Using Chess.com's self analysis

So, I have a few questions regarding analysis features: 1. It seems like the +X for white and -X for black in a given position is suggestive of the point lead the position should be able to convert to in terms of material advantage if both players avoid mistakes. 2. In game report mode where best moves are shown it's a bit easier to tell what's going on, but it seems that in certain positions, following the "best move" for both players will occasionally lead to the advantage overturning in close games. (following the best move for both players on a game that says its -.32 black say, 10 moves on with both players playing "best" moves ends up +.5 white How can this be? I find I'm analyzing deeper to see if the advantage actually holds. Finally, ocassionally the engine seems to go back and forth with it's evaluation depending on which move you select--while its considering one the other seems superior and vice versa. Or it can't come up with a best move period. I've gone so far as moving every piece every conceivable way occasionally with the engine telling me occasionally different things... I'm sure it's some combination of ignorance in general to how analysis engines work and specifically how chess.com's does... someone will probably understand my sleep deprived noob speak well enough to know what I'm jabbering after. I'll be much obliged to their quick schooling of mine brain face.

I failed at paragraphs. How can I edit?

I’m probably as much of a noob as you are, but I think that the x advantage the analysis gives doesn’t have to translate to material. It can be a positional advantage or a good attack on the king.

Also, if either side has an advantage that is below 1, it will usually be a drawn game with best play from both sides as neither side has enough of an advantage to do anything. That being said, chess.com analysis is way behind engines like stock fish and alpha zero so maybe I am incorrect