Welcome back, Cliff Hardy.
Notes are primarily by Hardy [I have added a few comments in blue - Rick]
Cliff Hardy (2176) - NN (1842),
1 0, Lichess, 2018
1.e4 e5 2.Nf3 Nc6 3. Bc4 Bc5 4. Bxf7+
4...Kxf7 5. Nxe5+ Nxe5 6. Qh5+ Kf8
7.Qxe5 d6 8.Qg3 Nf6 9.Nc3 Be6 10.0-0 Ke7?
Stockfish found that White could take the g-pawn here with the very strong 11.d4!! Bxd4 12.Qxg7+ Bf7 13.Bg5 c6 (or 13...Rg8 14.Bxf6+ Bxf6 15.Nd5+! Ke6 16.Qxh7 ++-, when Black's king would be perilously placed on e6) 14.e5!! Bxe5 15.Rae1 ++-, when White would be threatening to take the knight for free on f6 or to attack the pinned bishop with 16.f4, but I would never have been able to find this line in 1 0 chess (nor in a full-length game 😉).
11.d3? Rf8?! 12.Be3?
Rather than go for Morin's c3, d4 approach [see Morin - Guipi Bopala, Quebec Open, 2018 - Rick] to blunting Black's dark-squared bishop, I chose to blunt it with my own bishop, in an attempt to advance my pinned f-pawn. But since Black has still left his g-pawn vulnerable, Stockfish prefers 12.Na4 Bb6 13.Nxb6 axb6 14.Qxg7+, though Black would then have had a slight advantage after 14...Rf7.
12...Bb6?
12...Bxe3 would not have lost time in retreating and would have yielded Black a winning advantage here.
13.d4
By now the time for White to capture the g-pawn had well and truly past, since from here on it would just lose time with the queen for White and open the g-file for Black to use in attacking the white king.
13...Qe8 14.f4 Qh5 15.f5 Bc4 16.Rf4 Rae8 17.b3 Ba6
13...Qe8 14.f4 Qh5 15.f5 Bc4 16.Rf4 Rae8 17.b3 Ba6
Getting the Jerome pawns mobilised immediately with 18.Rh4 Qf7 19.e5 ++-, while Black's king is still sitting on e7, would have been much better.
18...Kd7 19.Bf2 Qf7 20.e5 dxe5 21.dxe5 Nd5??
The final losing move. 21...Nh5 would actually have been winning for Black, since then 22. e6+ Kc8 23. exf7?? would have been met with 23...Rxe1 mate!