| Edit:Updated stats here Someone in the MSP club suggested this and it piqued my curiosity so I did a bit of web scraping and here are the results. The chart has the top 10, the rest are given below.  When I say Favorites proportion I mean the percentage of viewers who put the anime in their favorites. As we can see this means that 11.66% of people who watched Steins;Gate put the anime in their favorites. Favorites Percentage = No of Favorites/(Total Members-Plan To Watch Members) *100 Top 30 Collected Data of Top 200 popular anime : http://s000.tinyupload.com/index.php?file_id=83876767409342094795 The data is only of the top 200 anime (excluding 3 which had special characters causing issues while Web Scraping) Observations
Questions 1)Is this metric meaningful? It seems harder to skew this metric with bots than the current mass rating of 1's to a show. It also seems to have an inverse sequel effect for some reason. 2) Does this mean Steins;Gate is MAL's favorite anime and not FMA:B? 3) Do you prefer this top 10 over MAL's current top 10? |
xLoopJun 29, 2019 12:05 PM
<Something> |
| 1. (a) Yes, data is important. (b) I would assume the inverse sequel effect to be because people generally see the first airing to be encompassing of the franchise, and most don't watch sequels. I myself don't want redundant favorites clogging up my limited space. 2. If you look at it like this yes, otherwise no. Judging which one is more relevant is up to each individual. 3. Yes, i prefer knowing which anime has the greatest effect on people to add to their favorites. Edit: zombie_pegasus said: What else isn't on the list, @-InfiniteLoop-? Replied below.LoGH has 14.5% of its non-ptw members putting it in their favourites, which puts it way above Steins;Gate. |
LunilahJun 12, 2019 5:21 PM
 I don't believe in the Devil. You should. He believes in you. |
| 1. Obviously, it's meaningful. This is pretty much just an upgraded version of the "Most Favorited" list. It's still possible to skew this metric, but really mostly in the upwards direction. You could theoretically bring the percentage down by spamming accounts that have the show on their list and not on their favorites, but that wouldn't really be efficient. I personally don't add sequels onto my favorites list. I have Bakemonogatari on there, but that's probably my 5th favorite Monogatari season. It's easy to just put the first season onto your list and have it represent the whole franchise. I assume many other users think this way as well. EDIT: Thinking about it, this can also work in the opposite direction. People inclined to only have one entry per franchise on their favorites may choose to pick their favorite season instead of the first one. Thus, some people have the first season on their favorites list, while others have whatever season was their favorite. This is a disadvantage for shows with numerous seasons. 2. Probably not. This metric only shows what anime people love, but not what anime they hate. All you're accounting for is how many people have the show on their favorites. For example, on a very small scale, say 100 people each watch Anime A and Anime B. For Anime A, 50 people love the show and put in on their favorites list; the other 50 people give it a 1/10. For Anime B, 10 people have the show on their favorites, 10 people gave the show a 1/10, and the other 80 didn't have very strong feelings about it. According to this metric, Anime A is 5 times better than Anime B. But is that really a fair statement to make? (This is obviously a very contrived example, but it illustrates my point). 3. I definitely don't think this is better than the current top 10 for reasons mentioned above, but it's certainly an interesting metric. As I mentioned before, it's much more accurate than the Most Favorited list, since that one doesn't account for popularity. Death Note is 3rd on that list simply because it's the most popular show, thus automatically having a good number of favorites. |
MythologicallyJun 12, 2019 2:17 PM
 |
| I'm just happy that Made in Abyss is in the top 30. Really interesting graph. |
 |
| Yes it's important and very interesting. I'm prefer the score, but it would be nice if we can see this stat on mal or grap+. Just few interesting from my fav. list: Sono Hanabira ni Kuchizuke wo: Anata to Koibito Tsunagi: Yuri hentai 1,47% :) Haibane Renmei 5,02% 2) No, but the Stein;Gate have more impact and that mean S;G a bit more important anime than FMA:B (I' think) |
| Good Luck ; Have Fun |
