What would be the ideal content rating system?

Question

Habr

We make a rating system in a project like 'Habr'. I would like to know your opinion.

Answer 1

Well, for example:
A special statistical technique known as Bayesian estimation (named after the author Thomas Bayes). It is designed to take into account not only the arithmetic mean of voters' ratings (average score), but also their number.
n is the required minimum number of votes cast for the element. 7.2453 - a certain average value taken as the basis of the method. The bottom line is that with a small number of votes, the estimated score will be close to 7.2453. As the votes cast increase, the role of the average score (arithmetic mean of the votes) will increase. An interesting feature is that, with the same arithmetic mean of votes, elements with a smaller number of voters have a higher calculated score. This distortion will be corrected with the arrival of new votes, although there is logic in it - if with a smaller number of voters the arithmetic mean is the same, then fewer low points are given.