softloud
generalise to other intuitions
This approach is pretty neat too because it could generalize not only to other distributions but also to other "intuitions". For ex, instead of specifying prop between desired mean / tau / gamma, could also do something like mean and fix desired value to hit by certain quantile
— Emily Riederer (@EmilyRiederer) August 14, 2020
Twitter doesn't give me enough characters for all of my random probability distribution musings! But I did want to say, this seems like an amazing generalization of the "triangle distribution" (parameterized by min / mode / max). I once took a simulation class where that was a popular high-level go to when you have no idea on the distribution 🤷.
Of course, that's obvious devoid of any scientific meaning and not a great abstraction of how things occur in nature. But being able to specify a couple of target quantiles for a more traditional distribution just as easily would be pretty neat!
