The above page describes how the top 250 is calculated where the m (prior) value is set at 25000 for top 250, and for bottom 100 , the m value is set at 1500.

My question is the following

1)Why conduct bottom 100 separately because the bottom 100 can be obtained after obtaining the Bayesian scores for the entire data set of movies while obtaining top 250 ?

The following website uses a only a single formula with m=13.

http://www.animenewsnetwork.com/encyclopedia/ratings-anime.php#best_bayesian

A example would help.

Lets assume the following case

I have 50 movies in the database. I need to calculate the top 5 and bottom 5.

I set the m value at 200 ( based on the finding that 70 percent of the movies i,e 70% of 50 = 36 movies have ratings above 200)

I now calculate the Bayesian scores for all the 50 movies

and obtain the top 5 and bottom 5 from from the rankings obtained for those 50 movies.

The above method is pretty much the way it should be done and is done by the anime website. Does it validate the question that i asked about imdb ?