Hacker News Reader

This is cool. But what I usually do is replace x/y with x/(y+5), and hope for the best :). The 5 can be replaced by 3 or 50, depending on what I'm dealing with.

(in less important areas than sorting things by ratings to directly rank things for users; mentally bookmarked this idea for the next time I need something better, as this clearly looks better)

View on HN

Heads up this weights all your scores towards 0. If you want to avoid this, an equally simple approach is to use (x+3)/(y+5) to weight towards 3/5, or any (x+a)/(y+b) to weight towards a/b. It turns out that this seemingly simple method has some (sorta) basis in mathematical rigor: you can model x and y as successes and total attempts from a Bernoulli random variable, a and b as the parameters in a beta prior distribution, and the final score to be the mean of the updated posterior distribution: https://en.wikipedia.org/wiki/Beta_distribution#Bayesian_inf...

(I saw first this covered in Murphy's Machine Learning: A Probabilistic Perspective, which I'd recommend if you're interested in this stuff)

Reply on news.ycombinator.com