-InfiniteLoop- said: Someone in the MSP club suggested this and it piqued my curiosity so I did a bit of web scraping and here are the results. Ah, thanks dude. My post was from year ago and at first I read my post again to remember what I said... You take my post seriously even more than myself :) The data is only of the top 200 anime (excluding 3 which had special characters causing issues while Web Scraping) Rip Madoka cuz I think its "Favorites proportion" is 6.34 which means #14 of this ranking. Don't care about Tokyo Ghoul season 2 and Lucky Star lol Observations
FMA:B is normier than Steins;Gate and that makes it becoming #4 of this list :P It's interesting and unexpected for me that Movies have less percentage rate. Also for the last one Naruto Shippuuden and Clannad After Story has more favorite percentage than S1s but for other anime it's true which Clannad After Story is obvious for what but idk about Naruto Questions 1)Is this metric meaningful? It seems harder to skew this metric with bots than the current mass rating of 1's to a show. It also seems to have an inverse sequel effect for some reason. 2) Does this mean Steins;Gate is MAL's favorite anime and not FMA:B? 3) Do you prefer this top 10 over MAL's current top 10? 1) Definitely yeah for me. 2) Hmm, don't know maybe 3) If you mean top 10 highest rated, no but if you mean top 10 most favorited yes but the problem of this system is encouraging elitism which isn't a good thing |
mhkr said: So does having any sort of statistics at all. Opinions themselves breed elitism. I see this as a non-issue.the problem of this system is encouraging elitism which isn't a good thing |
I don't believe in the Devil. You should. He believes in you. |
Lunilah said: mhkr said: So does having any sort of statistics at all. Opinions themselves breed elitism. I see this as a non-issue.the problem of this system is encouraging elitism which isn't a good thing umm, yeah but assume this become instead of what considered as "Most favarited anime" and then we have useless wars about it |
mhkr said: Sounds like a great topic to hash out a dialogue on. We know what the stats tell us, but not why it came to be, and if that offends people so what? I would hardly call it useless but definitely ironic, since it would be figuring out our feelings.Lunilah said: So does having any sort of statistics at all. Opinions themselves breed elitism. I see this as a non-issue. umm, yeah but assume this become instead of what considered as "Most favarited anime" and then we have useless wars about it |
I don't believe in the Devil. You should. He believes in you. |
| This is some interesting data. Also, i'm quite pleased to see that NGE and TTGL are at least among the top 10 Favorited, considering that they're not even close in MAL's official Top 10. |
| This is similar to this comment from two days ago: https://myanimelist.net/forum/?topicid=1786611&show=0#msg57786840 And it's been done in the past all across the database with similar results (I'm not quite sure how many shows this user in particular checked). The top 200 in popularity limitation you set here is quite arbitrary since the amount of people who completed Ginga Eiyuu Densetsu for instance is already sizeable. Also I don't think it's more meaningful, you can only have up to 10 favorites in MAL and the more you watch the less representative it gets. Other anime sites don't have a limit there which would make it at least slightly better. But at the end it's just another way to rank the most popular shows like every other top (with a potential exception being the aforementioned LoGH which would only piss off fanboys of other titles because "muh elitists" and "people just adds that to pretend, right?"). |
 |
| What really sticks out to me is how Bakemonogatari is the highest of the series by this metric. I think there's case a for Kizu III but I'm tempted to say this is how it should be. Have never understood why Owari 2 is highest by score, other than the sequel effect. |
YossaRedMageJun 12, 2019 3:54 PM
| “In individuals, insanity is rare; but in groups, parties, nations and epochs, it is the rule.” -Friedrich Nietzsche Aggregate scoring is bad for the anime fandom |
Unowen said: Praise be to LOGH.This is similar to this comment from two days ago: https://myanimelist.net/forum/?topicid=1786611&show=0#msg57786840 And it's been done in the past all across the database with similar results (I'm not quite sure how many shows this user in particular checked). The top 200 in popularity limitation you set here is quite arbitrary since the amount of people who completed Ginga Eiyuu Densetsu for instance is already sizeable. Also I don't think it's more meaningful, you can only have up to 10 favorites in MAL and the more you watch the less representative it gets. Other anime sites don't have a limit there which would make it at least slightly better. But at the end it's just another way to rank the most popular shows like every other top (with a potential exception being the aforementioned LoGH which would only piss off fanboys of other titles because "muh elitists" and "people just adds that to pretend, right?"). @shanimebib, the only thing about this i would criticize is that it doesn't account for those who have anime on favorites but haven't completed them. But even with that i don't think it would change much across the board. |
