Which score should be used as a reference when viewing only one site at film criticism sites "IMDb" "Rotten Tomatoes" "Metacritic" "Fandango"?

ByAndrew Hitchcock

When I thought about seeing a movie, there are various criteria such as "Who is the director?" "Who is starring," but one of them has "what kind of evaluation is getting"? It is best to compare the scores of various sites carefully and refer to the reviews as well, but, "Which site should you look for when it is time to know caku?", In short, which film critique The data scientist Alexandru Orteanu is investigating whether to refer to the site.

Whose ratings should you trust? IMDB, Rotten Tomatoes, Metacritic, or Fandango?
https://medium.freecodecamp.org/whose-reviews-should-you-trust-imdb-rotten-tomatoes-metacritic-or-fandango-7d1010c6cf19


Mr. Orteanu was based on the criteria that "Which site's score is closest to the normal distribution" in the selection. This is the experience that Mr. Orteanu has watched a movie, there are not so many works that are strongly impressed whether it is nice or bad, and the majority was an "average" work that makes it impossible to remember even the plot It is based on that.


The candidate site is "IMDb","Fandango","Metacritic","Rotten Tomatoes'S four sites. For the sampleData of 214 works released from 2016 to 2017Was used.

By the way, the reason for this way of dividing is partly due to Fandango's scoring system being changed to 2015, but in the survey score data of 4917 works of IMDb were found and distributed similarly to the data of 214 works From what I showed, Mr. Orteanu says that even 214 works are sufficient as a sample.

This graph shows the score distribution of 4 sites. Each site has a different score for each site, but Mr. Orteanu has less than the bottom 30% of the score for "terrible movie", less than 30% "average movie" score, less than 70% "good movie" score It is defined as a line on the graph as well.


IMDb is shaped like a normal distribution with a high mountain shape in the center, but it is characterized by no work belonging to "terrible movie" defined by Orteanu (low score is small).

Fandango is similar to IMDb, but it is characterized by a higher score. Fandango in the pastClearly the score is too highI have received an indication. This is the comparison graph at that time. Fandango does not score below 2.5. Originally the score was incremented by 0.5 and rounding up was rounded up so that the score of "4.1 on average" would be "4.5" rather than "4.0" was one factor.


Although it seems that changes have been made to the coefficients as a result of being pointed out, it seems that it is not yet enough, as you can see from the survey of Mr. Orteanu this time.

Metacritic is also a normal distribution type, but on the other hand, Rotten Tomatoes tended to be different from the other three sites. Rotten Tomatoes has "Tomatometer" showing the proportion of positive review by critics and "Audience Score" where the audience scores, where Tomatometer is used but it is close to flat from top to bottom , There was no missing score. Mr. Orteanu thought that this result is inappropriate because I thought that positive review and negative review would be about 7: 3 in many movies.

Based on these results, Mr. Orteanu concluded that "if you want to check the score quickly on one site," Metacritic "". However, Metacritic's meta score calculates scores by weighted average, and weighting is performed for each review in the range of 0 to 100 based on the quality of the review from the Metacritic management team and other reviews written Although I am aware that its weighting factor is private, it points out that it is a drawback that you do not know which review has impact on the score. Also, if it is a movie before 1999 that Metacritic can do, it will extremely reduce the number of reviews, and if it is a work other than English it is also a difficulty not to mention even the list.