Alas, Poor Bishop, I knew him well.
(Apologies to Will Shakespeare.)
In 1936 in America there was a chess game to which I didn't pay much attention as, some 3,548 miles away, I was in my cot, busy grizzling and cutting my baby teeth at the time. In this particular game, of which I show below the first four moves, there was a possible blunder/master stroke which has gone down in chess history - I don't know of anyone replicating the game which, as you will see, is quite unusual.
The event was in America - the Philadelphia Chess Tournament on Monday, 9th November, 1936. The two players were F. Arnold (White) and Milton Loeb Hanauer (Black). The opening used was the Budapest Gambit which, at that time, was still very popular and is still played a bit.
White's move 4. Bg5 seems to be the turning point with white apparently pinning down the knight but with no realistic immediate threat. Black responds with an unexpected and seemingly wild move Ne4, exposing his Queen with an unguarded Bishop now attacked by both a Black Knight and the Black Queen. The White Bishop - alone and unsuported, instead of retreating to a Bishop swap-off at e3, charges in (Bxd8) and goes for the material (Black Queen=9 points versus 3-and a bit for himself with the added bonus that Black will no longer be able to castle versus white losing tempo.) ... seems almost too good to be true.
My old Mum used to say, "In this World you don't get owt for nowt" (Translation: you don't get anything for nothing) and my Dad used to warn, "If it looks too good to be true then it probably is and he's got something up his sleeve - a clenched fist !"
As you have probably already worked out the "Clenched Fist" was Black's response Bxf2#
So this brings us to the built-in problem of "Blunder" or "Sacrifice" - or to put it another way ... the person you're playing is no fool so "What's in it for them?"
If the answer is "nothing" - then look again 'cos it looks too good to be true ...
But, if I'm the one doing the playing (attention span of a gnat and with possibly a three-move horizon) then, in a fictitious game, this might happen:
I intended the sequence:
White Pawn takes Black Pawn, supporting Black Pawn takes White Pawn, White Queen takes Black Pawn (Check) ... but, as I have been known to do, I reverse the order a trifle and play: White Queen takes Black Pawn with a potential response of Black Pawn takes White Queen ... Oops ... I think most people would recognise that as a 'Blunder' and unhesitatingly take the 'Sacrifice' with a happy grin.
In theory, Sacrifice moves should have the 'pay-off' beyond your move horizon. If you have a high move horizon, i.e. you are a good player, and you are playing seriously and you still can't see the 'pay-off' for the apparent sacrifice then it's a matter of how you rate your current opponent (even Nimzowitsch misjudged his opponent occasionally !)