I don't believe in the Devil. You should. He believes in you. |
| LoGH has 14.5% of its non-ptw members putting it in their favourites, which puts it way above Steins;Gate. I honestly don't think this data is all that important. Most of these are popular series which means a lot of their viewers have smaller pools of anime to stick into their favourites. In LoGH's case we all know about the 4chan circlejerk so of course it has a high percentage of favourites. @YossaRedMage Unlike scores which can be given indefinite numbers of times each, favourites can only be given to 10 different anime so people will often use one anime of a franchise to represent the whole franchise. Some choose the season or movie they liked the most from the franchise, but some will just use the starting point because they see it as representing the series as a whole. |
  |
Lunilah said: 1. (b) I would assume the inverse sequel effect to be because people generally see the first airing to be encompassing of the franchise, and most don't watch sequels. I myself don't want redundant favorites clogging up my limited space. Yes, this makes sense. I figured it might be something like this but it's interesting to see that MAL users prefer to put the first season in their favorites rather than the one which seems to have the higher score. There are exceptions of course like Clannad After Story having more than double the percentage of it's first season. Mythologically said: It's still possible to skew this metric, but really mostly in the upwards direction. You could theoretically bring the percentage down by spamming accounts that have the show on their list and not on their favorites, but that wouldn't really be efficient. Haha I never even considered this possibility but yes it's less efficient for them. Mythologically said: 2. Probably not. This metric only shows what anime people love, but not what anime they hate. All you're accounting for is how many people have the show on their favorites. For example, on a very small scale, say 100 people each watch Anime A and Anime B. For Anime A, 50 people love the show and put in on their favorites list; the other 50 people give it a 1/10. For Anime B, 10 people have the show on their favorites, 10 people gave the show a 1/10, and the other 80 didn't have very strong feelings about it. According to this metric, Anime A is 5 times better than Anime B. But is that really a fair statement to make? (This is obviously a very contrived example, but it illustrates my point). Yes, it's not optimal although it provides a new perspective and insight I suppose. Perhaps a weighted metric that's a mixture of both the score and this metric would provide a more balanced approach. Both the metrics are already collected in the file, the question is how to combine the two. I do feel this metric works well for popular anime and high rated anime but it's significance kinda fizzles out the lower we go. Unowen said: This is similar to this comment from two days ago: https://myanimelist.net/forum/?topicid=1786611&show=0#msg57786840 And it's been done in the past all across the database with similar results (I'm not quite sure how many shows this user in particular checked). The top 200 in popularity limitation you set here is quite arbitrary since the amount of people who completed Ginga Eiyuu Densetsu for instance is already sizeable. That's an interesting comment however it only takes into consideration the amount of people who completed the show. That would skew the data in favor of anime with a large number of episodes as those who don't like the show that much are unlikely to complete it. I also feel not taking into consideration the people who dropped an anime isn't accurate. It was just a small experiment, I was curious how the top ranking and top popular anime would stand when measured with this metric. I figured 200 would be a decent number to be statistically significant yet not too large to take too much time to web scrape the data. I can always increase this amount. I wanted to check if such a metric had any merit first. That said, scraping the entire MAL database in not feasible, what do you think a decent number would be? The reason I chose popularity wise is because the current top anime by favorite list is heavily skewed in favor of popular anime and this metric seemed like an apt solution to that problem. Ginga Eiyuu Densetsu was an oversight on my part, it's ranked really high and