| So I’ve seen many threads about everyone’s anime mean score, but none about standard deviation, so I thought I’d be that stat nerd and propose it. Resources Basics of standard deviation Find your standard deviation on MALGraph. Would post screenshots, but at work 1. Type in your username 2. Update your list if necessary (“Click to update” box above your profile picture). Might take a while if everyone does it :/ 3. Click ‘Ratings’ 4. It’s the third number to the right (Std dev) What’s your standard deviation? What would you expect the standard deviation of scores to be within the MAL community? I’ll do a quick analysis of my own, so that hopefully this might make some more sense to those not familiar with statistics: I have a standard deviation of 2.17 with a mean score of 5.88 and a population of 595 rated shows. Practically, this means roughly 68% of my anime (405 shows) falls within one standard deviation of my mean score, and 95% of my anime (565 shows) falls within two standard deviations of my mean score. 1 S.D. = 5.89 ± 2.17 = Scores between 3.72 and 8.06 2 S.D. = 5.89 ± (2.17 x 2) = Scores between 1.55 and 10.23 Of course this isn’t a perfect system to analyze anime scores, since we only use whole numbers on MAL and you’ve got a high possibility of going under 1 or over 10 once you get to 2 standard deviations (like I did), but hopefully we can gleam some interesting info from all this. For example, my standard deviation is probably higher than most, because I am pretty liberal about giving out 9s while also giving out quite a bit of low scores. That said, ~22% of my shows have a 6 score, which balances it out a bit. I wouldn’t be too surprised to find a handful of people with a higher deviation. Overall, I’d estimate the average MAL standard deviation to be between 1.25 and 1.5. I’d also say deviations between 1 and 2 are reasonable and conductive to a good scoring system imho. Hopefully this interests more people than just me and at the very least gives somebody a small math lesson for the day. |
is my standard deviation normal or high or low? |
 |
Salokannel2 said: I'd say a little on the high side, but pretty reasonable. Especially considering how sexy that bell curve is, lol. Most likely just higher cause you don't have as many 10s as a basic stat calc would think. Feel like that might be a reoccurring pattern for many.is my standard deviation normal or high or low? |
| Who cares if two thirds of MaL users rate in a 5* system.... |
| Wait, what? You referring to a 5-star system? I didn't think MAL had that option... Unless you mean that the majority of MAL scores only between 5-10. In that case, standard deviation would still give you results. You'd just have a low value. |
| Related: https://www.reddit.com/r/anime/comments/7tkra3/finding_the_most_and_least_polarizing_anime_on_mal/ Also uses standard deviation, but on the scores for a specific anime not the scores given by a specific user. I'm not sure I totally agree with the methodology, because "polarizing" implies a bimodal distribution that standard deviation alone doesn't necessarily cover (like, there's a difference between a scattered distribution and a bimodal distribution), but it's still interesting to read. |
| Avatar character is Gabriel from Gabriel DropOut. |
| Anyway Total rated: 111 Mean score: 6.66 Std dev.: 2.34 FWIW: 1. I use a 1-9 scale. Actually I use a -2 to 2 scale, then map it to 1/3/5/7/9, then fill in the gaps with tweaks as I see fit. 2. I don't rate everything. (150 unrated items) 3. I rate on my own personal opinion -- a 6 is not a "pass", it is a "I mildly liked it". Also, standard deviation for a user's score is less meaningful than it would be if the scores were normally distributed. But user scores are generally NOT normally distributed. |
| Avatar character is Gabriel from Gabriel DropOut. |
| Total rated 390 Mean score 7.13 Std dev. 2.15 4,98 ~9,28 2,85~11,43 interesting sounds about right, I do have a lot of 8, 9 and 10s and I rarely give something beyond 4 unless I can't find anything positive or I find the show really bad, really interesting indeed |
GlennMagusHarvey said: Thanks for the link fam. Interesting idea at the very least.Related: https://www.reddit.com/r/anime/comments/7tkra3/finding_the_most_and_least_polarizing_anime_on_mal/ Also uses standard deviation, but on the scores for a specific anime not the scores given by a specific user. I'm not sure I totally agree with the methodology, because "polarizing" implies a bimodal distribution that standard deviation alone doesn't necessarily cover (like, there's a difference between a scattered distribution and a bimodal distribution), but it's still interesting to read. |
