Hacker News Reader

Or more generally, look for critical points in the model and see if it still holds. Max/min values (or odd combinations of max/min for different variables) are good candidates, as are zeroes, and anything which makes part of an equation go to zero.

I've always thought this should be a very effective way to explain a point to someone, but in practice it rarely seems to work....maybe that saying applies, something about you can't use logic to change the mind of someone that didn't use logic to arrive at their conclusion.

That might be because you're trying to use it to argue politics, where it's less applicable; hard cases often make bad law, and you can easily end up with a straw man. It works better in science and engineering.

Also a basic programmer skill. Check a normal value, limits, and if you find it some values that may lead to unexpected results like division by zero.

Right. The Laffer curve https://en.m.wikipedia.org/wiki/Laffer_curve applies this idea

Yeah, that's somewhat related to sensitivity analysis that's already in the list.

Though I think I'd agree that it's technically a different model, but related.