yet has a relatively very low popularity. It's an outlier of sorts, I'm contemplating expanding it to maybe top 500 popularity wise and manually adding the link of such outliers. Unowen said: Also I don't think it's more meaningful, you can only have up to 10 favorites in MAL and the more you watch the less representative it gets. Other anime sites don't have a limit there which would make it at least slightly better. But at the end it's just another way to rank the most popular shows like every other top (with a potential exception being the aforementioned LoGH which would only piss off fanboys of other titles because "muh elitists" and "people just adds that to pretend, right?"). It's a double edged sword, the limited number of favorite slots means an anime with a high favorite percentage must really be something special. The significance of this metric might wane when it comes to comparing mediocre anime but for highly regarded anime I think it provides some interesting insight at the least. @Lunilah - The data consists of the top 200 popular anime excluding Madoka, Tokyo Ghoul root A and Lucky Star because of their hyperlinks containing special characters that seem to be causing some encoder problems with the html parser I'm using. My program prints an exception message when it's traversing these links so I'm easily able to identify them so manually adding such exceptions is possible. I suppose people find the top 200 popularity to be less (?) or maybe I should have parsed the top 200 top rated anime. Expanding on the size of the former is possible and the latter option is also feasible, I stated why I went with popularity above. Anyways I'm open to suggestions. |
xLoopJun 12, 2019 4:54 PM
<Something> |
-InfiniteLoop- said: Ah right, popular, not ranked. That's my fault for forgetting something i've already read.@Lunilah - The data consists of the top 200 popular anime excluding Madoka, Tokyo Ghoul root A and Lucky Star because of their hyperlinks containing special characters that seem to be causing some encoder problems with the html parser I'm using. My program prints an exception message when it's traversing these links so I'm easily able to identify them so manually adding such exceptions is possible. I suppose people find the top 200 popularity to be less (?) or maybe I should have parsed the top 200 top rated anime. Expanding on the size of the former is possible and the latter option is also feasible, I stated why I went with popularity above. Anyways I'm open to suggestions. The prime argument i can think of for doing top ranked over most popular is, statistically speaking, more people will be exposed to great anime they weren't aware of. I would definitely prefer both though. Top 500 ranked is also anime rated 8.05+ which is aesthetically pleasing. Edit: @-InfiniteLoop-, i wonder if it's possible for you to plug the data in somewhere to be able to compare it with peoples existing lists, would definitely facilitate the idea of finding new anime. |
LunilahJun 12, 2019 6:11 PM
I don't believe in the Devil. You should. He believes in you. |
| i havent updated my favorites like ever. I imagine a lot of people dont, but most people throw down ratings so thats probably a more meaningful thing to take into account if you do care |
| https://combosmooth.itch.io/ - I make free-to-play browser games for PC and I sell pixel art animation here |
| I actually scrapped the data based on top 500 rankings like one day later (June 15th) but I became so overwhelming busy with my new job that getting even 4 hours sleep was a blessing. Anyways just gonna dump the results here for the few who might still care about it. New top 10 based on data of top 500 ranked anime  So yeah Legend of the Galactic Heroes is number 1, good for all you die hard fans. Aria is an anime I've heard about a few times that completely flew under my radar, I feel more inclined to giving it a watch now. Yojouhan Shinwa Taikei is an anime that I've never in my entire life heard about. Some charts I also made a couple weeks back. Tell me if there's any confusion on what they're representing. FP means favorite percentage. Note: The data is based on the top 500 ranked anime however it excludes anime with extremely low popularity/favorites so in total there is 416 anime. Most of the excluded anime were specials, niche movies (like a dozen detective conan movies), OVA's etc. I have a list of the excluded anime. They're basically negligible/garbage data. Also I manually inserted the few with special symbols so yes Madoka and a few other popular ones are included in the list. Anyways here's some charts I made because I like looking at them. Just a cool curve showing trend of FP. The x-axis is the rank based on FP  FP on y-axis, MAL score ranking based on those in the data set on X  Scatter graph showing absolute MAL score and FP correlation  New top 50 I think Code Geass is the only anime with both it's seasons in the top 50. That's a testament to it's impact on the anime community I suppose. The data file which nobody will bother downloading Just tagging a few people who might be interested : @Zee530 @Shymander @Lunilah @Mythologically @mhkr - Added Madoka, it's ranked 23 Anyways I sold my soul so my days of being active on MAL are over |
