Print

Please login or register.
1 Hour 1 Day 1 Week 1 Month Forever
Login with username, password and session length

[Orange Wizard]

Yeah, I like the simple "would recommend/would not recommend" system instead of mucking about with X/10 ratings where no-one can agree what they're even rating.

[Flying Dice]

That's not very nuanced though: look at Steam game ratings, the only time they tell you anything meaningful is if the game's a lemon.

[Culise]

The same thing with e.g. having a rating like "originality" or "necessity". You'd get a lot of people claiming NGNL and Mahouka are 10/10 necessities to watch, and I'm not sure that the rating would be useful. 50 different harems that all look the same? Well the girls are all equally cute in all the shows, so they're all necessities right?

I let the idea sit in my head for a bit, and I'm thinking the most practical way to implement a rating system is in something I'm calling the "Top 90" system. Basically, every person has a list 90 names long.

Basically, you have a finite number of ratings you can use to describe any show.

You can give ten shows 10/10
and then ten show 9/10
and then ten shows 8/10
...

and so on until you have a list of 90 shows stretching til 2/10, and the 1/10 section is infinitely long, allowing it to hold every other show in existence regardless of quality.

So basically, if you want to rate something 10/10, you're gonna have to bump every other name on the list down one notch. I'd think that if all your favorite shows are competing against eachother for your favor, which in reality they are due to entertainment being a business and time being finite, then a person's opinion becomes more honest as it is impossible to highly rate every half-arsed show in existence.

It'd also give a better idea of how experienced an anime watcher someone is, if they can't fill out the top 20, then you can safely assume they don't have nearly as well-developed an opinion as someone who is struggling to make a cut to place something into their 2/10 (aka the 81-90th slot) section.

Stretched across many people, averaging would give a better idea of who considers what to be actually worth watching.

I think it seems like a rather poor idea, because you aren't only allowed to enjoy a few dozen shows at all period.  Let's say you only watch and (more importantly) rate shows that you absolutely enjoy, simply on the grounds that life's too short to waste on shows you hate; you don't rate the shows you hate, because you set them aside and never bother finishing them, and you're ethical enough to not downrate shows you didn't finish watching.  Let's say you watch 11 shows.  Suddenly, one of the shows you love gets bumped down, not because of its own merits, but because the rating software simply can't handle more than ten 100% shows by design.  You watch 21 shows?  Suddenly one of those shows you absolutely loved is an 8/10.  31 shows?  7/10, or a C.  41 shows?  6/10, and if you're wondering if there's more than ten 10/10 animes in the world?  There must be at least ten-thousand anime series that have been made in various formats (I got bored and checked AniDB; far from exhaustive, but it has over 7300 records).  Even a cursory application of Sturgeon's Law suggests far more than 30 of those fall in the range from mediocre to decent. 

I'd suggest, if you really wanted to limit positive reviews (though that still falls into the trap of telling people they can only like this many things), unlocking more positive reviews as you fill out the lower ranks.  Then, however, you end up running into the issue of incentivizing the creation of false reviews in order for people to be allowed to give the reviews they really want to give. 

[Execute/Dumbo.exe]

Well, in an ideal situation, you wouldn't be basing the reviews off of themselves individually, but relative to the top 10 reviewed shows, so as long as you knew someone only rated shows they enjoyed, you could always figure out that even something at 2/10 is still alright, it's just not as good relative to the shows topping it.

[Culise]

Unfortunately, that seems like it would completely negates the point of the rating system, which is to pose a single (albeit poor) standard.  Most people don't look at ratings by looking at the user who submitted them, then the show; they look at the show itself to find its rating.  Any aggregator of those ratings is going to find it difficult to track that sort of thing, and when you run into something with, say, a hundred or a thousand ratings, it will not help matters at all if you have to ask for each individual rating whether this "2/10" means "this show is awful" or "I just happened to like more than 70 anime and this one dropped from 8/10".  In other words, for each rating site, you'll need to be familiar with the general watching and reviewing habits of the entire user base, and after that still have to consider the reviewing habits of the people who watched this specific show. 

EDIT:
Basically, my problem with the idea is that it seems to do little beyond adding another point of uncertainty regarding the meaning of ratings, which is the last thing something that's already as uncertain as a rating system based on an arbitrary point scale and as malleable as the enjoyment of a very broad fan base (in terms of tastes and interests) really needs. 

[Execute/Dumbo.exe]

True, but then don't most people have reviewers they normally go to on habit? You know, the ones you agree with most.

[Haspen]

Each new episode of Legend of Arslan is putting it further from 'LoGH in Persia' and more into 'Medieval FMA' territory.

Not that I really mind v:

* Haspen preps popcorn for more episodes.

[Culise]

True, but then don't most people have reviewers they normally go to on habit? You know, the ones you agree with most.

I don't (EDIT: Or rather, I do, but through the power of word of mouth rather than formal site reviews; there's no ten-point scale when talking to someone), but that's just anecdotal evidence, so I wouldn't consider it admissible.  In either case, people who have favorite reviewers are more likely to read actual text reviews written by those reviewers, which are much more nuanced and simply more fun, rather than simply looking at the point scale.  In other words, a fundamentally inaccurate point scale isn't going to affect them as much one way or the other. 

