There are speculative sacrifices in master games.
Then our study of the masters game was faulty! Each masters game with sacrifice I have found precision calculations - there have been positional sacrifices - but those are not speculative - they knew exactly what they were doing!
MIght I suggest reading up on Rudolf Spielmann? He seemed to disagree alot with what you are saying, and he was no doubt of grandmaster level. (As strong as that would have been in the 30ies.)
crissy, I belive that 'speculative' and 'unsound' are synonyms in chess sacrifice lingo. A speculative sacrifice is one that is unsound but the opponent may not find the correct path to demonstrate the refuation. The terms do not reflect whether or not the player pragmatically benefits from the sacrifice, only its theoretical status.
Yes, I rather thought so - it's problematic though. There are other issues as well that aren't described very well by a sliding scale of ? - ?! - !? - ! and so on, or their verbal equivalents. As far as I remember reading, the Immortal game is 'unsound', though presumably AA didn't think it was, and I guess he wasn't acting speculatively either. Really we need to introduce the idea of whether it worked or not at the time under the heading of 'un/successful'; 'un/sound' is fine, after the event, and I would reserve 'speculative' for certain players' styles. I think Tal would be the outstanding example of a speculative sacrificer.