I am looking for a precise definition for a blunder. Computers often flag moves as blunders for any move that loses a point advantage or more. But what if I am completely winning and I decide to simplify the game by "sacrificing" my material plus. The computer may say that I lost more than a point advantage, but am I not more likely to win? There is also the scenario where a move may improve your position as evaluated by the computer but greatly decreases your chance of winning. Thus, I am interested in opinions on defining a better definition of a blunder than the computer uses.
The first English definition of blunder I see is "a stupid or careless mistake", and that's how I have always perceived the term in chess. It's not the severity of the error that matters so much as how silly and avoidable it is. If you played a move that takes the evaluation from 0 to -10 because you missed all the implications of a spectacular sacrifice that needed to be calculated out to 10 moves to verify its soundness, that's not a blunder, but if you played a move that takes the evaluation from 0 to -1 because you just didn't notice that a pawn was hanging, that is a blunder.
This is an interesting question. It seems that the higher the skill level, the less it takes to consider something a blunder. A GM may consider getting double pawns in a certain position a blunder while weaker players may just see it as a mistake or may not even call it a bad move. So how you call a blunder depends on your skill level because higher level players can better take advantage of all ranges of mistakes compared to weaker players.
A blunder is a move that gives your opponent enough of an advantage to win, or to get a draw after you should have won.
As someone else pointed out, it depends on the skill level. At a high level, the "loss" of a move or a key square may be a blunder. For intermediate players, maybe the loss of a pawn, and for beginners, maybe one or more pieces.
That is, these "advantages" are large enough, relative to the skill level, that upon seeing one of them, you can reasonably predict the likely winner.