Tea Ratings not calculated properly in JS
I have been wondering how the top rated teas are figured. It seems odd that a tea with only 9 ratings would come in right behind one with 94. The other day I looked at it and the top were all shifted around from where they had been and later were back, similar to the error you saw.
Steepster uses a Bayesian Algorithm to determine aggregate ratings.
Jason is one of the Admins here and he explains it a little in the link above.