So around 3 hours ago, I decided to learn Jerome Gambit Theory to play it against my friends and or my dad. I found some theory, but it wasn't all that extensive for the extremely dubious lines.
That's where my genius idea came in to find traps in the Jerome I'll never be able to use anyway. How? Using low depth Lc0 (pure positional, depth 4) and Stockfish (degraded because max depth 13).
For reference, the line which was played had a evaluation at 1 point of ~-7. But stockfish found a ridiculous perpetual. Ofc, I'm thinking at this point i should see if chess.com thinks it's brilliant as well.
Turns out, chess.com didn't think a rook sac was brilliant enough. But I've noticed in the past that it'll give random brilliant moves if the material is very imbalanced and you just keep checking.
So I made with the help of the engines a few more checks. That somehow gave the game 4 brilliant moves.
I was not satisfied.
So I made the decision to play the game on, checking if black could block with the queen and still draw the game. After a bit of analysis, it turned out that black did still have a draw.
At this point it was just about creating a situation where 1 player needs to make a single sacrifice to keep the game drawn. And Lc0 is perfect at creating these situations, since if Lc0 thinks it's winning, and stockfish thinks it's a draw, it'll most likely require a move that's hard to find at low depth (brilliant moves)
And low and behold, around 20 moves later, I saw an opportunity arise. Make a move for white that looks to be losing, black makes the best move, and the only drawn move for white is to sac a bishop.
Around 15 moves later, I can end the game in a theoretical draw and put it in chess.com's analysis. Low and behold, 5 brilliant moves by 1 player in 1 match.
https://cdn.discordapp.com/attachments/758702903673356328/823712773811863552/5_Brilliant_moves_in_Jerome_Gambit.jpg
If you want to see the pgn and chess.com's analysis, check here: https://www.chess.com/analysis/game/pgn/3WjoJuqjA6