[continued from previous post]
My second Jerome Gambit, in the second round of the "Italian Game Classic" tournament at Chess.com, was a battle. I flailed around for a while, trying to put together a decent plan. In the end, it was the pawns.
Italian Game Classic, Chess.com 2019
12.Bg5
This move is often thematic in the Jerome Gambit. However, when the game was over and I submitted it for Chess.com's computer analysis, the verdict was that while 12.Bg5 was "good", 12.h4 was "best". However, in the same position, I would choose my move, again, instead of weakening my Kingside.
12...Qe5
Unpinning the Knight, moving the Queen to a more active post - and offering to exchange Queens. With Black's dark squared Bishop still holding back my f-pawn, I wasn't sure what my best way forward would be - but it certainly wasn't exchanging Queens.
13.Bf4 Qd4 14.Be3 Qb4 15.Bxc5 Qxc5
Never mind what the Chess.com computer said about the past few moves (I am sure that you can guess) - I have gotten rid of the pinning Bishop, and Black's Queen will have to stay on the a7-g1 diagonal to keep my f-pawn at home.
16.Rae1 h4 17.Qf4 Qe5 18.Qd2 Be6
19.f4
Ta dah!
"Good", but 19.h3 was "best". Maybe so, but how does White follow up? The Chess.com computer's move would have led me further into the wilderness.
19...Qd4+ 20.Kh1 h3 21.g3 a5
This move reminded me, fondly, of the Fidelity Chess Challenger 7, a dedicated chess computer (running on the 8-bit Z80 cpu) which came out in 1979. Whenever it assessed its position as good, but didn't have a way forward, it would often launch one of it's Rook pawns.
I was skeptical of my opponent's move when he played it, as it seemed like it gave me a free tempo, but the Chess.com computer later labelled it as "good". I figured that it was time to strike in the middle, and, of course, played another "inaccuracy".
22.e5
The computer preferred 22.Ne2. That is an interesting move, and Stockfish 10 (at 35 ply), afterwards, could barely differentiate it from 22.Nd1 and 22.f5 (they were within 4/100th of a pawn of each other!). That last move held the possibility of White following up with Rf1-f4, and then g4 and a later g5, so I should have given it consideration.
22...Nd5
Okay, this is a "blunder", but it is helpful to understand why.
When White's pawn advances in the Jerome Gambit, the question for Black is often "to exchange or not?" Black decided not to, in our game, but he overlooked the better 22...Bd5+, probably because the piece can be snapped off right away with 23.Nxd5. After 23...Qxd5+ 24.Kg1 Qd4+ 25.Kh1 Qd5+, etc, Black would have a draw - but why would he want it? Instead, 25...Nd7 would keep Black's advantage, as after 26.e6, the pawn could easily be blockaded.
Black's move gave me a tempo, and I spent it on the "Jerome pawns".
23.f5 Bg8
24.e6
I could not figure out how to make 24.f6 work, and, therefore, settled for blocking in Black's light square Bishop - which often gets blocked in (or neglected) on the Queenside. The g-pawn looked like it had a future in a supporting role.
Black's next few moves show that he wasn't sure how to continue, either.
24...Nxc3 25.bxc3 Qf6 26. g4 Qh6
It is true that 27.Qxh6 now would turn the advantage back over to Black, but I was not interested in anything so peaceful. Besides, there was another tempo to find for my "Jerome pawns".
27.g5 Qh5 28.Qf4
Certainly, 28.f6 also would work here, but the text targeted Black's pawn at d6, giving White's Queen entrance behind enemy lines. The f-pawn would have its day.
28...Qe8 29.f6 g6 30.Qxd6+ Black resigned
(I just noticed that I still had all 8 of my pawns.)
