| I would love to give 8.5 or 9.5 scores. |
 |
Sphinxter said: I'd rather have a scoring system not based on an absolute number but simply ranking everything you know from best to worst. Care to explain? I don't think that would work, without numerical values. |
Optigisa said: Why not? Based from this the system can automatically compute an average percentile score.Sphinxter said: I'd rather have a scoring system not based on an absolute number but simply ranking everything you know from best to worst. Care to explain? I don't think that would work, without numerical values. As in "the average user considers this particular thing better than 62% of the media he knows". It automatically normalizes it based on this. |
 It is obvious that "obscenity" is not a term capable of exact legal definition; in the practice of the courts, it means "anything that shocks the magistrate". — Bertrand Russell |
| First off what I dislike at the 1-10 scale is, that you can't rate with the median. An 11 point system (0-10) would be far more logical, so 5 would actually be the average. |
MandarBloodeagleJun 26, 2019 6:55 AM
 |
Sphinxter said: Why not? Based from this the system can automatically compute an average percentile score. As in "the average user considers this particular thing better than 62% of the media he knows". It automatically normalizes it based on this. Seems complicated and pointless. I would rather stick to a numerical scoring system. Also, a non-numerical system that would probably work would be "Great", "Good", "Masterpiece" though I don't know how that will accumulate a mean score as well without the use of numbers. |
| I would like decimals too, since Mydramalist uses that. There are some series I'd like to rate it with a .5 |
 «  »🔮 Tojo NozomiWashi Washi |
sunsetocean said: I would like decimals too, since Mydramalist uses that. There are some series I'd like to rate it with a .5 There is a site called Mydramalist? I have to see this for myself |
| I would keep the default settings for users in the scoring system as it is, but extend the options to be able to go and use .5, use quarter increments or use tenth increments and separate options of using the 0 score |
 |
Optigisa said: sunsetocean said: I would like decimals too, since Mydramalist uses that. There are some series I'd like to rate it with a .5 There is a site called Mydramalist? I have to see this for myself yeah that site's for Asian dramas/movies. I just watch mostly k-dramas though, they take up more time than watching anime (usually 1 hr per episode) the longest I've seen was 1 hr 50 min for some Thai dramas lmao |
sunsetoceanJun 25, 2019 2:27 PM
« »🔮 Tojo NozomiWashi Washi |
 30count30count30count30count30count30count30count30count30count30count30count30count30count |
 |
| The simple implementation of .5 scores would do wonders for my scale. Using 10 whole numbers for hundreds of shows takes away the nuance in how I try to gauge my overall enjoyment. |
 |
| I WANT A TWO ROUND DECIMAL SYSTEM, JUST LIKE MAL, THAT WAY THE JUDGEMENT IS PRECISE EG I RECENTLY SAW ERGO PROXY, I WANTED TO GIVE IT A SCORE 9.12 , HOWEVER I COULD ONLY GIVE 9 OR 10, AND THE ANIME IS TOO ELITE FOR GETTING A 9.00 SCORE , BUT NOT THAT GOOD , THAT IT RECEIVES A SCORE OF 10 DECIMAL RANKING WILL ALSO UPLIFT THE OVERALL SYSTEM OF MAL RANKINGS |
 |
| @_9_, i don't know if you're still replying to this thread or not, or if someone has mentioned this already. In a practical sense and otherwise the 1-10 shouldn't be changed, it's as close to as ideal you can get and as a result it's a staple of scoring systems, a percentage already exists within the aggregate of 1-10. The Japanese version of MAL called Anikore has a 2 tier system that shows the average of all reviews and overall score (everyone). All scoring uses a 5 star system but it also includes 5 main categories; Story, Art, Voice Acting, Music, and Characters then it averages satisfaction out of 5 stars automatically. The average of all reviews is within those 5 stars including each category but they calculate overall score blanketly to an out of 100 points with decimals. Rotten Tomatoes has the same 2 tier system reversed, critics (reviewers in Anikore's case) is 1-10 and audience is out of 5, but it also has the Tomatometer for critics and Popcornmeter for audiences. The former acts as a percentage for how many approved critics gave something a positive review, and the latter for the percentage of users who rated something positively. But to note both the Tomatometer and Popcornmeter have 2 categories each; All Critics/Top Critics and Verified Audience/All Audience. To answer the question: If i could choose i would merge something like Anikore and Rotten Tomatoes. They both have 2 tier systems which i think MAL definitely needs and can implement without changing/ruining anything as they already have a reviews section, however i'd add a