IMDB score rank algorithm

The IMDB website is currently the most authoritative, systematic, comprehensive movie website on the Internet, including almost all movies, and TV dramas after 1982. The IMDB’s information includes many information, actors, packages, content introduction, grading, comment, etc., to a large extent, to the IMDB score, for personal buying discs. Its unique film scoring system is deeply popular, and the registered users can score and review any movie, and the website will calculate the weighted average of the film according to the average division of the movie, the number of votes, etc. This is carried out in this way of TOP250 (best 250 movies) and Bottom100 (worst 100 movies).

Due to the update and the income score of the movie, TOP250 and BOTTOM 100 must be a literal list, but most of the positions (or popular) the location of the film will be relatively unchanged, so this TOP list is also There is the meaning of peeping the taste of the popular movie. When you select the best 250 movies, just consider the results of the official voter. The score system uses a 10-point system, the minimum of 1 point, up to 10 points for Excellent (outstanding). It is worth noting that although many movies are scored high in the data system, they cannot participate in the rankings due to the minimum 1250 votes required to reach TOP. As a result, many excellent films in many songs (at least in the United States) are unable to be included. Despite this, this list still has an important reference value to the selection of the movie.

There is a problem here: the average score of the movie and the cooler movie is really comparable? For example, a Hollywood blockbuster has 10,000 viewers vote, and a small cost of literary film has only 100 viewers vote. How is the voting result of these? How can I fix a movie true quality fairly?

A reasonable idea is that if you have to compare the good and bad movies, you should at least ask the same more viewers to watch and score. Since the number of people’s audience is small, you should try to add some viewers to it. At the bottom of the ranking page, IMDB gives its calculation method.

IMDB has two scoring systems, one is a routine score for each film, one is a special score (weighted average) to enter TOP 250 (weighted average).

According to the TOP250 scoring criteria published on the IMDB website:

The IMDB TOP 250 uses the weighted Rank-WR of the Bayesian statistics, and the formula is as follows:

Weighted Rank (WR) = (V ¡Â (V + M)) ¡Á R + (M ¡Â (V + M)) ¡Á C

The use of Chinese is: Weighted average division = [Regular voter voting number ¡Â (often voter voting number in TOP250)] ¡Á ordinary method calculated the average divide + [Minimum ticket required to enter TOP250 (often The number of voters votes + the minimum number of tickets needed to enter TOP250)] ¡Á The average score of all movies


R = average for the movie (mean) = (rating) (the average) calculated using a normal method)

v = number of votes for the movie = (votes) (the number of voters, it should be noted that only frequent voters will be calculated, this will explain in detail below)

m = minimum votes Required to be listed in the top 250 (Currently 25000) (the minimum number of tickets required to enter the IMDB TOP 250, only three or two people vote, it is useless, and it is currently 25,000 votes)

C = The mean vote Across The Whole Report (Currently 7.0) (Currently average score of all movies, currently 7.0)

Carefully study this formula, you will find that IMDB adds 25,000 votes for each movie, and these votes are 7.0. The reason is that all movies have at least 25,000 votes, then they have the first 250 selection criteria; then assume that this 25,000 votes are the average score of all movies (ie, assume that this movie has this movie Average levels); Finally, with existing audience voting for amendment, long-term perspective, the weight of V / (V + M) will be larger and larger, and the score will slowly close the real situation.

This makes the difference in voting people between different movies, making movies with fewer voters may also rank the front.

How to discharge the brush behavior to the final score, the key is that the “Regular Voter Votes” is worth this value, only the “Regular Voter” vote will be calculated within the IMDB TOP 250, this It is IMDB defense because some kind of movie fans affect TOP 250 results, and TOP 250 is limited to the main method of voting in the voter voting. In order to prevent users from cheating, the criteria of the “Regular Voter” have been unapproved, but we are not difficult to guess, it requires a requirement for the user’s voting history, such as the number of votes, or meets a certain voting qualifier, or a vote for one year, or And the user’s registration time, usage frequency, etc. may also be too much voting in the time of filtering.

The score on the IMDB is completely commended by netizens, with the own preferences of participating netizens participating in the evaluation, the more netizens participating in the score, the more reliability of IMDB scores.

Seeing this formula, we can understand why IMDB ranking is the most authoritative, although it is a website mainly based on an English and American audience, but is still highly respected by the world’s fans, this is the objective and impartiality of science It is not possible. Other so-called list of rankings can be forged, can spend money, or a business game, but IMDB is not, so this is its value.