[pcnix]

This is similar to calculating probability from an odds ratio - if x = 1 and if n = your odds ratio, p = n/(n+1)

A neat explainer for how this helps with Bayesian updates (which is the update in trust that you're describing) is here - https://www.lesswrong.com/posts/QGkYCwyC7wTDyt3yT/0-and-1-ar...