Proving that the Star System is useless

[Deleted member 84247]

I posted this in discord first, but I decided to do it here. Yes, I made up all of the numbers using intuition, bite me.

I have to say this somewhere, or it will never leave my brain. I am willing to guess that you can mathematically prove that the star system on SH is useless.

In statistics the average is the sum of all variables in a data set, divided by the amount of variables. In a 5 star system you can calculate that the average is 3

(1+2+3+4+5)/5 = 3

Now, keep this in mind. The median (middle number of ascending values) is also 3. But I will use intuition and thought experiment based off numbers I imagined.

I assume most novels mostly get a 5 star rating. The most common occurring number in a data set is called the Mode. Anyway, I bet if you calculated all of the numbers for ratings on SH you would get a median of 5 (Based on a sample like this: 1, 1, 2, 3, 4, 4, 5, 5, 5, 5, 5, 5, 5, 5,). The mode would also be 5 because it's the most occurring number, and the average would be closer to 4. You could make the statement that "most novels on scribble hub are better than average" and be correct but dishonest.

 

[RepresentingWrath]

Imagine I posted that one crying emoji.

 

[StringSkeleton]

steam has a thumb up/down system which i think is superior.
There cant be grade inflation if the only options are wether you recommend or not

 

[LilRora]

You could make the statement that "most novels on scribble hub are better than average" and be correct but dishonest.

This is why it is there is evil, there is greater evil, and then there's statistics.

None of this is correct, but statistics make it look like you're making sense. To speak of novels, you would need to collect fractional ratings from all books, not the ratings that make those up, and even then it would hardly mean anything because ratings are subjective.

 

[Deleted member 84247]

This is why it is there is evil, there is greater evil, and then there's statistics.

None of this is correct, but statistics make it look like you're making sense. To speak of novels, you would need to collect fractional ratings from all books, not the ratings that make those up, and even then it would hardly mean anything because ratings are subjective.

That part you quoted was a joke. The true statement would be "most novels on scribble hub are better than the average of a 5 star rating system." It's still dumb to say though

 

[SRB]

I like your funny numbers, vampire girl.

 

[Deleted member 84247]

I like your funny numbers, vampire girl.

To be fair, this is dumb. A real analysis would take polls of people who rate stories. There are people who don't rate stories and those who do. But I think the star system is unintuitive exactly because you can do this.

 

[beast_regards]

Are you trying to tempt me to mention that site again?

The problem with the five star rating system is that the readers are more inclined to select the story within the range of 4 to 5 stars as opposed to 3 to 5 stars. (or 2.5 to 5 stars in the system that allows half-star ratings)

This means that the system offers more options for a negative rating than for the positive one.

 

[MatchaChocolate69]

the star system on SH is useless.

thumb up/down system which i think is superior.

Based

 

[Tyranomaster]

I said it before, and I'll say it again, the "average" value rating is subjective, but more focused around what people would consider an average story, not a story by someone trying to write a story. If you look at local McDonalds on google reviews, you'd expect something between a 2.5 and a 4. A 2.5 is technically average. What that really implies, however, is that the experience was on par with eating the average person's cooking, not eating at a place you paid money for.

It doesn't do a great job of informing the average user, but it does give some amount of valuable feedback. I'd expect anything above a 4.2 to be readable with minimal grammatical and spelling errors. Something at a 4.8 or 4.9 I expect to be enjoyable even if it's not my normal genre choices. Other than that, it's more a gauge of total audience appeal, rather than a hierarchical ranking of individual prowess. Below a 4, and you're likely to have some number of poor story telling aspects that draws people out (spelling, grammar, plot holes, or so much more). It's an average, therefore it is representative of the average reader, not representative of an individual reader. So you, as an individual can't actually gauge the amount of enjoyment you'll get from an average that you deviate from.

In other words...

 

[Deleted member 84247]

I actually have a different proof that only uses thought experiments. Imagine you have two novels on SH. Both have equal readers and views. They each have the same number of ratings.

