I see my score is 4.18

[Dearelliot]

Can I see how the scoring was computed...Only a few comments

 

[StillStunned]

Well you see, when a mummy score and a daddy score love each other very much...

More seriously, no. You can get an inkling from looking at the number of votes and working out what combination of votes would be needed to get to your score. I think one or two people have developed formulas to help. Otherwise, unless you watched them accumulate in near real time, it's a mystery.

 

[PapaRomantic]

Total number of stars divided by total number of votes. If a story has 20 stars total and has 5 voters, the score average would be 4.00.

 

[Bamagan]

Can I see how the scoring was computed...Only a few comments

The site does not provide vote histograms or similar breakdowns. Depending on how much you like doing math and statistics, it's possible to extrapolate probable vote distributions, at least when the vote total is relatively low.

 

[SimonDoom]

Your score is the mean of all scores, 1 through 5, given to your story by the readers that voted on it. The Site does not disclose the vote breakdown.

 

[NotWise]

The general method of backing out individual scores is to multiply the score by the number of votes to get the total number of stars, and then try to construct a unique, or at least more likely, combination of that many votes that would give you that many stars.

It's hard to do.

 

[Tio_Narratore]

The vote distribution isn't that hard to track, but it is tedious. Check often, and each time it changes calculate the change in the average by the number of voters.

 

[Fatdog25]

This article explains one process for analyzing scores. It involves some math.

I'm not that mathematically inclined so I can't attest to its accuracy.

How to Analyze Your Scores

 

[Voboy]

It's a rabbit hole you can disappear down obsessively, writing code and running scripts to figure it out, or it's a reasonably confusing mass of numbers that quickly gets unwieldy.

Or? It's not something to worry about. Just take your score as it comes and treat it as one data point, not an end in itself.

All three options are valid.

 

[WCSGarland]

Yesterday one of my stories had 1.65 with 56 votes, today it has 1.65 with 55 votes. Once they have more than a handful of votes, I lose track.

 

[Writer61]

Can I see how the scoring was computed...Only a few comments

The short answer is "no", unfortunately.

When first published, if you check frequently (like hourly), you can get a reasonable idea of the distribution using a simple spreadsheet, but as soon as there are multiple votes whilst you are not monitoring it is practically impossible to know what happened.

Two three-star ratings have the same effect on your average as a one and a five, but you would probably view those scenarios differently. You can't know what happened, so there is no point worrying about it.

That said, I do keep track of the initial hours of my stories; I can't stop myself.

 

[lovecraft68]

Keep in mind that although people get obsessed with scoring over 4.5 for an H, anything over a four is good. The equivalent to 80%+ or a B on a report card.

Don't sweat it. Score hawking can get to be too big a distraction sometimes and affect how you write, because you're more worried about numbers over the story.

 

[StillStunned]

Keep in mind that although people get obsessed with scoring over 4.5 for an H, anything over a four is good. The equivalent to 80%+ or a B on a report card.

And even that's relative. I have a 2P POV story that's rated 3.98 3.97 today. I'll take that as a win.

 

[lovecraft68]

And even that's relative. I have a 2P POV story that's rated 3.98 3.97 today. I'll take that as a win.

Sure, technically anything over three means the majority liked it.

But for the most part someone sees a score under four and they label it as a 'bad' score.

I've said before, we get one bombers, we get people who leave crappy comments, but far and away the average vote here is 4+ and the average comment decent.

As a whole-minus the anomaly of LW-the site isn't what I'd call harsh, its actually rather soft in voting and comments on average.

 

[Writer61]

Sure, technically anything over three means the majority liked it.

Technically, it means the majority who could be bothered to vote liked it. This doesn't tell you much when the reads/votes ratio is low generally (it's around 1.25% for my stories),

 

[BobbyBrandt]

I'll use this as an example of the relative value of a story's rating:



The rating is good, but was the story successful with the readers? The votes per view ratio is off the charts, but the low number of views after almost two and a half years is the telling factor, for me at least. The feedback has been good, though very sparse.

If I judged the story strictly by the score, it would be a win. If I judged it on its market appeal, it would be a big loser.

