Mostly my gut and previous observations. If he's daring scum then he'll probably slip up some time and it'll be easy to catch him. Don't see him as a strong lynch candidate atm.
The choice of No-lynch and my choice to not say that players are town is deliberate, yes. (A bit more on this later)
I'll discuss this more later on in this post too, but remember my ISO on Shifty? That's because I identified that as what was most important to me today, and I set out every part of my position and justification on why. I put high effort into what I considered to be an important read and important pressure.
I used to associate the two together and have learned my lesson. Which is another reason that I'm now more hesitant to throw out a townie label willy-nilly. I think a no-lynch label is far more accurate of my intentions toward a slot than "this is town."
I don't think I've equated or made a comparison between my read on Unknown and your read on me. I think that you make a lot of good points.
Anyway: since you're driving me to evaluate myself and my own play, I'll say this: I admit that I am lazy on Day 1. I think Day 1 reads are inconsistent, I've not been good at them, there are lurkers who lurk to live to Day 2, shitposters who have the same motivation, no flip yet and associations to consider, no night actions to read and evaluate, etc. These are all reasons for why I am far 'lazier' on Day 1 than I am on other days. Since so much of my scumhunting style relies on intent and motivation, it becomes so much easier on Day 2 and beyond to read those things with the wealth of new information that arrives. So yes, I'm lazy on Day 1 and prefer to focus my attention on the things that catch my attention (i.e. Unknown earlier in the day, Shifty now) rather than providing general reads list that often just end up getting revised after the flip and night, hammer vote, so on. I want to choose good pressure and lynch targets, and the information from Day 1 can then be kept in mind and factored accordingly beyond. Basically: I have some gut feelings and develop some feel for who I'd rather keep alive and who I want to pressure / lynch, and then I move on and adapt my reads to become more defined and reliable with some hard info for reading interactions and stuff.
This has worked for me, so this is how I've approached this game. If my results get worse, I'm certainly open to changing it.