Some possible suggestions to help reduce the negative effect of 1/10 ratings

[zStella]

Hi, I just had some ideas for how to reduce the effects of people giving unjustified 1/10's on series.

1) For people with a history of negatively voting series (maybe 70% or more of their ratings are all 3 or lower and they've rated like, 30+ series), they should have their ratings silently excluded. It would look to them like they've rated, but other readers wouldn't see their ratings counted. (The percentage and rating score requirement is up for tweaking of course)

2) Similar to 1), exclude low ratings from people who haven't even read a single chapter (or a decent portion of one). A lot of 1/10 voters don't even read the actual series; they simply throw on a 1/10 rating because it's tagged as "isekai" or "smut" or they don't like what they read in the description of the series. (Some have even shamelessly admitted to doing that before, which really baffles my mind)

 

[justforthelulz]

I think censorship of votes is not a good idea. There really are a lot of mediocre and crappy manga/manhua/manhwa/etc. out there and on MangaDex.

As for requiring a minimum of one chapter read before voting, I guess it would somewhat reduce the kneejerk 1/10 ratings,
by making it slightly more inconvenient by having to click a few extra times and some scrolling, but in the end, what is this all for?

 

[zStella]

@justforthelulz The idea behind 1) is that it specifically targets problematic raters who have a habit of rating series poorly. Again, the variables are up for tweaking; it could be raised to 80% or 50+ series, etc. to reduce false positives or to give people the benefit of the doubt

Essentially, if somebody is constantly bashing on everything they read, they're probably not providing constructive ratings. And if all they ever read are bad series, something is clearly wrong lol

 

[justforthelulz]

@zStella
It's not uncommon for some readers to leave overwhelmingly negative comments in multiple series within the same genre (i.e. isekai or harem, which have a reputation for being incredibly derivative).
I'd say Bile Fascination has gripped them particularly hard.
And, if we're being honest, I don't necessarily always disagree with their points. I, too, enjoy reading the occasional trainwreck/shitshow/clusterfuck.

 

[MrIncognito]

One feature already on the way is requiring a title be put on one of your lists before you can rate it, the concept being that people who would rate titles low for trolling reasons or for not being even remotely in their demographic would balk at the idea of having every single one of those on their account

 

[MavB_Ver]

Great, now do it to reduce 10/10 ratings, to balance the extreme ratings. Fair enough, right?

 

[vinsegas]

Yeah, as the one above me says. We need to reduce the 10/10 ratings to. Its not only a problem for 1/10 ratings.

 

[Holo]

The Bayesian average is intended to reduce the impact of 1-bombing a title.

 

[BzzBzz]

> Everybody always rates on the 5 point scale: 1, 5, 7, 8, 10.

That's perfectly understandable.
1 - I don't like it
5 - average boredom
7 - good
8 - good, a bit better than 7
10 - masterpiece

😇

 

[AB2000]

Rating can be tricky indeed.

There are some people rating every series with a 1, and even more people rating every series from 5 to 10.

When there are thousands of readers, then a bunch of pranky people can't really change the rating much. But since lots of people just give good ratings to everything, then the rating is also difficult to statisticaly interpret.

When there are just a few readers, individual ratings have more of an impact. But, hey, what should we think of a series when there are, say, only 17 readers and 3 ratings? Even if they all say they hate it or all say they love it, statistics on such small numbers aren't very significant.

Moreover, ratings only show opinions of registered users of who took the time to rate. Unregistered users can't rate at all (maybe they rate the same series on Manganelo or other websites, though, but then it's a different data set).

In the end, rating is an extremely rough estimation of how people like a series or not.
So I think we should leave it as it is. Statistics are not a very trustworthy mathematical tool for opinions, because that's the essence of statistics. Just know that ratings shouldn't be trusted too much.

 

[schlo]

@AbyssalMonkey it's not just that the average is 7, there are only 24 titles rated at under 5 using their Bayesian rating. Honestly I would go with a 1-2 or 1-4 rating system, force people to make a hard choice.

Admittedly I don't rate because I consider it mostly useless but do find the issues with ratings and statistics in general interesting.

@MrIncognito do they know how many seriously use their MDList? I quit using mine and the number I am reading or "waiting for the next chapter" probably out numbers my mdlist by 30 to 1. I assume they aren't tracking if I have read it since it doesn't show up to me as a read chapter on manga that I don't have marked as follows. So would I and others not be allowed to rate? What if I am reading it on another site, would I still be able to rate?

@zStella
It may be that some of those people only vote on things that they have strong feelings about. I'm curious, how do you think people should rate? Base on enjoyment? Perceived quality? Something else? I like isekai trash. However if I were to rate it on quality I would rate many that I enjoy as probably twos and threes but if it were enjoyment probably 7+/-. On the flip side of 1's I can't believe the number of people who rate stuff as 10 even if it is not a common rating for them. I can't think of any that I would rate as a 10 with the possible exceptions of few oneshots.

 

[MrIncognito]

@schlo
If you don't have it on any follow lists, rating a manga would automatically add it to "followed"

 

[schlo]

@MrIncognito ok, makes sense. I don't rate so haven't had it happen.