Chapters....the unfair advantage

[TheNewcomer]

Top 250 with a drop down list where you can select each year?

I’d use them if I could do that at least.

 

[RejectReality]

If they're complete stories ( ala my Magic of the Wood series ) then they're complete stories. Doesn't matter whether they're connected, share characters, etc.

Each submission has a beginning, middle, and end. A chapter requires reading what came before and what comes after to make any real sense of it. Different animals.

All right, that makes sense...but what about stories that are a series but not necessarily chapters? There are a few of them that are broken down into parts with chapters as well, or some that have different titles yet are meant to be read in order as a series of events.

Maybe top authors in each category?

 

[AfterDusk]

If they're complete stories ( ala my Magic of the Wood series ) then they're complete stories. Doesn't matter whether they're connected, share characters, etc.

Each submission has a beginning, middle, and end. A chapter requires reading what came before and what comes after to make any real sense of it. Different animals.

Follow up question: what would be used to determine the overall average of the chartered story? Granted math is not my highest subject, but it seems there could be a huge discrepancy between the following calculations:

Averaging a story/series based off of the averaged score for each chapter

Averaging a story/series based off of a total number of votes/scores from each chapter

 

[NotWise]

Averaging a story/series based off of the averaged score for each chapter

Averaging a story/series based off of a total number of votes/scores from each chapter

The site probably has the information needed to use the second method, which I think is the better one. I don't know if the information is organized so it can be easily done, but that's a different problem.

Calculating a lumped score for a story series might discourage writers from tying stand-alone stories together as chapters.

 

[SimonDoom]

Follow up question: what would be used to determine the overall average of the chartered story? Granted math is not my highest subject, but it seems there could be a huge discrepancy between the following calculations:

Averaging a story/series based off of the averaged score for each chapter

Averaging a story/series based off of a total number of votes/scores from each chapter

The former method would be better than the latter. Assuming a sufficient number of votes, there'd be no reason to give greater weight to the chapter that got more votes, because after 100 votes an extra 50 votes would be unlikely to change the score that much. The logical thing to do would be to calculate the mean score for all of the chapters, without regard to vote numbers.

 

[TxRad]

The former method would be better than the latter. Assuming a sufficient number of votes, there'd be no reason to give greater weight to the chapter that got more votes, because after 100 votes an extra 50 votes would be unlikely to change the score that much. The logical thing to do would be to calculate the mean score for all of the chapters, without regard to vote numbers.

Since the story needs at least a certain number of votes to get on a top list then the number of votes does play a part but not with the score.

 

[SimonDoom]

Since the story needs at least a certain number of votes to get on a top list then the number of votes does play a part but not with the score.

The way to deal with that would be to use only those chapters that reach a certain vote threshold (which I don't think is very high) in the calculation of the mean.

 

[AfterDusk]

The way to deal with that would be to use only those chapters that reach a certain vote threshold (which I don't think is very high) in the calculation of the mean.

Then it seems like, by omitting certain chapters, the average would be skewed and not a fair representation of the entire body of work (because it doesn't include the entire body of work).

In computing the average, should more weight be given to chapters that had more votes? As mentioned before, later chapters tend to have higher scores/fewer votes so it would seem like that could give the overall story a higher rating than it deserved.

(I'm pretty much throwing questions out there to be a snark, not because I actually care one way or the other)

 

[SimonDoom]

Then it seems like, by omitting certain chapters, the average would be skewed and not a fair representation of the entire body of work (because it doesn't include the entire body of work).

In computing the average, should more weight be given to chapters that had more votes? As mentioned before, later chapters tend to have higher scores/fewer votes so it would seem like that could give the overall story a higher rating than it deserved.

(I'm pretty much throwing questions out there to be a snark, not because I actually care one way or the other)

Honest snark is fine. It's sneaky snark I can't stand. And to be frank you don't sound very snarky.

Thoughts:

1. It's unlikely in a multi chapter series that some chapters will reach the threshold and others won't, because the threshold isn't very high and it's unlikely the chapters will vary that much.

2. As a matter of probability, after the first 100 votes, it's unlikely that the next 50 will change the score much. So there's no need to weight stories differently based on how many votes they got. It makes no difference. It's also more complicated to weight the stories and calculate the mean that way.

 

[NotWise]

1. It's unlikely in a multi chapter series that some chapters will reach the threshold and others won't, because the threshold isn't very high and it's unlikely the chapters will vary that much.

The threshold for the all-time top list appears to be 100 votes. The common pattern is for early chapters in a series to get a lot of views and votes, and for the count to drop for new chapters. In my Unlikely Angels series (since removed) Chapter 21 had less than 1/10th the votes of Chapter 1.

For smaller categories, it's easy to imagine that only the first chapter would ever have more than 100 votes.

2. As a matter of probability, after the first 100 votes, it's unlikely that the next 50 will change the score much. So there's no need to weight stories differently based on how many votes they got. It makes no difference. It's also more complicated to weight the stories and calculate the mean that way.

I don't know that it's so complicated to treat all the chapters as if they were one story. The site has the data, and the script would only have to be written once. If you don't, then the early chapters where more people voted -- and the score is usually lower -- is underweighted compared to later chapters that typically have higher scores that represent the opinions of a rather few people.

 

[RejectReality]

All the data is available to Lit for every vote cast, so doing it as a whole shouldn't be an issue from a programming standpoint.

Whether that's a problem with processing power... Think about some of the incest serials... Several hundred scores per chapter x however many chapters x however many make the toplist, on top of already querying 500 stories per category.

If doing the full math drags things down, then an average of averages for both score and vote totals isn't going to be all that far off, really. It's not going to change much during a 24 hour period except for extremely low vote stories ( which won't make most toplists ) or serials with new entries out that day. That's why I suggested calculating the series average once per day in a timed task before the toplists are populated for the day.

Really, the volatility of day-1 voting could be detrimental. Better that you have the previous average up there the day the new chapter comes out, followed the next day with the new average, after things have smoothed out on the new chapter through a full day.

We don't know if Lit is storing an average score for stories, which can then be pulled by other queries without doing the full calculation every single time an average is needed. It does make sense, though. That average only needs to be changed when a vote is cast or swept, so the rest of the time, you could just yoink a stored value for every other query that is simply for display.

If I didn't mention it earlier, having a tab for one-shot and a tab for serials on each toplist would the the optimal way to solve the problem described in the OP. I still think each serial should be condensed to a single entry with an average in such a case. That would allow one-shots to shine in their own tab of the toplist, and perhaps expose some lesser-known serials to readers on that tab, rather than having 100 chapters of a single story taking up 1/5 of the serial tab toplist.

It exposes more authors to new readers, and gives readers another tool to find stories more tailored to their story length taste.