Right now black can't promote their pawn without losing it, as soon as he tries, white takes it. Blocking it with the bishop sets up the only situation where the new queen is a factor in the game... Which happens to be a checkmate in 2 moves.
Right, so moving the bishop allows black to discover check with the rook by promoting the pawn. It's not quite mate in two because white can delay the inevitable by covering its king with the knight.
My question is, why is it better to move the bishop to b4 than to take the pawn at f4?
1 is you lose your bishop to the white pawn next. Then the King has more mobility to get away from your queen/rook
And, you still haven't broken the stalemate of the white rook keeping your pawn from promotion. White just never moves the rook and black can't move his rook and that leaves a knight and two pawns against a king and two pawns... With a white rook covering 7, while the black rook just stares on. The bishop sacrifice at that moment gives black a checkmate in 3 (you're right not 2).
How do you see it playing out if black takes the pawn?
(There are a couple things white could do here, including taking the pawn at g6+, but nothing that really puts white ahead, because black is eventually looking to promote the pawn for discovered check. White can't take the pawn with his rook before it's promoted because he loses the rook. He can't take the promoted pawn either because he needs to get out of check from the black rook. Uncovering the king by taking the bishop with his pawn would work to white's disadvantage against a rook-queen attack.)
Then white's only move would be:
Nf2
Then black could either take the white rook with the new queen or try to maneuver into mate with the rook and queen.
I don't see a huge difference - black loses - or at least risks losing - the bishop either way, and also is able to promote the pawn either way. The white rook can't take the new queen because he is in check from the black rook.
My solution opens up the king's defenses a little and gains black a pawn
That is true. White doesn't have to play Rxb4 after Bb4 either.
Once white stops the discovered check with Ng2, black can either capture the rook at b7 or try for mate. White can't move to the king to g1 or h1 because the new queen is covering that row.
In either case, you are moving the bishop out of the way for a discovered check. My question is wouldn't it be better to get a pawn out of it?
That is true. White doesn’t have to play Rxb4 after Bb4 either.
It doesn't matter whether white takes the bishop or not, in any case black can promote without being taken by white's rook, for example 1...Bb4 2. Ng2 b1=Q, and white can0t take the new queen.
Once white stops the discovered check with Ng2, black can either capture the rook at b7 or try for mate.
I don't think I follow you. Do you mean after 1...Bxf4 2. Ng2? How can black capture the rook? If the pawn promotes, the rook simply takes the new queen (1...Bxf4 2. Ng2 b1=Q 3. Rxb1). And trying for mate without promoting the b2 pawn seems unnecessarily complicated to me (I mean, the position is -5 in favour of black according to the engine, so probably there is some path to mate, but it isn't obvious)
White can’t move to the king to g1 or h1 because the new queen is covering that row.
White moves the king before the pawn promotes: 1...Bxf4 2. Kh1. And again, if you promote, the white rook simply takes the new queen 1...Bxf4 2. Kh1 b1=Q+ 3. Rxb1