Suppose you do trade of your Bishop you need to get your King to the E/F-line ASAP so that white pawn is under attack. I actually think the trade of wasn't that bad especially because his got an isolated pawn except if he trades of his B pawn, but you got a rook+King to stop 2 pawns + rook which is possible imo.
Then you get an endgame of 4 pawns vs 2+ king which you might actually win?
Or you get this as an alternative situation
Here is a game I recently lost. I've added my own annotations, including my thoughts throughout the game. It's always better for me to hear others' ideas, so if you have any comments, please post them! Thanks!
Overall ideas:
1. I got distracted during moves 18-20 by the white bishop and switched my focus from the whole board to a couple pieces. These moves should have gone something like: (This is posted in the analysis above) 18...Rfd8 19. Bb4 Rd5 20. Rad1 Rad8 21. Rxd5 Rxd5.
I don't have to worry about the pin on my knight here. My bishop is fine where it is. And if I end up having to move my bishop, the knight is still protected by the rook.
2. 28... Bxd4 was the worst move of the game for me. I should not have traded my bishop, as it would have been useful in the end game. And white's knight didn't pose as much of a threat as I thought it did. It should have gone something like: (This is posted in the analysis above too) 28... Rd8 29. Nc6 Rd7.
And now my bishop can come back go g7 and then f8 to help attack. Also, it's more difficult for white to get two passed pawns with the c-pawn. He has to go to more trouble to defend his c-pawn so that he can play b3.
30. Rd3 Bg7 31. b3 Bf8 32. c4 bxc4 33. bxc4 Rxd6 34. Rxd6 (if not, he loses his knight) Bxd6. And I am much better with my bishop against his knight.