1) 6. .. Bb4 is not better, don't believe what Eric Schiller told about it. After 7.exf6 Qxf6 White can play 8.Be3 or 8.Qe2
2) I think you are wrong that Stockfish recommends 11.Nb5 because after 11. .. Nxf4 !! White is lost and that is the reason 11.Qe2 has to be played. Maurits Wind analyses this move ( in Kaissiber 27 ) to find out that after 11. .. Qe6 ( with an exchange of Queens ) White still has compensation as well as with 11. .. Kd8
My game went as above. The database showed 11Nb5 being played in all 3 games, and Stockfish evaluated that position as white up about 0.8. I thoughtt white had compensation, but I won as black.
My game went as above. The database showed 11Nb5 being played in all 3 games, and Stockfish evaluated that position as white up about 0.8. I thoughtt white had compensation, but I won as black.
Stockfish 11? Depth 20? Yeah it’s kinda weak. Use Lichess analysis board if possible, that’s Stockfish 14 and goes past depth 20.
Stockfish evaluation :
The move 11. .. Nxf4 was never found before I published it in UON13 , October 2005. It is very hard to find after the board so if Black has not studied the Halloween you can safely play 11.Nb5 because all other black moves give White a decisive advantage. You find the complete analyses at https://halloweengambit.jouwweb.be/
Below is a game I played as black. The database I looked it up in gave 11.Nb5 as being played with a significant advantage to white according to Stockfish. That would probably be difficult for black to defend. My opponent played 11.Qe2+, allowing an exhange of queens, and probably an error. Apparently 6..Bb4, giving back the piece is considered better.
There are 52 games in with the Halloween Gambit in the database I was using, but none in the master database. The piece sacrafice looked odd when I saw it, but it may be fine to play at blitz or bullet, altleast below the expert level.