<Something> |
Lunilah said: Edit: @-InfiniteLoop-, i wonder if it's possible for you to plug the data in somewhere to be able to compare it with peoples existing lists, would definitely facilitate the idea of finding new anime. I have no idea how to do this, better to ask the creator MAL graphs about it. They already have all the data so it shouldn't be difficult for them to do if they wanted. |
<Something> |
| Mo Dao Zu Shi at 22. Nice. Can't wait to see it's performance for 2nd season. |
| It doesn't matter if you like LoGH,Monster etc.If you are a jobless or college/school dropout living in your mom basement, you are still an unintelligent loser. Taste in anime does not make you a better person. |
| I'd attribute the Geass thing to how divisive it is. There's a bunch of people who think season one is way better and a bunch who think the opposite. Can't tell if I'm surprised about Tatami being in the top 10 or not. I'm kinda surprised by Koe no Katachi being above Kimi no Na wa., if only by a tiny bit. The first curve is quite interesting, in terms of how quickly it drops off. Looks like there are only a select few shows that people truly love since FP drops by about 50% around 25 to 30 entries in. The second curve isn't too interesting; consistent downward trend with some outliers. Same thing with the third curve. |
|
| Yeah this is quite amazing, i love new information. Praise be to LOGH. Edit: @lt_wassile, if you needed anymore convincing about LOGH, post #19. -InfiniteLoop- said: Going to make a note of that then. Lunilah said: Edit: @-InfiniteLoop-, i wonder if it's possible for you to plug the data in somewhere to be able to compare it with peoples existing lists, would definitely facilitate the idea of finding new anime. I have no idea how to do this, better to ask the creator MAL graphs about it. They already have all the data so it shouldn't be difficult for them to do if they wanted. |
LunilahJun 29, 2019 10:23 AM
I don't believe in the Devil. You should. He believes in you. |
| About Gintama does it just include 1st season? How is Gintama performance if we include it's other season. |
| It doesn't matter if you like LoGH,Monster etc.If you are a jobless or college/school dropout living in your mom basement, you are still an unintelligent loser. Taste in anime does not make you a better person. |
Gorochu said: About Gintama does it just include 1st season? How is Gintama performance if we include it's other season. I don't think it's easy to do, probably impossible without access to more data. If you just add up the numbers its gonna be overstated because of people who have multiple seasons in favourites being counted multiple times. |
 |
Are you kidding? Data is life yo! Thank you for the wicked graphs :) |
 |
-InfiniteLoop- said: The data file which nobody will bother downloading @mhkr Anyways I sold my soul so my days of being active on MAL are over I don't know how exactly appreciate your data but really thanks dude, I love this data; though it would be better if you made a new topic for it or at least edit your first first post so people find the new post easier... |
Lunilah said: jajaja hey there!! quite pleasant that u remembered our convo and mentioned me here to see the post, thanks!!Yeah this is quite amazing, i love new information. Praise be to LOGH. Edit: @lt_wassile, if you needed anymore convincing about LOGH, post #19. -InfiniteLoop- said: Going to make a note of that then. I have no idea how to do this, better to ask the creator MAL graphs about it. They already have all the data so it shouldn't be difficult for them to do if they wanted. I've just read the whole post, I was planning to start watching LoGH when I finish what I'm watching now, I guess I'll have to rush things more jajaja looks like a once in a lifetime series :D PD: I love statistics and all kind of s*it like this jajaja |
lt_wassileJun 29, 2019 11:22 AM
| Enjoy the ride... |
| Thank you for that. At one point I calculated the top 50 or so manually for my own amusement (since I don't know coding) but then got sick of it, nice to see someone doing a more extensive one and publicizing it. 14% is crazy, 1 of every 7 people considers LotGH to be one of the best shows they ever watched. I can't wait my current shows to finish just to watch that honestly, it was on my mind for quite some time but now I'm more motivated than ever. |
