I don't think there has ever been or will ever be a consistent way of approaching potential in as much as we don't set specific criteria based on youth caps etc. A lot of it is left to the individual researcher and in the main I'll trust them to get on with it.
What we do do however is urge our researchers to set a 'fixed' potential ability rather than a 'minus' one once a player has established himself in the first team.
But then you raise a good point comparing Middleton and Johnstone because they are quite similar. Neither are fully established first team players yet, and yet one researcher has chosen to give a fixed potential and another a minus one. I would say that in a sense, both are slightly wrong but at the same time it's difficult to find the right way.
I would say that being on a periphery of the first XI each week, they have the potential to be good enough to be regular first team starters. At the moment for Celtic, that's maybe 120-137 and for Rangers, 116 - 130.
So what do you give them? -7 has the potential to mean their development would cap just about where they currently are, but could also make them potential stars for both teams. A score of -75 puts them from being at worst 120 (which would be ok) to 150 (which on the current evidence, would be a bit of a leap).
You could say "Why not give them both 135" and that would be fine, although I'd like there to be a sense of variability in their potential when they are yet to be fully established.
As for the other ones you mention...
I don't think McBride should have a fixed one, and I'll look at changing that. Ditto Henderson.
I'm ok with Wilson having -75 as he's still young enough for that to become a reality, but he would need to show signs big of development this season (and be a standout at Forfar) to justify carrying a potential like that, fixed or minus, into Football Manager 20.
Does all that make sense?