Gregor Klevže
Member
OP
Yea thats happend because we use Bayesian estimate rather than standard average rating suma_votes / num_votes. <br />We dont use this formula because some article have 100 votes of 4 and one article have 1 vote of 5, will be this article with ONLY one vote higher rated! <br />So we deciede to use BAYESIAN ESTIMATE. which also counts how many votes article gets .... <br /> <br />Detailed description: <br /> <br />We are using the same true Bayesian estimate formula used by the Internet Movie Database, for calculating average ratings. The only difference is that our rating system uses multiple dimensions (presentation, graphics, usability, technical, and personal slant), so we apply the formula to each dimension by using the respective mean value of that particular dimension (C), take the sum of individual results, and divide this sum by the number of dimensions, which is five. <br /><pre> <br /> weighted rank (WR) = (v / (v+m)) * R + (m / (v+m)) * C <br /> <br /> where: <br /> R = average for the design (mean) = (Rating) <br /> v = number of votes for the design = (votes) <br /> m = minimum votes required to be listed in top 25 (currently 6) <br /> C = the mean vote across dimension <br /></pre> <br />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 design has only a few votes above the minimum required votes to be listed in top 25 (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. <br /> <br />Here is an example for a single dimension: <br /><pre> <br /> WR = (6 / 10) * 5.33 + (4 / 10) * 7.18 = 6.07 <br /> | | | | | | <br /> v v+m R m v+m C <br /></pre> <br />The formula normalizes the average rating of a relatively low rated design from 5.33 to 6.07 since the number of votes (v=6) is only slightly above the minumum required votes (m=4) and the mean across the dimension (C=7.18) is quite high. If, in the future, this particular design gets more votes, the difference between R and C will increase as the number of votes increase. The idea is that the more the votes, the more representative the average rating is. Please note that this example is only for one dimension. Different dimensions have different means (C). <br /> <br />Or find more about this formula on google :)
Edited
