Just as a small reference: I have to say today's list includes a lot of great content :) So you are doing something right already... 

nout

I suspect I'll need to use some kind of [binomial proportion confidence interval](https://en.wikipedia.org/wiki/Binomial_proportion_confidence_interval).

k00b

Users who have more in common with their votes, trust each others votes more.

But I need someway to determine this relationship with a potentially small sample size

Are you structuring it such that users with stronger connection have more weight in their votes or the opposite

Daily discussion thread

saloon

Going to hopefully push some of the little fixes I have locally today. Then going to go back to work on trust algo.

Ideally I can come up with some kind of model for how much a user trusts another user's upvotes given what they've each upvoted in the past.