Novel A is GL, and novel B is NTR. Which one do you think will have higher rating?

Now, to make the rating system useful again you have to say something like "apply a different rating scale to each genre respectively." So you get things like a maximum potential of one genre or theme getting like 4.7 stars while another one is capped at 4.2, given a high rate of views.

 

[Tyranomaster]

I actually have a different proof that only uses thought experiments. Imagine you have two novels on SH. Both have equal readers and views. They each have the same number of ratings.

Novel A is GL, and novel B is NTR. Which one do you think will have higher rating?

Now, to make the rating system useful again you have to say something like "apply a different rating scale to each genre respectively." So you get things like a maximum potential of one genre or theme getting like 4.7 stars while another one is capped at 4.2, given a high rate of views.

People inherently do that already though. Unless you're completely unfamiliar with a genre, a search in that genre will show average ratings. This applies in all aspects of life. The best WNBA basketball team, still loses to mediocre NBA steams. People have internal gauges for all sorts of value judgements. A 4.0 rated McDonalds is a good McDonalds. A 4.0 rated $$$ steakhouse is a bad steakhouse.

It's about your expectation going into anything that adjusts what the rating means. Sure, if you compare apples to oranges, you'll find more tangy oranges than apples. I don't think anyone honestly believes, however, that a high rating on google would hold more meaning to food connoisseurs than an endorsement by Chef Ramsay. Likewise, a highly rated story on Amazon wouldn't be more valuable to me as an author judging the quality of a story than an endorsement by Sanderson. It just is what it is. Ratings always are subjective, because the universe doesn't have a universal qualitative measure for "The goodness that you've aligned the scribbles on a page".

 

[Deleted member 84247]

People inherently do that already though. Unless you're completely unfamiliar with a genre, a search in that genre will show average ratings. This applies in all aspects of life. The best WNBA basketball team, still loses to mediocre NBA steams. People have internal gauges for all sorts of value judgements. A 4.0 rated McDonalds is a good McDonalds. A 4.0 rated $$$ steakhouse is a bad steakhouse.

It's about your expectation going into anything that adjusts what the rating means. Sure, if you compare apples to oranges, you'll find more tangy oranges than apples. I don't think anyone honestly believes, however, that a high rating on google would hold more meaning to food connoisseurs than an endorsement by Chef Ramsay. Likewise, a highly rated story on Amazon wouldn't be more valuable to me as an author judging the quality of a story than an endorsement by Sanderson. It just is what it is. Ratings always are subjective, because the universe doesn't have a universal qualitative measure for "The goodness that you've aligned the scribbles on a page".

You right. This is also why I don't like star systems, but I can't complain because I'm also retarded. I rate everything 5 stars or nada

You right. This is also why I don't like star systems, but I can't complain because I'm also retarded. I rate everything 5 stars or nada

Except for K5 she got a 1

 

[RedMuffin]

Novel A is GL, and novel B is NTR. Which one do you think will have higher rating?

Hah! That's obviously the NTR. Ask Anon, he can vouch for me.

 

[Tyranomaster]

You right. This is also why I don't like star systems, but I can't complain because I'm also retarded. I rate everything 5 stars or nada

That's fine to do. It's up to people interpreting a rating system to understand how other people use it, not up to the people using it. It's actually somewhat difficult to interpret the ratings usually, but it's possible to navigate if you're clever about it. Too many upstart authors freak out about ratings that have no interpretable meaning to them though.

I'll say it again. There is almost no distinguishable difference between a 4.2 and a 4.8 rating to the average reader. They'll act like there is, but there isn't.

 

[Deleted member 84247]

That's fine to do. It's up to people interpreting a rating system to understand how other people use it, not up to the people using it. It's actually somewhat difficult to interpret the ratings usually, but it's possible to navigate if you're clever about it. Too many upstart authors freak out about ratings that have no interpretable meaning to them though.

I'll say it again. There is almost no distinguishable difference between a 4.2 and a 4.8 rating to the average reader. They'll act like there is, but there isn't.