 |
-InfiniteLoop- said: It's not only meaningful, the only meaningful "best of" lists on MAL are "popular" and "favorited". I completely ignore the numeric ratings, which have zero predictive power for my shit-taste. And your refinement makes that even better. Infinite thanks for that work!1)Is this metric meaningful? -InfiniteLoop- said: Absolutely, favorites is what you want your taste to be associated with. Much unlike tactically loaded and insufficiently agreed on numeric ratings. In your favorite lists you don’t lie - it would hurt your own self-projection.2) Does this mean Steins;Gate is MAL's favorite anime and not FMA:B? -InfiniteLoop- said: By a mile. Rationale: There was a recent suggestion and discussion about calculated ratings derived from ordered lists/trees vs. assigned ratings as used now. I think the former are far superior and fairer. The "favorites" is such a list, it is pretty much the head of a linked list of preferred anime, disregarding the more technical aspects. Pure "like", if you will. I think investing more research into derived ratings will lead to way better rankings and prediction models than the complete assign-number approach.3) Do you prefer this top 10 over MAL's current top 10? |
 |
-InfiniteLoop- said: 1) I think less people put movies as favorites, so yes but actually no.Edit:Updated stats here Someone in the MSP club suggested this and it piqued my curiosity so I did a bit of web scraping and here are the results. The chart has the top 10, the rest are given below. When I say Favorites proportion I mean the percentage of viewers who put the anime in their favorites. As we can see this means that 11.66% of people who watched Steins;Gate put the anime in their favorites. Favorites Percentage = No of Favorites/(Total Members-Plan To Watch Members) *100 Top 30 Collected Data of Top 200 popular anime : http://s000.tinyupload.com/index.php?file_id=83876767409342094795 The data is only of the top 200 anime (excluding 3 which had special characters causing issues while Web Scraping) Observations
Questions 1)Is this metric meaningful? It seems harder to skew this metric with bots than the current mass rating of 1's to a show. It also seems to have an inverse sequel effect for some reason. 2) Does this mean Steins;Gate is MAL's favorite anime and not FMA:B? 3) Do you prefer this top 10 over MAL's current top 10? 2) Favorite doesn´t mean best so YES Steins;gate is MAL favorite anime, but FMAB is what MAL considers the Best anime ever ( Like my I know that FMAB is probably the best thing ever but I still preer CG, SG and TTGL. 3) What I don´t like of the current MAL top 10 is that season and movies of a show are counted deifferently, I would like if the calculated the average of a series and then put the score on the top: for example Code Geass S1: 8.77 + S2: 8.94= 17.71/ 2= 8.85 so it then would be top 25 with Haikyuu!!. That way you get rid of all the Gintama in the top and give some other show more recongition.( Obviusly counting only the cannon stuff to make this avarege. So something like the MHA movie wouldn´t count for the final score, but movies like the LOTGH, or Gintama, or the Kizu movies would count for the final score cause they are cannon. 1)Ginga Eiyuu Densetsu 2)Steins;Gate 3)Hunter x Hunter (2011) 4)One Piece 5)Fullmetal Alchemist: Brotherhood 6)Gintama 7)Aria The Origination 8)Yojouhan Shinwa Taikei 9)Clannad: After Story 10)Code Geass: Fukkatsu no Lelouch 11)Monster 12)Shoujo Kakumei Utena 13)Neon Genesis Evangelion 14)Tengen Toppa Gurren Lagann 15)Code Geass: Hangyaku no Lelouch 16)Cowboy Bebop 17)Mushishi 18)Ping Pong the Animation 19)Versailles no Bara 20Nana 21)Shigatsu wa Kimi no Uso 22)Mo Dao Zu Shi 23)Mahou Shoujo Madoka Magica 24)Death Note 25)Shinsekai yori 26)Code Geass: Hangyaku no Lelouch R2 27)Hajime no Ippo 28)Mononoke 29)Detective Conan 30)Natsume Yuujinchou 31)Kaiba 32)Kenpuu Denki Berserk 33)Koe no Katachi 34)NHK ni Youkoso! 35)Naruto: Shippuuden 36)Haikyuu!! 37)Kimi no Na wa. 38)Kino no Tabi: The Beautiful World 39)Ashita no Joe 40)Princess Tutu 41)Maison Ikkoku 42)3-gatsu no Lion 2nd Season 43)Baccano! 44)Katekyo Hitman Reborn! 45)Made in Abyss 46)Bakemonogatari 47)Yu Yu Hakusho 48)Great Teacher Onizuka 49)Kemono no Souja Erin 50)FLCL Now this I feel is the most accurate thing ever. |
WhgediaJun 29, 2019 2:49 PM