abhutrash said: Wait, what? You referring to a 5-star system? I didn't think MAL had that option... Unless you mean that the majority of MAL scores only between 5-10. In that case, standard deviation would still give you results. You'd just have a low value. It doesn't, some people just prefer to use this and goes for 1 - 5 or 5 to 10 like @romagia for example |
| Std Dev : 0.89 Mean score : 6.83 Total rated : 289 |
 |
blumenbalt said: Oh damn, I legit didn't know people do this. Interestingly, people using this system could find a comparative deviation by multiplying by 2. Not perfect math, and actual statisticians would shun me for suggesting something like that, but it'd do in a pinch.abhutrash said: Wait, what? You referring to a 5-star system? I didn't think MAL had that option... Unless you mean that the majority of MAL scores only between 5-10. In that case, standard deviation would still give you results. You'd just have a low value. It doesn't, some people just prefer to use this and goes for 1 - 5 or 5 to 10 like @romagia for example |
| Aaaahhhh, Statistics, ahhhhhh, the class I barely passed. God help my soul. |
 𝔚𝔞𝔫𝔫𝔞 𝔱𝔬𝔲𝔠𝔥 𝔶𝔬𝔲, 𝔚𝔞𝔫𝔱𝔦𝔫' 𝔶𝔬𝔲 𝔴𝔦𝔱𝔥 𝔞𝔩𝔩 𝔪𝔶 𝔪𝔦𝔤𝔥𝔱 ______________________ |
-Lofn- said: *insert Obama "Then perish." meme*Aaaahhhh, Statistics, ahhhhhh, the class I barely passed. God help my soul. |
abhutrash said: I mean there are ppl who rate from 1-10 and there are dummies who rate 6-10 and they call it a 5 star system.Wait, what? You referring to a 5-star system? I didn't think MAL had that option... Unless you mean that the majority of MAL scores only between 5-10. In that case, standard deviation would still give you results. You'd just have a low value. |
 Mmm... Sweet and juicy statistics. Looks like I've a lack of 5s. I guess I only give those if I've like no opinion on the show whatsoever (which is pretty rare). Total rated: 315, Mean score: 6.88, Std dev.: 1.77 |
Standard Deviation: 1.87 I'd guess that ~2 is pretty normal SD for users that use the full scale. |
| Total rated 196 Mean score 5.47 Std dev. 1.84 Only says that I almost never use 1&10 score which is obvious to me and almost rate everything 4-7... |
 Looks pretty standard. The ~60 not rated are a bit distracting, though. |
 |
 I could calmly use it as a staircase, although I never minded my average. |
| Total rated : 374 Mean: 6.66 Std. Dev: 1.75 What does this mean anyway??? My score deviates an average of 1.75 from general ratings??? Or is it deviation in my own ratings?? |
| I don't get what is the practical use of this. If this is a way to compare my score with other users why should I care? Should I change my scoring system or what? I have seen people give 1/10 to good shows just because they didn't liked some small detail and give 10/10 just because random shit suited they mood in that exact moment. Why people should always compare to other people and trying to be like them? (Well maybe this is a good idea in sports...but here...) Also why people should look at MAL Graph to learn what genres they like? How can they be unaware their own tastes, how absurd is this? (And this statistic is messed up because of the unweighted genre tags.) |
 Here's mine based on your range I'd guess the deviation is on the lower end. Lanor-sama said: abhutrash said: I mean there are ppl who rate from 1-10 and there are dummies who rate 6-10 and they call it a 5 star system.Wait, what? You referring to a 5-star system? I didn't think MAL had that option... Unless you mean that the majority of MAL scores only between 5-10. In that case, standard deviation would still give you results. You'd just have a low value. I mean 6-10 is a 5 star system, there are 5 numbers. 6, 7, 8, 9, 10. |
GamerDLMOct 1, 2018 1:41 PM
| I'm guessing my standard deviation would be 0 |
I see no problem on my end beside a couple things. |
 |
Hatred said: this is so pointless tho i couldn't care less about my standard deviation or whatever the fuck alshu said: Ah yes, because this thread is so much more pointless than the 30 threads about the same topics we've been discussing over and over on MAL. Cmon guys, it's just banter, no need to be pretentious buzzkills.I don't get what is the practical use of this. If this is a way to compare my score with other users why should I care? Should I change my scoring system or what? I have seen people give 1/10 to good shows just because they didn't liked some small detail and give 10/10 just because random shit suited they mood in that exact moment. Why people should always compare to other people and trying to be like them? (Well maybe this is a good idea in sports...but here...) Also why people should look at MAL Graph to learn what genres they like? How can they be unaware their own tastes, how absurd is this? (And this statistic is messed up because of the unweighted genre tags.) |