This is why I am not a statistician. Wait...Actually, a statistician would publish these numbers in the news.

 

[Tyranomaster]

This is why I am not a statistician. Wait...Actually, a statistician would publish these numbers in the news.

Yeah. The thing about star ratings is people assume it means that a higher rating means a better story. All it means is that the people who read and rate this story give it a high rating. Those two things may seem the same, but they are very different at the fringes.

This is a bit off topic, but actually, the royal road system does a slightly better job of getting "objective" ratings from people. Many people rail on how their story gets trolled into oblivion with low ratings there, but one thing they prompt people to do when they review is to give mini-ratings on things like grammar, characters, etc. That removes some of the subjective feeling from it, and puts the rater in a mindset to actually rate something. The reason stories get lower ratings is because people reviewing go from a praise mindset to a rating mindset, and minor flaws they wouldn't have thought of are suddenly remembered.

However! That also introduces its own bias towards the specific categories as a result. Perfect grammar, Style, Characters, and Story are equally rated, with enjoyment not being a category. All rating systems do is tell you information about the people who rate. You can try to read into what that means at a deeper level, but statistics always only show exactly what is being asked, everything else is interpretation.

 

[Deleted member 84247]

Yeah. The thing about star ratings is people assume it means that a higher rating means a better story. All it means is that the people who read and rate this story give it a high rating. Those two things may seem the same, but they are very different at the fringes.

This is a bit off topic, but actually, the royal road system does a slightly better job of getting "objective" ratings from people. Many people rail on how their story gets trolled into oblivion with low ratings there, but one thing they prompt people to do when they review is to give mini-ratings on things like grammar, characters, etc. That removes some of the subjective feeling from it, and puts the rater in a mindset to actually rate something. The reason stories get lower ratings is because people reviewing go from a praise mindset to a rating mindset, and minor flaws they wouldn't have thought of are suddenly remembered.

However! That also introduces its own bias towards the specific categories as a result. Perfect grammar, Style, Characters, and Story are equally rated, with enjoyment not being a category. All rating systems do is tell you information about the people who rate. You can try to read into what that means at a deeper level, but statistics always only show exactly what is being asked, everything else is interpretation.

True. People will always be biased, so the statistical analysis most often becomes a biased conclusion. You can tell me to rate things different based on different genres. To me it doesn't make sense, but many are perfectly okay with that. Or even rating things different on different sites. For example, a 4.0 here is probably worse than a 4.0 in RR if given the same sample size. And this is even a biased conclusion.

 

[laccoff_mawning]

I'm going to pick a few books for study.

First, the top of the trending list right now, "A strange new life.".
148 5-stars, 9 4-stars, 2 3-stars, 1 2-star, and 1 1-star.
Second, at last place on the tranding list right now, "Slime Girl."
146 5-stars, 38 4-stars, 18 3-stars, 8 2-stars, 10 1-stars.
Third, a random story picked from the find-random-story button: "The sidekick story - I reincarnated as...".
6 5-stars, 1 4-star, 1 3-star, 1 2-star, 1 1-star.
Fourth, the most recently updates: "The Last Rae of Hope".
1 5-star. No other ratings.

So yea. In all four cases, both the mode and median ratings are 5-stars.

 

[Deleted member 84247]

I'm going to pick a few books for study.

First, the top of the trending list right now, "A strange new life.".
148 5-stars, 9 4-stars, 2 3-stars, 1 2-star, and 1 1-star.
Second, at last place on the tranding list right now, "Slime Girl."
146 5-stars, 38 4-stars, 18 3-stars, 8 2-stars, 10 1-stars.
Third, a random story picked from the find-random-story button: "The sidekick story - I reincarnated as...".
6 5-stars, 1 4-star, 1 3-star, 1 2-star, 1 1-star.
Fourth, the most recently updates: "The Last Rae of Hope".
1 5-star. No other ratings.

So yea. In all four cases, both the mode and median ratings are 5-stars.

My question becomes. What can the reader learn from it when they see it?