mafd12 said: Very good observations, all worthy of integration into a model. Allow me to add two more I would consider for that “über-list”'s inputs.I've come to some conclusions: 1. The adding of the Currently Watching category change a lot the statistics. (...) 7. Some pretty new shows have a good opportunity for being there, (...) to do a list like this, but for making it fair, the amount of time and job spent would be insane. 8. Popular characters: Otherwise mediocre shows may gain a significant share of popularity from excellent characters (e.g. Rem / Re:Zero). I definitely watch/pick shows just because they star characters being loved much by many people. Shows people pick favorite characters from should also be integrated. Rationale is the same as my previous posting: you want to be associated with them, so you don't lie and let your heart speak, not some tactical voting. 9. Franchises: Lump franchises together. From the Top 30 presented here, I favorited two shows for which I simply picked another season (Monogatari and Mushishi). I always was torn between favoring the best season of a franchise, or the one most others favored (usually the first season). This dilemma could and should easily be put to an end. |
| |
Imaishi said: Gorochu said: About Gintama does it just include 1st season? How is Gintama performance if we include it's other season. I don't think it's easy to do, probably impossible without access to more data. If you just add up the numbers its gonna be overstated because of people who have multiple seasons in favourites being counted multiple times. But then he include Clannad AS and Code Geass Season 2. 1) Not as meaningful as top rated anime on MAL since it's purely popularity only include people who like but not people who dislike like those giving it low rating so it skew the data.It's the same as taking people who rated anime full 10/10 while ignore other rating(most people give their favorite 10/10). This data is not balance. Top rated anime in MAL > this data  2) No, due to reason I mention in 1. 3) Again, no for the reason in 1. |
Papa_ScorchJun 29, 2019 5:57 PM
| It doesn't matter if you like LoGH,Monster etc.If you are a jobless or college/school dropout living in your mom basement, you are still an unintelligent loser. Taste in anime does not make you a better person. |
| Late reply, didn't login due to work and Stranger Things. Anyways my priorities have shifted away from web scraping and MAL for the time being. Mythologically said: I'd attribute the Geass thing to how divisive it is. There's a bunch of people who think season one is way better and a bunch who think the opposite. Can't tell if I'm surprised about Tatami being in the top 10 or not. The first curve is quite interesting, in terms of how quickly it drops off. Looks like there are only a select few shows that people truly love since FP drops by about 50% around 25 to 30 entries in. The second curve isn't too interesting; consistent dovwnward trend with some outliers. Same thing with the third curve. That's a really interesting hypothesis about Code Geass. I personally preferred season 1 overall but season 2 has probably one of the greatest anime endings. I've never watch Tatami before, what's special about it? Great observation about the first chart. As for the second and third, yes a correlation with score when we aggregate the data was the expected result. If there wasn't some correlation that would mean something was amiss. I suppose the outliers are the object of interest, after all we probably care more about the individual records rather than the aggregated data when it comes to anime. In the 3rd chart the variation seems to be really high for scores around 8.6 and above and then it sorts of mellows out Gorochu said: About Gintama does it just include 1st season? How is Gintama performance if we include it's other season. Since MAL treats each season as a seperate entry therefore each season is treated as a separate record, the rest of the data is in the csv file. I'll post the Gintama rank by FP and FP here for you Title Rank FP Gintama 6 10.25% Gintama' 140 2.98% Gintama. 230 1.82% Gintama Movie 2: Kanketsu-hen - Yorozuya yo Eien Nare 240 1.69% Gintama': Enchousen 244 1.63% Gintama.: Shirogane no Tamashii-hen 303 1.15% Gintama.: Shirogane no Tamashii-hen - Kouhan-sen 305 1.1% Gintama.: Porori-hen 312 1.05% Gintama Movie 1: Shinyaku Benizakura-hen 368 0.55% Jesus Christ there's way too many different Gintama shows and this is excluding the specials. Still not as bad as the amount of Detective Conan movies. Anyways whether you want to aggregate all the data by franchise or not is upto you. mafd12 said: Well, I've had this idea in my head for some time, so I've checked this proportion frequently, but with a change from this proposal, I only include Completed and Watching, because I highly doubt that someones who put a show On Hold or Dropped it would put it in their favorites, there must be some cases, but must a very few. Ignoring the dropped section would skew the data in a FP positive direction. Also would skew the data in favor of longer running anime. The same would occur for doing it with On Hold to a smaller extent. mafd12 said: 1. The adding of the Currently Watching category change a lot the stadistics. 