Baka_phnx said: Total rated : 374 Mean: 6.66 Std. Dev: 1.75 What does this mean anyway??? My score deviates an average of 1.75 from general ratings??? Or is it deviation in my own ratings?? It's a deviation in your own ratings. I don't know if there's a way of comparing your scores to the MAL scores besides doing it manually. |
changelog said: Gotchu fam.I took math for IB and I still remember nothing about standard deviation. In all fairness, statistics wasn't a very large part of the course. Could you check mine for me, please? I'm pretty busy anyways. I don't think my SD will be very high since I rate most of my shows 8 - 10. https://graph.anime.plus/changelog/ratings,anime 1.77 sounds about right. Surprised you don't have any 3s tho? |
abhutrash said: Ah yes, because this thread is so much more pointless than the 30 threads about the same topics we've been discussing over and over on MAL. I am not against the topic, just don't understand why you guys are using statistics on everything. Please explain! Is standard deviation giving you some important new angle, demonstrating issues or something? |
alshu said: abhutrash said: Ah yes, because this thread is so much more pointless than the 30 threads about the same topics we've been discussing over and over on MAL. I am not against the topic, just don't understand why you guys are using statistics on everything. Please explain! Is standard deviation giving you some important new angle, demonstrating issues or something? Personally I just find it interesting to see other users statistics and compare them. No particularly good use for the data comes to mind although I suppose it can show whether a user is a relatively erratic scorer or something. |
alshu said: It's essentially just interesting. Often rating anime can be a passive process, since we all know our own systems and don't have to think too hard about what constitutes a 6 versus a 7. That said, everyone has a different way of looking at things. Stats like standard deviation and mean can give some insight as to how others view ratings, and I'm all about understanding the perspectives of others. Not because I feel I need to change my own perspective to match, rather it helps me understand them as a person.abhutrash said: Ah yes, because this thread is so much more pointless than the 30 threads about the same topics we've been discussing over and over on MAL. I am not against the topic, just don't understand why you guys are using statistics on everything. Please explain! Is standard deviation giving you some important new angle, demonstrating issues or something? Is there a practical use to data like this? Not necessarily. But there's not a practical use to a majority of general discussion you'd have in a casual environment. Not everything has to be meaningful; sometimes we just make conversation that piques our interest momentarily. |
 My mean score is gradually on the rise since I'm more picky with the anime I watch nowadays. |
 It's an entirely different kind of flying, altogether! It's an entirely different kind of flying. |
| ...okay, I take back my comment about user score distributions; they DO tend to be relatively normal-looking. "normal" in this case being the specialized stats term. |
| Avatar character is Gabriel from Gabriel DropOut. |
GlennMagusHarvey said: Yeah, I'm surprised about this too. ...okay, I take back my comment about user score distributions; they DO tend to be relatively normal-looking. "normal" in this case being the specialized stats term. |
| Jees, don't you have some more interesting stuff to do then this? Why everyone so obsessed with scores? |
| ~4500 total rated 5.86 average score 1.55 standard deviation now what? |
| I probably regret this post by now. |
UrbanSpaceman said: The answer is no; I was stuck at work ¯\_(ツ)_/¯Jees, don't you have some more interesting stuff to do then this? Why everyone so obsessed with scores? |
| I'm already learning this in college, and if you have low Std deviation, your data will be closer to your mean score. And it mean you're most likely to give a rating score to anime/manga closer to the your mean score in mal So, if you have higher std deviation. You're usually give rating score much widely. For example. We take 3 anime to be rated. If your rated 3 of the anime with 5-5-5 or 7-7-7, your std deviation will be 0. But if your rated with 5-6-3, you'll have a std deviation number. Thats all |
 |
| Total Rated - 575 Mean Score - 7.08 S.D - 1.56 1 S.D. = 7.08 ± 1.56 = Scores between 5.52 and 8.64 2 S.D. = 7.08 ± (1.56 x 2) = Scores between 3.96 and 10.24 When rating I normally follow a method that I picked up on from playing a lot of fifa. All shows start at a 6 which is a rating that is neither good nor bad just a rating for shows that don't stand out. As the show progresses the number goes up or down based on what I believe it has done well/badly. As I normally only rate shows I finish and I am not afraid to drop shows if I dislike them enough it makes sense my lower deviation would be just between a 5-6. That was an interesting topic to think about |
