerome-looking position! My big challenge was could I make use of the time that my opponent spent on 10...c5, 11...h6 and 15...Nh5 ? The answer is "yes" - and "no".

17...Kd7 18.e5 Rhf8 19.d4

Hoping to open up central files against the King and Queen.

19...Rxf4

A surprise.

20.Rxf4 Nxf4 21.Qxf4 Qg5

Another surprise; and, actually, not the best move.

I exchanged Queens here, as I thought it would mean at least a drawn endgame, with some chances to win. It seemed to be a good idea at the time.

22.Qxg5 hxg5 23.d5

My original idea had been to exchange Queens and then play 23.exd6 Kxd6 24.dxc5, but then I saw that 23...cxd4 for Black, instead, messed up things. I then looked at 23.dxc5 dxc5, but I was concerned that the open position would favor Black's Bishop and make things drawish (or worse).

I finally came upon an idea that was paradoxical (and wrong) - a line where I thought I could get serious play in a closed position.

23...Bf5 24.e5+ Ke7 25.Rf1 Rf8 drawn

I had totally overlooked the simple idea of Black using the Rook for protection of the Bishop. Instead, I was caught up with the ideas behind 25...Bxc2? 26.Rf7+, etc. (I lost our first game by similarly overlooking a move, thinking he had to block a check with his Rook - when I would have good play - when he simply used his Knight, and I was busted.)

My opponent suggested that we had reached a draw, and I agreed by offering one, which he accepted.

By the way, the proper response to 21...Qg5 was 22.Qf3!?, continuing the attack on Black's King, by making use of a diagonal as well as the central files, e.g. 22...Rb8 23.dxc5 Qf5 24.Qd1, but that is computer analysis that I need to study another day...