3. The movies, OVAs and Specials have a much more lesser chance to get a good proportion. 4. Between less members, it's very probably to have a better proportion between favorites and members. 5. Ironically, those entries which huge number of members also have a higher probability than the usual show, just not as high as those with a reduced number of members. 1)Currently watching is included, ignoring currently watching would cause ongoing long running anime such as One Piece to be excluded. 3)For non-standalone movies, yes that's true. Such movies have less significance towards the overall anime community though so maybe that's for the better. 4)Not necessarily but the same skew would occur for scores as well 5)I don't really agree with this. If a show becomes super popular and also has a high FP then it's probably because it's a good quality show. The high FP would be justified in such a case. This is unlike MAL's favorite system where they only take into consideration the number of favorites skewing the data in favor of high popularity anime. @inim - Thanks for linking that thread, it was a really interesting read @Whgedia - Combining according to franchise would be difficult by automation due to MAL's structure. Distinguishing by canon and non-canon would make things even more complex. |
<Something> |
Clannad AS and Code Geass s2 are like all the rest treated as individual records. Clannad is ranked 97 with a 3.67% FP. Like I said, check the file if you want access to all the data. You're free to aggregate all the shows by franchise if you want. petran79 said: I notice that manga readers have far higher favorites number. 25+ percent share. Don't people like watching anime? What?? I can't tell whether you're being sarcastic here or not. There are a myriad of reasons why your statement doesn't make sense. For starters: 1)The favorite list of anime and manga are mutually exclusive so doing a horizontal comparison is inaccurate 2Many people on MAL do not read much manga, quite a lot have read less than 10 manga 3)There's a large proportion of people who generally only read manga of an anime that they liked a lot, skewing the data in the positive direction (would need to verify this though) |
<Something> |
petran79 said: -InfiniteLoop- said: What?? I can't tell whether you're being sarcastic here or not. There are a myriad of reasons why your statement doesn't make sense. For starters: 1)The favorite list of anime and manga are mutually exclusive so doing a horizontal comparison is inaccurate 2Many people on MAL do not read much manga, quite a lot have read less than 10 manga 3)There's a large proportion of people who generally only read manga of an anime that they liked a lot, skewing the data in the positive direction (would need to verify this though) Fewer numbers do matter and majority of favorites manga on the list do have anime adaptations However I notice that Goodnight Pun Pun, the highest rated manga with no anime adaptation, is sitting at top 10 of favorites and has a rating of 36%! (20.044 favorite votes among 50.086 users). Not an easy manga for newcomers either 55.086k users is not the number taken but the total number who have completed/dropped/hold/reading the manga which would be 91,744. That would give us 22.06% |
<Something> |
| @-InfiniteLoop- I believe snk season 3 part 2 number is impressive as well. 134k users votes with 8225k users have it in their favorite list. Compare to snk season 3 part 1 223k users votes with 4700k users have it in their favorite list. snk season 2 499k users votes with 9600k users have it in their favorite list. |
| Steins:gate I really liked, Hunter X Hunter is OK, the rest of the series I have rarely watched. Also, if a series is more readily available, like if the series ends up on Adult Swim, for example, it may be watched by more viewers and is more likely to be an anime viewers favorite series, (ie Hunter X Hunter on Adult Swim, and I think it's on Netflix also.) |
More topics from this board
» 🖊️ Hall of Sensei: Who Teaches Best! ( 1 2 3 4 5 ... Last Page )nirererin - Sep 23 |
290 |
by Ohayotaku
»»
1 minute ago |
|
» Upcoming Dubbed Anime ( 1 2 3 4 5 ... Last Page )Kenny_Stryker - Dec 17, 2017 |
3948 |
by appledude1221
»»
7 minutes ago |
|
» Waifu War V5 (Anniversary-Edition!) (Nomination Phase) ( 1 2 )TheMinkalex - Sep 28 |
60 |
by Sheol01
»»
19 minutes ago |
|
» What's the point of watching nonseasonal when no one talks about it?Dragevard - Yesterday |
32 |
by iluvmozu
»»
20 minutes ago |
|
» Do you have a comfort character?swirlydragon - Sep 29 |
19 |
by Maou_heika
»»
38 minutes ago |