would or would not recommend regardless of score result, perhaps replacing the Tomatometer as a percentage for recommending. I would also keep 1-10 universal for simplicity and do away with Anikore's overall score out of 100 points and RT's out of 5 for audiences. Adding the 5 separate metrics Anikore uses i think would be ideal, but i would add animation to it or conflate it with art so it would at least be Art & Animation for specificity reasons. Lastly i would definitely add requirements to being allowed to review something, like a character requirement and maybe some other things. Edit: Just wanted to note that Anikore uses a 9 point scale since they use .5 decimals in their 5 star system, 9 because they don't include 0.5 and start from 1. |
LunilahJun 25, 2019 2:00 PM
 I don't believe in the Devil. You should. He believes in you. |
| My apologi Windrage said: Joe_Bloggs402 said: Mathematically, yes. But when using a 3 star, you usually mean that it is good, worth your time, etc. The sentiment lines up with a 7. A 4 star is something that excels. Not to the extent of a 5 star, but something that surpasses expectations. These are not descriptors I would give to a 7. Actually, it all depends on how you see every number. 3 is in the middle of the 1-5 rating system, so that's why for me for me 3 stars means average, in others words, not too good but not necessarily bad either, just okay for the most part. Now let's take a look at the 10 point system and see which number would mean average: 5.5 is in the middle so the average would be 5 or 6. I've already talked about it above. If you look at things this way than you can definetely rate an anime that excels with 7, because 7 is already above average. My apologies. I've been using a localised phrasing for "average" whereby it kind of means "bellow average". I suppose it all depends on the personal ratings systems of the person doing the review. What I stated lines up with a lot of what I've observed, but your system makes logical sense. |
  |
Sphinxter said: I'd rather have a scoring system not based on an absolute number but simply ranking everything you know from best to worst. That would be neat! Well, I'm basically doing the almost same thing with numbers for sorting purposes. For the rest I don't care much about, I just throw a five-sided table top-RPG dice and see what happens. ;) |
My Rereads | My Favorite Mangaka | Soylent Green Buy stuff, or buy at least your favorites if you can afford it - Don't be lame. D W-W&N 🤍 MAL is a far cry from being my main literature site, that's why I won't be doing mere single chapter updates anymore. Manga- aka Baka-Updates + 2 French ones are miles better and more thorough/complete + less restricting. The latter are perfectly suited for Manga/Light Novel/Books collectors such as myself. Less stressful/tiring this way. -g- |
| I'd like better a 1-5 stars system, I find that easier to use. |
| Biggest thing that bothers me is the descriptors for the numbered ratings. If five is average then is a 4/10 really bad? I don't know if something barely below average deserves that label. I think a bad anime would be more like a 3/10. People complain about how people prefer to give high scores, yet the wording for the ratings is so negative below five. I would prefer something with slightly less tiers and less negative labels. I would do away with an 'average' tier as it doesn't make sense to me. An anime is good or it isn't, with varying degrees either way. When people say something is completely average they often mean bland and generic, which is 'bad' to me, certainly not 'average'. I'm probably in the minority but I only need three tiers of 'less than average'. 3 for bland and dull, 2 for actually bad and disappointing, and 1 for highly lacking across the board. This would work best for me: 8 - Masterpiece. 6 - Excellent. 7 - Great. 5 - Good. 4 - OK. 2 - Disappointing. 1 - Highly Lacking. |
| “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 |
| 5 Star system is more than enough This could be expanded to be 5 panties, or 5 Miyazaki heads, or maybe 5 OK hand signs to give special weight to the rating if the user wants to. |
| I usually perform a ritual sacrifice so that I can talk to gods and ask them for a score And what ever they tell me I write. |
 |
| I want decimals so I can give Steins;Gate 0 the 0.1/10 score it deserves |
  |
karkki_ said: I want decimals so I can give Steins;Gate 0 the 0.1/10 score it deserves Okay, so definitely no decimals for you xD |
 |
| You now how in some games there is this class D and than C right up to SSS. That would be cool. There are also titles for military, nobility and royalty which can from, peasant to king, cadet to general or mayor to Duke. Let's not forget that so far we have been using the cardinal scale with 10 being the best score, why not use the ordinal scale and 1 being the best score. Rating by genres might be fun, which genre would be highest, since an anime sometimes consists of at times five genres. And my last idea is rating by separating new from old, an anime like Boogiepop 2000 and Boogiepop 2019, which is better. Same with seasons, sequels, prequels, alternative settings, versions. |