BIGMatthew182Oct 1, 2018 4:52 PM
  |
abhutrash said: Overall, I’d estimate the average MAL standard deviation to be between 1.25 and 1.5. I’d also say deviations between 1 and 2 are reasonable and conductive to a good scoring system imho. Mine is 1.05 Makes sense since my mean score is 6.85 Am I too nice? |
 |
| Total rated: 514 Mean score: 5.58 Std Deviation: 1.77 1 S.D. = 5.58 ± 1.77, between 3.81 and 7.35, 375 or 73.0% of anime with 1 S.D. 2 S.D. = 5.58 ± 2(1.77), between 2.04 and 9.12, 488 or 94.9% of anime within 2 S.D. abhutrash said: It probably is. Having numbers not too unreasonable will definitely make for a better scoring system.Overall, I’d estimate the average MAL standard deviation to be between 1.25 and 1.5. I’d also say deviations between 1 and 2 are reasonable and conductive to a good scoring system imho. Hopefully this interests more people than just me and at the very least gives somebody a small math lesson for the day. Still though, I think mean score is still really important and S.D. cannot be full read without taking that into account. Don't mind me taking you as example, but your best rated anime is rated within 2 standard deviations of your mean. That means your upper range is not as wide as your lower range, which could mean you don't have much space to spread out the highly rated anime. In other words, it could be that your higher scores of 9 and 10 are not special enough to put anime that you feel are truly special to you. A low standard deviation with a high mean could mean that you bump too many anime up from the lower ranges to a 6 and 7. Making those score ranges just a dumping ground for most anime. You can think of it being kind, but you can also think of it as cruelly lumping good anime with or near anime that should really be in a lower score range. I prefer to think of it as the latter. |
BurningSpiritOct 1, 2018 5:22 PM
| I'm not a lolicon, you're just projecting your tendency to lewd 2D characters. If your favourite character is Tsutsukakushi Tsukiko, you are my soul mate. Been a long time since I've been here, I'll continue expressing myself freely and believe everyone should too. My MAL Interview |
UrbanSpaceman said: Stats are fun.Jees, don't you have some more interesting stuff to do then this? Why everyone so obsessed with scores? abhutrash said: What about standard deviations greater than 2?Overall, I’d estimate the average MAL standard deviation to be between 1.25 and 1.5. I’d also say deviations between 1 and 2 are reasonable and conductive to a good scoring system imho. |
| Avatar character is Gabriel from Gabriel DropOut. |
My SD is 2.21. Makes sense. I have 118 anime rated as of this time. That's a small ass number so my 10s, 9s, 2s, and 1s have a larger effect on that calculation. At 500 shows, I expect to be entering the 1.xx SD simply because I'd probably have a lot more shows rated 4-7 compared to what what few shows I'd probably rate 10, 9, 2, and 1. |
 You gave up your freedom of speech when you clicked Agree to the User Agreement This is not a public platform. My gaze is the measure of all things: I stopped considering "anime" a helpful tag Recommended Essays Exploring Actually Excellent World-Building |
BurningSpirit said: A low standard deviation with a high mean could mean that you bump too many anime up from the lower ranges to a 6 and 7. Making those score ranges just a dumping ground for most anime. You can think of it being kind, but you can also think of it as cruelly lumping good anime with or near anime that should really be in a lower score range. I prefer to think of it as the latter. Suppose that depends on what you call "good". Different people have different tastes. I do try to evade watching shows I most likely won't enjoy. Anime isn't my job, merely a hobby. |
|
I haven't updated my MALGraph in about a month, but I don't plan to since I rather finish the seasonals that I already started before doing so Total rated 1088 Mean score 5.37 Std dev. 2.30 |
 |
More topics from this board
» Upcoming Dubbed Anime ( 1 2 3 4 5 ... Last Page )Kenny_Stryker - Dec 17, 2017 |
9068 |
by animeboilolz
»»
4 minutes ago |
|
» Anime Misandry ( 1 2 )ColourWheel - Apr 21 |
91 |
by Zarutaku
»»
19 minutes ago |
|
» Characters you like but everyone else hates.Spunkert - May 17, 2023 |
48 |
by Styx_Thistle
»»
20 minutes ago |
|
Poll: » Bare feet or pantyhoses?Absurdo_N - 12 hours ago |
24 |
by Spunkert
»»
22 minutes ago |
|
Poll: » Best kind of relationship in anime?IpreferEcchi - Apr 22 |
15 |
by Spunkert
»»
25 minutes ago |