Inspired by a recent video where Magnus Carlsen was surprised that the engine thinks his position was a draw. Even top players may find a position considered winning can actually be drawn with perfect play, but is too hard to defend in practice, but that doesn't mean their evaluation is unjustified. Conversely a weak player like myself may find it hard to convert what engines think is a big advantage because I can't find the only winning move(s).
I'm interested in finding a way to make engine evaluations also show how hard it is to find the best move sequence for a given position, for example whether most reasonable-looking moves according to common principles are playable, or if only one/a few moves (especially if counterintuitive ones) can win or defend the position. And by extension, how difficult it will be for the opponent to defend against a candidate move that I can make, not just how objectively good it is with perfect play (kind of like the "dark forest where 2+2=5" idea. And possibly how easy it is to play a wrong move as well, for weak players like me :)
Even if it's as simple as adding some grey area in the evaluation bar to show difficulty or likelihood of keeping/losing an advantage, but possibly much more sophisticated than that. I'm willing to try writing some software if none are available.
Similar questions may have been asked before but no good solution seem to be given: Are there any engines that present evaluations based on probability of finding moves?