Jerome Gambit: Adding to Theory

I recently received an email from chessplayer Rahul T Reji. (I have added a few diagrams.) He adds to Jerome Gambit theory.

I have been analyzing a particular position in the Jerome Gambit and believe I may have found a move worth considering. I have pasted below the PGN detailing the line for your review. You will notice that although black does not make any major blunders, the game slips away from its grasp very quickly. I would love to hear your thoughts on it and whether you think it holds any potential in practical play. It is currently the second most popular choice on chess website lichess.org, the most common being 9.Qe3

1.e4 e5 2.Nf3 Nc6 3.Bc4 Bc5 4.Bxf7+ Kxf7 5.Nxe5+ Nxe5 6.Qh5+ Ng6 7.Qd5+ Ke8 8.Qxc5 d6 9.Qb5+ 

The most common response to d6 is the move Qe3, which is also the best engine move. However, I found that playing Qb5+ and then Qb3 leads to a slightly more advantageous position (when playing against humans) the Queen on the b3 square controls the f7 square, preventing the very common castling "by hand" maneuver (Kf7 Re8 Kg8), while also restricting development of the bishop on c8, as it would lead to the loss of the b7 pawn. 

9...c610.Qb3 

Computer gives this position -1.9, whereas after 9.Qe3, computer gives -2.1 

10...Nf6 11.O-O Qc7 12.Nc3 Bd7 13.d4 Ke7 14.f4 Rae8 15.f5 Nf8 16.e5 16.Ng4 

Very natural move, which is also a blunder. In order to avoid complete collapse, black had to sacrifice a pawn here - with dxe5

17.Bg5+ Nf6 

only move 

18.exf6+ gxf6 19.Rae1+ Kd8 20.Bxf6+ Kc8 21.Bxh8  Black Resigned

I have a little to add to Mr. Reji's work.

But, as you will see, only a little.

I appreciate that he could shine a light on a pretty hazy bit of the opening. He is right, it does hold potential for practical play, for the reasons he points out.

A couple of years ago, in the post "Jerome Gambit: Easy Come...," I looked at the game angelcamina - FutureWorld_Champion, 1 0 bullet, lichess.org, 2023, which featured 10.Qc4 instead of 10.Qb3, but noted

In earlier games angelcamina has retreated his Queen to b3 or b4

Currently, The Database show 12 of angelcamina's games with 9.Qb5+ c6 10.Qb3 but none of them have appeared on this blog.

Eleven years ago, in the post "Tangled", I shared the game Wall,B - Guest7556673, PlayChess.com, 2014, which did see 10.Qb3; as did "My House! My House! My Kingdom for a House!" with the game Wall,B - Yunfan, Chess.com, 2011.

After the latter post, I pointed out

This is something new: in almost 24,000 games in The Database, this position does not show up at all!

To update, The Datbase currently has over 116,000 games. 34 of those games go 1.e4 e5 2.Nf3 Nc6 3.Bc4 Bc5 4.Bxf7+ Kxf7 5.Nxe5+ Nxe5 6.Qh5+ Ng6 7.Qd5+ Ke8 8.Qxc5 d6 9.Qb5+ c6 10.Qb3. In 1/3 of the games, angelcamina played White.