May 7, 2014 10:04 PM
Are new anime overrated?
Anime Relations:
Clannad
Today, as of May 8, 2014, I decided to conduct a practical investigation in comparing new anime's scores with the global scores. There are going to be a lot of estimations here and there, but for the most part the method and data should be believable.
I have selected all the TV anime series aired after January, 1, 2013, and passed the minimum required votes (50) to have a Bayesian score. I used the anime search function, using "aired after 2013" and "TV series" as filters, and copied each page of anime information into excel. Somehow, despite the messy formatting, I was able to retrieve all the anime scores. There are 247 titles in total (again, I might have missed a few). The average score is 7.19.
Now, we can compare this to the global score for TV series, which is kind of a "calculated guessing" process, lol. Again, I resort to the anime search function with "TV series" as a filter. Obviously, it's impractical to calculate the average by copying each page (20 entries per page) into excel again, since there are a total of 2356 TV series entries with scores. This is easily found by sorting by "score", and then use bisection method to find where the total entries. By assuming that TV series have a normal distribution in score, I plan to estimate the average by taking the middle, and see how it deviates on both sides.
The 1178th entry has a score of 7.18. 75-percentile entry is 6.74 (0.44 from average), and 25-percentile is 7.61 (0.43 from average). It seems that it deviates evenly in both sides. Now, obviously I don't know any easy way of collecting the score for 2356 entries (I have to copy more than 100+ pages into excel), and I am too lazy to select a big enough random sample from the series to construct a sample deviation (yes, all for the sake of standard deviation). [Actually, technically it is possible. Just use any internet random number generator between 1 and 2356 and do it 50 times. Then, find those entries' scores. Maybe I'll do it tomorrow.
Anyways, sufficient to say, coincidence or not, the median of scores of all TV series (7.18) (of an assumed normal distribution; which is supposed to be equal to the average) is very close that of the new anime (7.19) aired after January 2013.
Based on the sketchy results from this investigation, MAL community as a whole does not overrate new anime. To be honest, I am relieved. While it is true that newer anime on a whole have more pleasing graphics, the "average quality", if you will, remains the same as it always has been.
I have selected all the TV anime series aired after January, 1, 2013, and passed the minimum required votes (50) to have a Bayesian score. I used the anime search function, using "aired after 2013" and "TV series" as filters, and copied each page of anime information into excel. Somehow, despite the messy formatting, I was able to retrieve all the anime scores. There are 247 titles in total (again, I might have missed a few). The average score is 7.19.
Now, we can compare this to the global score for TV series, which is kind of a "calculated guessing" process, lol. Again, I resort to the anime search function with "TV series" as a filter. Obviously, it's impractical to calculate the average by copying each page (20 entries per page) into excel again, since there are a total of 2356 TV series entries with scores. This is easily found by sorting by "score", and then use bisection method to find where the total entries. By assuming that TV series have a normal distribution in score, I plan to estimate the average by taking the middle, and see how it deviates on both sides.
The 1178th entry has a score of 7.18. 75-percentile entry is 6.74 (0.44 from average), and 25-percentile is 7.61 (0.43 from average). It seems that it deviates evenly in both sides. Now, obviously I don't know any easy way of collecting the score for 2356 entries (I have to copy more than 100+ pages into excel), and I am too lazy to select a big enough random sample from the series to construct a sample deviation (yes, all for the sake of standard deviation). [Actually, technically it is possible. Just use any internet random number generator between 1 and 2356 and do it 50 times. Then, find those entries' scores. Maybe I'll do it tomorrow.
Anyways, sufficient to say, coincidence or not, the median of scores of all TV series (7.18) (of an assumed normal distribution; which is supposed to be equal to the average) is very close that of the new anime (7.19) aired after January 2013.
Based on the sketchy results from this investigation, MAL community as a whole does not overrate new anime. To be honest, I am relieved. While it is true that newer anime on a whole have more pleasing graphics, the "average quality", if you will, remains the same as it always has been.
Posted by
bunny1ov3r
| May 7, 2014 10:04 PM |
Add a comment