Beer Ratings Quality Assurance : RateBeer.com


HOME	RATINGS	MY PAGE	FORUM	BEST BEERS	BEER INFO

Beer Ratings
Quality Assurance

RateBeer has created a number of ways to assure that our ranking systems are among the best of their kind. We use a weighted average (Bayesian) to calculate beer and brewer scores so that the validity of any average is increased (as the score tends toward the actual average) by the number of samples.

A rating may not count toward a total score for a number of reasons.

1) The rater has fewer than 10 ratings
A user must have more than 10 ratings before his or her ratings affect the real score of a beer or brewer.

2) The rating is obviously bogus
Obviously bogus ratings are deleted immediately and sometimes so are the user accounts of those who post them.

3) The rating was made by a brewer or brewer affiliate
Ratings by brewers or their affiliates of their own products is a strictly forbidden at RateBeer. The ratings and the accounts of those making the ratings are swiftly deleted and account renewal is forbidden for life. 'Blackballing' brewers for blatant misconduct in this regard may be considered in extreme cases.

RateBeer has been known to use extreme measures for ensuring the validity of our ratings -- including conducting IP address traces. If you suspect a 'troll' may be submitting harmful or abusive ratings, please contact a site administrator immediately.

Joe Tucker
Executive Director
RateBeer.com
joet@ratebeer.com

Membership: Sign Up Now!
About RateBeer
Advertise With RateBeer

Bayesian Formula
We use the same true Bayesian estimate formula used by the Internet Movie Database for calculating average ratings.

weighted rank (WR) = (v / (v+m)) * R + (m / (v+m)) * C

where:
R = average for the design (mean) = (Rating)
v = number of votes for the design = (Rate Count)
m = minimum votes required to be listed in the top beers list (varies according to average of ratecounts for top 50 beers)
C = the midpoint of the scale (2.75 in our case)

This formula normalizes scores, that is it pulls a particular score (R) to the mean (C) if the number of votes is not well above m. In other words, if a particular beer has only a few votes above the minimum required votes to be listed in top 50 (m), the average score is decreased a little if it is above the mean, or increased a little if it is below the mean in accordance with the normal distribution rule of statistics.

The idea is that the more the votes, the more representative the average rating is.

Interface design by