 |
| I feel like 1-5 is too condensed to rate over hundreds of shows. Whenever I want to rate something I finished I encounter a dilemma.Let me give you an idea via an example. I might give a show like Highschool DxD like a 8/9 because i enjoyed it and at the same time if I encounter a well-made anime like Ergo Proxy I have to give it a 8/9 because it deserves that much amount of respect.The two shows are in different dimensions and it makes no sense when both end up with the same score,i mean like you can't put a comparison between them in the first place. My preference/solution would be to be able to have the option of two(maybe more) lists(which could be 1-5 rating based) and sort the shows by rating in their individual lists.I feel that it would be neater if i could be able to say that Messi is my favorite soccer player and Lebron is my favorite basketball player instead of trying to assert that Messi is better/worse/as good as Lebron. It's two different sports people. I would implement this by having one list for solely enjoyment and binge-factor as criteria and the other where i critically examine the show's elements. Does anyone else feel the same way? Let me know :) |
 |
| Unlike so many feel fags who only score anime on enjoyment. I score it based on character, plot , character development, art style, animation so when I give a score to anime it is ABSOLUTELY objective and it isn't up for debate. |
 |
| I score anime based on my personal enjoyment there are no strict rules. |
 |
| Any system with >= 20 steps. We are used to percent, so 100 steps seems natural. The interval it's normalized into doesn't matter (1..100, or 1..10 with one decimal is just the same). Rationale: I frequently have anime to be "on the edge" between two ratings, usually because they are well made but from a genre i do not like, or not so well made but I like the genre. This are typical dot-fives to me I currently can't express. Having too many steps adds too little for too much confusion. |
 |
FoxFire75 said: I share that uncanny feeling for both the 4/10 and by extension 6/10 ratings described by @FoxFire75. My solution is to forget the official worded labels MAL uses and use own wording.YossaRedMage said: That's an interesting point and having thought a bit about it I think I really am hesitant to give 4s because of that. I tend to think in words, not numbers when ratingBiggest thing that bothers me is the descriptors for the numbered ratings. If five is average then is a 4/10 really bad? Coarsely speaking 10-9 = Genre bending/innovating and meeting personal taste and emotions 8-7 = Good and very good. 4-6 = Average, either slightly above or below. Most anime land here. 3 = Either exceptional camp / trash, or boring stuff which failed to be a 4/10 2 = Camp and trash 1 = Annoying |
| |
Sphinxter said: +1, relative ranking is the shit. Regrooming the list while it grows using that method is always an option, and if you try to keep a normal distribution doing it, it leads to numeric changes occasionally. No rating is forever.I'd rather have a scoring system not based on an absolute number but simply ranking everything you know from best to worst. "I don't know everything, I just know what I know." (Tsubasa Hanekawa) |
| |
| I don't have any problem with the currest scoring system but I would prefer using the percentage,also the score with the stars from 1 to 5 is good |
| Signature removed. Please follow the signature rules, as defined in the Site & Forum Guidelines. |
FoxFire75 said: inim said: The goal should be to find a system everybody can agree on and so would be used by everybody the same way, as in using the whole range of stars, points, percent or whatnot. But that might be impossible imo, so there is no real solution, is there ?I share that uncanny feeling for both the 4/10 and by extension 6/10 ratings (...) wording. I don't think there is a real solution, you can't herd cats. If there was one, it had been long patented. The alternative is to apply chinese democracy and make them behave. The most widely used fix to the dilemma probably is using the impressively fine tuned magic data-mining and stochastics offer. You can normalize even the weirdest user's (ok, all 10 is pathological I guess, but you know what I mean I hope) deformed normal distribution. You can work with cutoffs for the worst and best ratings. You can give weight to users based on past rating behaviour and distribution. My personal favorite would be if we can select a peer group (friend list can do that, but is destroyed by it for the original purpose). Then, you can view ratings limited to only those of your friends averaged to your individual view. I just select friends based on a compatible rating approach, that is the trick. This is my kluge currently, and it's predictive power for anime I probably will like is amazing. But of course this all required the return of the official API so you can write the code doing it yourself, or integration into the official functionality of the MAL site. Keep your fingers crossed, I certainly do. |
