I wouldn't worry about unit designations until at least one winning season at a given sport.
But do take notes and keep track of 1 star, 2 star & 3 star plays or whatever, you'll learn a lot. After enough data, you'll start to see a trend about whether your 3 star play is really that much better than your 1 star play, etc.
For me there isn't that much difference between a good play and a great play. So, I just flat bet everything 1% of BR. It's more important to distinguish a play from a no-play than trying to apply Kelly's system, which most people get wrong anyway.
Also, you'll notice the guys that assign unit/stars/whatever to plays or either touts or wannabe touts or somehow picked up that habit along those lines. Basically touts will maybe give away a 1 star bet for free but you've got to pay for their 5 star pick, it's just a marketing gimmick.
So, in conclusion, maybe shelve the unit designation until you've got good data history showing that your different play designations really do hit at meaningfully different rates. If you can confidently determine your 60% ATS plays from your 65% ATS plays, then go ahead and dive into Kelly's system, here's the Wiki:
https://en.wikipedia.org/wiki/Kelly_criterion