Hacker News Reader

Comment by Automatt-2 | Hacker News Reader

>So if a product has a large number of ratings, but they are 50/50 positive and negative, then it's not going to rise to the top simply because it has the most reviews.

What I'm saying is that it will rise to the top because even though only 50% of the ratings are positive, it will have many many more positive ratings than the nearest next item. That 2nd item may have a much higher percentage of positive ratings, but a much lower number of total ratings.

View on HN

It seems to work as described in the article:

    require 'rubygems'
    require 'statistics2'

    def ci_lower_bound(pos, n, power)
        if n == 0
            return 0
        end
        z = Statistics2.pnormaldist(1-power/2)
        phat = 1.0*pos/n
        (phat + z*z/(2*n) - z * Math.sqrt((phat*(1-phat)+z*z/(4*n))/n))/(1+z*z/n)
    end

    puts ci_lower_bound(60, 100, 0.10)   # => 0.517809505446319
    puts ci_lower_bound(500, 1000, 0.10) # => 0.474027691168875
    puts ci_lower_bound(50, 100, 0.10)   # => 0.418847795168265

The top ranked result has 60 positives out of 100, and beats 500 out of 1000 - almost 10 times as many positive votes.

Reply on news.ycombinator.com