Besides, the whole reason this system was proposed as a hypothetical wasn't because of someone's favorite reviewer blanketing the world in 10/10s.  It was because of the unwashed hordes who have no critical thinking skills who do that. 

[Execute/Dumbo.exe]

True, but then don't most people have reviewers they normally go to on habit? You know, the ones you agree with most.

I don't, but that's just anecdotal evidence, so I wouldn't consider it admissible.  In either case, people who have favorite reviewers are more likely to read actual text reviews written by those reviewers, which are much more nuanced and simply more fun, rather than simply looking at the point scale.  In other words, a fundamentally inaccurate point scale isn't going to affect them as much one way or the other.

Yeah, that makes sense.
Hmm, guess both methods of marking have the problem of subjectivity, don't they?

[Culise]

Yeah.  To be fair, any time you're trying to rate personal preferences on a fixed scale, there are going to be huge subjectivity issues if only because personal preferences are themselves quite subjective.  Ten-point scales tend to have it a bit worse because most people are inculcated for literally decades in a grading system where 7/10 is average and 5/10 is awful, which creates conflict with people adopt a flat linear scale where 5/10 is average and 2/10 awful.  Five-star scales with half-star granularity tend to be a bit better, even if the actual numbers are identical (2.5 stars = 5/10), because it helps divorce the rating scale from that grading system and makes it more likely that the reviewers are using a single standard for what the numbers mean.  Or, at least, something closer to a single standard; you still have to worry about people who, say, calibrate against Mars of Destruction (which I've never heard of, but I assume to be poor) versus calibrating versus something like Gankutsuou (which was decent enough, but had an art style that could occasionally induce headaches). 

Then you run into things like acquiescence bias (the tendency to agree rather than disagree, even outside of peer-influenced biases) or its converse, extreme responding (the tendency to vote to extremes rather than medians), and the aforementioned peer-influenced biases, all of which aren't as easy to eliminate...

[Furtuka]

Wait there's gonna be ana anime for that new monster hunter rpg?


...wat O_O

[JoshuaFH]

Atleast in my mind, the ideal rating system only very loosely calculates based on the quality of a show, as quality is both an already very ambiguous thing to quantify and as more and more things are created the total 'amount' of quality also increases and the definition of quality becomes more and more ambiguous.

My beef is that, not just anime but games, movies, and all other media, are an extremely over-saturated market, and that the ideal rating system calibrates for that by being being much more harsh. My 'top 90' system is honestly very arbitrary, it could be a list as long or short as the person making it wants, but what was important was that even if you're given 100 equally good shows, you're not allowed to like them equally, some have to be cut from your recommendations, and so the matter of refining and honing one's opinion-ating ability becomes an increasingly important task.

[Reelya]

I just started watching Sore no Seiyuu, a behind the scenes look at the role of voice acting in anime, from the perspective of 3 cute girls trying to make it in the industry. First impressions is that this show is trying really, really hard to cash in on the popularity of Shirobako, which is the same deal but about the whole anime production process. Shirobako is one of the top-selling series right now on DVD/BD so you can see why someone might want to clone it.

But IDK, since Sore no Seiyuu comes from a 2011 manga. It could be that after Shirobako became a hit, studios were suddenly interested in adapting similar material. One thing that hammers home how much the new show seems to be copying the format of Shirobako, is that both main characters have some cute+creepy dolls/plushies. And whenever some technical industry stuff is mentioned, both shows cut away to animations of the dolls discussing the issue in question:

Spoiler: Shirobako info-plushies (click to show/hide)

Spoiler: Sore no Seiyuu info-plushie (click to show/hide)

So, you have similar subject matter, similar character set-up, and similar "info-plushies". The main difference so far is that Shirobako has much better artwork (it's by PA Works) whereas Sore no Seiyuu looks ugly by comparison. The verdict would be, don't watch Sore no Seiyuu unless you already saw Shirobako and want a different take on the same industry.

[Reelya]

btw Culise. Mars of Destruction is only 19 minutes long. You should give it a look.

http://myanimelist.net/anime/413/Mars_of_Destruction

It's one of those things that generates hilarious reviews:

Sound: 10/10
The voice acting is absolutely perfect. When the characters yell each other's names repeatedly you will FEEL the emotion. The OST recorded by Beethoven himself is also marvelous. He did an excellent job, as always.

Characters: 10/10
The characters are all equally well-done, from Green Hair Girl to Blue Hair Girl; it's so easy to tell them apart! I especially enjoyed the doctor's dialogue, as I would not have known that the girl whose head exploded was dead without his helpful reminder that "she won't make it."

[Arx]

That was not as bad as I expected, given the reviews.

Admittedly, it could have been fades between slightly-above-averagely drawn comic panels with no voice acting set to dissonant jazz played on a damaged tin whistle and it would have been better than I expected.