What is actually the best move here?

On move 10 I castled long and it got flagged as inaccurate. Furthermore it suggested I castle long instead. What's the idea here?

I just hope you've learned your lesson!

@Prinke said in #1:
>

There is no discernable difference between the two pathways you present, so there is not a decision response in existence that would enable any rational choice making action to be performed.

If you turn evaluation on at depth 27 it suggests 10. h4 is +2.2 while 10. O-O-O is only +1.5. I went up to depth 31 and the eval changed a bit but stockfish still preferred h4, I can't figure out how to lower the depth but I'd stick a fiver on O-O-O being preferred at some slightly lower depth like 25 or 26.

I noticed that too. One is SF 14 and one is SF 15 though.
I find it odd that it would evaluate a move as inaccurate and suggest the same move. Is it dope or nope?

I'm no computer guy but my first thought was some kind of print error, where the computer means to say one thing but prints another instead. Like if I mislabel sugar as salt so when you add it to the dish it comes out wrong. After the eval though I'm guessing it's a sort of print error, where it pulls the best move from one version of stockfish but prints it from the other?
SF15: best move is (print line 1)
Display: search for line 1, found line 1 on SF14, (print line 1)

I could be way off, just speculating.

Castle is best move in 10th move