the whole concept of predicting human behavior on very sparse and backward-looking data is problematic, not just in this one instance. even in cases where we ha

the whole concept of predicting human behavior on very sparse and backward-looking data is problematic, not just in this one instance. even in cases where we have a ginormous amount of data (equity markets, self-driving cars) and a recognizable set of rules/heuristics does it fail to predict human behavior well enough to depend on predictive models. we're infinitely complex relative to these models. they just don't seem to have enough degrees of freedom to pin us down (yet; who knows in a thousand years). in the case of landlords, it's tantalizing to pay a few bucks and get some comfort that you've done your due diligence, rather than facing the uncomfortable reality and esteem hit that there is no certainty predicting future behavior. and renters pay for that ignorance and moral hazard.