| |
| MAL rating is fine. I use it as personal rating anyway, so who cares how I score something. Also do not understand people who say they need 0.5 system. Why not use the scale 1-20 then? A 8.5 would be 17 then. For me personally the best rating would be something like a layered tree. It has no hard scale (like 1 to 10) but instead it has a depth which can be infinte. I can also put anime I think are equal in rating on the same level. If you for example watched FMA and think this is the best thing you've ever seen so far, then it is your 10 on MAL currently. In my "tree" it would be the very top of the tree. But then some time later you watch something AOT that tops FMA, then you would put it on top so it becomes the best anime. Or you choose to place it on the same level as FMA. Same goes for every level of the tree. I often feel that anime I watched have almost same enjoyment level, so I end up rating them equally, but they are not quite equal. In a tree I would place it just below the other anime. Using a 100 scale is too oversized and I would have trouble to givea show, let's say a 88 instead of 87. |
FoxFire75 said: RandomFriday said: For me personally the best rating would be something like a layered tree. It has no hard scale (like 1 to 10) but instead it has a depth which can be infinte. I can also put anime I think are equal in rating on the same level. So like show A is the 17th best of the 300 shows you've watched, show B is 17th best too and show C is 24th best of 300 and so on ? Can't that be expressed in a numerical value as well ? Yes you got the idea. The difference is that the layers are not fixed, there is no hard scale. You can always place a show D that is 18th best of the 300, this would be just under A and B, but push the C to 25th place. You would have watched 300 anime, but maybe only 30 layers, since a lot of them share the same layer. The tree would have an average depth of 30 and average width of 10. |
| Mine is the best, everything else is crap. |
FoxFire75 said: I like this method, it's a refinement of my prefered way to appraoch hard rating problems. These family of methods has proven wildly successful in prioritizing and weighting tasks in software engineering, e.g. as Planning Poker in the Scrum method.RandomFriday said: You have 146 completed shows, I got 260 ... so your 130th best is quite bad compared to my 130th best which would mean average. The difference is that the layers are not fixed, there is no hard scale. You can always place a show D (...) just under A and B, but push the C to (...) The tree would have an average depth of 30 and average width of 10. So in the end we are back to percentages. I don't think if you have that beautiful tree data structure deriving meaningful numerical equivalents over all users is hard. The main difference here is that the rating displayed will be calculated, not assigned. And no two users sharing an identical "completed" list isn't a real issue either, because you can limit the view to those who have seen a particular show, and normalize the number of shows they have seen (and/or the depth of the trees etc.) for a good numeric equivalent. The one problem I see here is the ever same, people will try to play the system. They will want (and probably find) a way to upvote their favorite and hurt the shows they hate. That's very human and a problem to any rating system. They will not love the limit that they can "just order stuff", but have no way to "say what is right". |
| |
FoxFire75 said: RandomFriday said: Yes you got the idea. The difference is that the layers are not fixed, there is no hard scale. You can always place a show D that is 18th best of the 300, this would be just under A and B, but push the C to 25th place. You would have watched 300 anime, but maybe only 30 layers, since a lot of them share the same layer. The tree would have an average depth of 30 and average width of 10. I can see how that would work for each person's list - but if you want to compare them you still need to express your 14th best of 300 as a number, right ? You have 146 completed shows, I got 260 ... so your 130th best is quite bad compared to my 130th best which would mean average. So in the end we are back to percentages. If you want to express globally you would have to check of the depth/layer of an anime for all users. Yes in the end you need to introduce some kind of fixed scale, since variable