> keep person homeless because nobody will want to rent a flat because someone's kids stained the walls in the rented apartment.
If Evan's kids stain the wall, then Evan's kids are more likely to stain the wall. Currently, the expectation of someone being more likely to stain the walls is "priced in" to the rental prices everywhere. People who don't stain walls pay a little more, and people whose kids stain the wall cost their owners a little more, but on average it comes out (hope you can own a lot of domiciles).
If we start to be able to measure this, doesn't this mean we can more accurately price it? People less likely will pay less, and people more likely will pay more. I think you're suggesting that "Owners will just use the likelihood of damages as a threshold, and if you're 0.01 above the Unsafe/Volatile threshold you'll not be rented to at all", but that leaves a pretty big hole in the market. If there's not enough supply, then people are being kicked out instead of Evan _now_, and if there is enough supply, then you'd rather rent to Evan at higher rates to cover your risks than to not rent at all, right?
The current situation kinda feels like an "insurance" thing, with pooled risk.