I'm asking more about why a specific move is called a blunder. If you're playing as black, and white uses the Fried Liver attack, when you use the Bishop and do Bxf2+, Stockfish on Chess.com says that's a blunder. I'm curious as to why. I know it's (probably) giving up a valuable bishop, but White's King playing Kxf2 allows for Qh4 (thus removing the Queen from the threat of the rook on f7), and also possibly puts pressure on white's Bishop on c4.
Is it a blunder solely because there's no equal or greater exchange? Is it something else?
After 1.e4 e5 2.Nf3 Nc6 3.Bc4 Nf6 4.Ng5, now 4....Bc5 is a mistake by Black. If White goes 5. Nxf7 then following up with 5...Bxf2+ is the best move in that position, so definitely far from a blunder. Maybe the engine is running at a low depth or you may have made some error when inserting the position.
Anyway engines don't classify moves as "blunders", they just return a number as the position's evaluation. chess.com classifies moves according to those numbers into different categories (blunder, brilliant move, inaccuracy and so on...) but you shouldn't pay too much attention to those since they're arbitrary categorizations of something that exists on a continuum.
There is much inconsistency in the way this question is posed. The Fried Liver is associated with 4.Ng5 d5 5.exd Nxd5 6. Nxf7 (a sacrifice) The Traxler is associated with 4.Ng5 Bc5 and now perhaps 5. Nxf7 (a grab of material) In the first case Bxf2+ is not possible, so OP must be asking about the second case, when 5.Nxf7 Bxf2+ has always been regarded as the correct play and White usually declines with 6.Kf1.
The alternative in the Traxler of 5.Bxf7+ has always been regarded as safer although not necessarily stronger. Beliavsky held the resulting position against both Karpov and Anand, drawing with Karpov and even winning when Anand blundered. AFIK, theory has not yet spoken the last word.