depth of every single user cannot be translated in one common tree with variable depth. Lets say your tree is 30 levels and anime A is on level 9, then effectively anime A is 9/30=0.3 from the top. Where 0 is top level and 1 is lowest level. For another guy this anime could be on the level 35 of 50 which is 0.7. So you sum up all "scores" (which is in the end the relative position of the anime in users tree) and create the average. There it would be (0.3+0.7)/2 = 0.5. Now you can interpret it however you want. You can multiply this by 10 to get MAL scores, so the anime A will now have the score of 5 (0.5*10). I think by having this tree it is easier for me to maintain my list. The relative position of an anime in my tree determines its "rating" for the absolute scale of all trees. |
| No matter what scoring system you implement there's still gonna be people who are not gonna use it fully. |
 |
RandomFriday said: That sounds like the way things are now since we can only rate based on our own experience. Can't compare experiences explicitly since you can ever only know your own.FoxFire75 said: I can see how that would work for each person's list - but if you want to compare them you still need to express your 14th best of 300 as a number, right ? You have 146 completed shows, I got 260 ... so your 130th best is quite bad compared to my 130th best which would mean average. So in the end we are back to percentages. If you want to express globally you would have to check of the depth/layer of an anime for all users. Yes in the end you need to introduce some kind of fixed scale, since variable depth of every single user cannot be translated in one common tree with variable depth. Lets say your tree is 30 levels and anime A is on level 9, then effectively anime A is 9/30=0.3 from the top. Where 0 is top level and 1 is lowest level. For another guy this anime could be on the level 35 of 50 which is 0.7. So you sum up all "scores" (which is in the end the relative position of the anime in users tree) and create the average. There it would be (0.3+0.7)/2 = 0.5. Now you can interpret it however you want. You can multiply this by 10 to get MAL scores, so the anime A will now have the score of 5 (0.5*10). I think by having this tree it is easier for me to maintain my list. The relative position of an anime in my tree determines its "rating" for the absolute scale of all trees. |
I don't believe in the Devil. You should. He believes in you. |
Lunilah said: The difference is: you can not rate, only order. Moving one new item up means to push another down. There always is a cost, the law of equal exchange in full effect ...That sounds like the way things are now since we can only rate based on our own experience. |
| |
inim said: I understand that, but it's like Friday said you still have to introduce a fixed scale eventually. Even if you don't you can still calculate one.Lunilah said: The difference is: you can not rate, only order. Moving one new item up means to push another down. There always is a cost, the law of equal exchange in full effect ...That sounds like the way things are now since we can only rate based on our own experience. |
I don't believe in the Devil. You should. He believes in you. |
Lunilah said: inim said: I understand that, but it's like Friday said you still have to introduce a fixed scale eventually. Even if you don't you can still calculate one.The difference is: you can not rate, only order. Moving one new item up means to push another down. There always is a cost, the law of equal exchange in full effect ... Yes you introduce the fixed scale to analyze and scale ALL trees. But all individual trees of the users do not have any inherent score. The system evaluates your tree and those of all other users and creates a scale - the overall scoring system. You have to visualize the data somehow. The easiest and most understandable way is to have a fixed sclae of scores, like MAL 1-10, or giving 1-5 stars, or having 8 bananas or w/e. The solution I proposed just takes away the pain and responsibility from the user to actually give an anime a score. Well at least I have sometimes trouble to give anime a certain score. You as user just simply order your anime. |
More topics from this board
» Do You Consider Yourself a Generous or a Harsh Scorer?Alpha_1_Zero - 2 minutes ago |
0 |
by Alpha_1_Zero
»»
2 minutes ago |
|
» If MAL didn't existAlpha_1_Zero - 37 minutes ago |
10 |
by eblf2013
»»
3 minutes ago |
|
» If your favorite anime exploded in popularity to the point that everyone in socmed talks about it, would you like it or not?Rinrinka - 1 hour ago |
12 |
by Ratris_Decision
»»
4 minutes ago |
|
» What anime song (intro, outro, j-pop) of the ringtone do you use right now?Kiryotsu - 9 hours ago |
19 |
by Zimmu
»»
7 minutes ago |
|
» Visual Novels — General Thread ( 1 2 )Shizuna - Yesterday |
52 |
by -YaoiBoy-
»»
23 minutes ago |