 

[pink_silk_glove]

Keep in mind that although people get obsessed with scoring over 4.5 for an H, anything over a four is good. The equivalent to 80%+ or a B on a report card.

Not really. Someone did a score data crunch a few months back and the mean score (50th percentile) was I think 4.42 if I recall, and my personal avg score of 3.8ish ranked around the 12th percentile, so a 4 would be somewhere around a dismal 20th percentile. No school that I know hands out a B grade for a score like that.

LW aside, anything below roughly 4.2 is a pretty bad score.

 

[TheLobster]

Not really.

Yes, really. In any sane school grading is based on the score itself, not stack ranking versus other students.

The only error LC68 made was saying that 4.5 equals to an 80% score. It’s actually 87.5%, so quite close to an A.

 

[SimonDoom]

Yes, really. In any sane school grading is based on the score itself, not stack ranking versus other students.

The only error LC68 made was saying that 4.5 equals to an 80% score. It’s actually 87.5%, so quite close to an A.

That doesn't follow.

On an exam, a 4.5 out of 5 would typically mean someone got 90% of the questions right, or, in the case of an essay, performed at that equivalent level. Traditionally, that has correlated with an A. With modern grade inflation, I'm not sure one can say that anymore.

I'm not sure what you mean by "sane school." Many "sane" schools grade on a curve, at least in significant part. If a test is easy, then scoring a 90 out of a 100 as a raw score doesn't mean that much. If you have a pretty good idea of the quality of the overall pool from year to year, then grading on a curve will give you a pretty good and accurate assessment of the quality of performance of individual students within the overall pool. Raw score is only helpful if you are confident in the test construction and the way results are tabulated. I was a teacher for a little while, and I found this to be true when grading students. Some of the tests I created were easier than others, whether I wanted it that way or not, so raw score often was an unreliable guide. Percentile was much more likely to be a useful guide.

None of this has anything to do with a 4.5 at Literotica. We know from the data and from comments here that people who like a story are much more likely to vote than those who don't, so while a single 4 might mean "that voter liked that story," a mean score of 4.5 means no such thing, because it is not representative of all those who read the story. It is skewed high.

I think PSG is correct about this. Percentile analysis gives you a much better picture of a story's quality than does raw mean score, especially because so many non-quality elements skew the scores, such as the weirdness of Loving Wives and the way chapters are scored. If you see Chapter 38 of a story and it has a 4.83, all you know is that readers stuck it out through 38 chapters, and they're almost certain to be readers who like the story--otherwise, why would they still be reading it after that long?

You can't compare a 4.5 LW story with a 4.5 romance story. But you can make some comparison of both stories by looking at their percentile rank within their categories, making further adjustments by number of chapters and by story length (longer stories score higher, all things being equal--again, probably because of the attrition factor).

Judging from my own experience as a reader here for over 20 years, while I build in a very significant "error rate" in assessing scores, I generally do not regard a mean score of 4 as indicating the story is "good." I probably will not regard it as good, unless there's something about the story (it's very short, it's LW, it's in an unpopular category like Humor, it covers a weird subject, etc.) that gives me reason to believe the score will not represent its quality.

 

[designatedvictim]

It's a rabbit hole you can disappear down obsessively, writing code and running scripts to figure it out, or it's a reasonably confusing mass of numbers that quickly gets unwieldy.

Or? It's not something to worry about. Just take your score as it comes and treat it as one data point, not an end in itself.

All three options are valid.

Yeah.

Once you reach a certain number of votes, 150 or so, it can be problematic calculating what score you may be getting on any one vote.

My most-voted story now has 334 votes, averaging 10-15 reads/hour and maybe a vote a day for the last month, or so, since it fell off the New page.

Since it tallied vote #327, about four days ago, the site rating has been 4.66. No variance out to two decimal places over seven votes.

I seem to be be getting a mix of 5s and 4s, but which is which is currently impossible to tell.

Even how the site software rounds that value can impact what you see. Does it truncate past the second decimal place? Does it round? If it rounds, does it round up at .005 or round down? You can even pick the rounding rules, so if the previous digit is even, it rounds down, if it's odd, round up.

It's enough to give you headaches.

Still, I've been very fortunate in that all but one of my seven stories (or parts of stories) has rated over 4.5. Two don't have enough votes to rate an H, and one of those, at 4.00, has only four votes after five months (and lost a vote in a recent sweep), so I'm not holding my breath on that one.

I track this stuff - that's what scripts are for - but, as said, it can be such the rabbit hole.

Your best bet is to try to back away from it, however tempting it is to follow.

As LC68 said up-thread, over a three should be fine, but the perception for most people is that under a 4 is not well-liked.

 

[TheLobster]

Raw score is only helpful if you are confident in the test construction and the way results are tabulated.

Shouldn't you be? Sure, relative percentiles can give you evidence as to how well a particular test is calibrated, but using those percentiles to post-hoc adjust the results so that they match some imagined ideal of how many A's, B's, and F's there "should" be?

Yeah, that is so unfair that I might as well call it insane.

Percentile analysis gives you a much better picture of a story's quality than does raw mean score, especially because so many non-quality elements skew the scores, such as the weirdness of Loving Wives and the way chapters are scored.

If those non-quality elements skew individual scores, is it not the case that they skew the percentiles as well?

Just look at LW: it is widely acknowledged that there are at least two factions there, BTB and RAAC, and their influence is felt by anyone who ever dared to post there. How big are those factions? How much of a percentage of whole readership each one takes? We don't know that. And if one is significantly bigger, then a story that panders to that factions will be unduly propelled in terms of both raw score and stack ranking, whereas the one that offends the larger gang will tank by both metrics.

Other categories may not have this kind of faction warfare, but they still suffer from the systemic issue that uncategorized kinks aren't spread evenly among readerships in each category.

You can't compare a 4.5 LW story with a 4.5 romance story.

Indeed, you can't compare a 4.5 romance story with a different 4.5 romance story. But you don't need to. A score is what a score is, just like a test result is a test result is a grade.

 

[designatedvictim]

That doesn't follow.

Keep in mind the voting is from 1-to-5, not 0-to-5.

 

[pink_silk_glove]

I'm not sure what you mean by "sane school."

He means that I'm an idiot and he's putting me in the corner with a dunce hat.

 

[pink_silk_glove]

Yeah, that is so unfair that I might as well call it insane.

Leave the math to those who understand it.

 

[SimonDoom]

Keep in mind the voting is from 1-to-5, not 0-to-5.

True. But this only highlights the fact that a 1-5 vote system is not in any way commensurate with a 0-10 range for a test score. The system is completely different. What is being measured is different.

If those non-quality elements skew individual scores, is it not the case that they skew the percentiles as well?

It depends. No, if, for instance, you separate categories before calculating percentile. A 90% LW story may be roughly commensurate with a 90% Romance story, but a 4.5 LW story is not in any way commensurate with a 4.5 romance story.

Shouldn't you be? Sure, relative percentiles can give you evidence as to how well a particular test is calibrated, but using those percentiles to post-hoc adjust the results so that they match some imagined ideal of how many A's, B's, and F's there "should" be?

Yeah, that is so unfair that I might as well call it insane.

It's unfair, I agree, if it's used to match an "imagined ideal," but that in fact does not describe the reality. Experienced teachers have a pretty good what the range of skill is likely to be in a class, and a percentile ranking usually does a pretty good job of giving you an idea of what ultimate grade they deserve.

In the case of Literotica, there is no ultimate grade, except in the case of the completely random and meaningless "red H," which is so ridiculous that I don't see how anyone can justify it from the standpoint of conveying useful reader information OR fundamental fairness. The only two things it has going for it are a) tradition, which counts for a lot, and b) author expectations, because many authors would be upset if their red Hs were taken away. But a category-adjusted percentile-based red H would be far fairer and convey more useful information to readers. If a reader knows that an LW story with a red H is in the top 80% of all LW stories, the reader can decide what that means for himself.

The problem with the raw score system is that readers, and even authors, clearly do not understand what a 4.5 means, as evidenced by comments in this forum. For some categories, it's below the median.