Re: Voting feedback

#3
It currently works by the stories that get the most sit higher on the list than those that don't. The idea was to have another list of member-validated stories for people to find stories they might like to read. You can find more details about it here: https://www.royalroad.com/blog/39/experimental-new-voting-system


Zonzog Wrote: It is redundant, unless first page is omitted, since it does not help you find something good you would otherwise miss.

Interestingly, I made a very similar suggestion to the admins to do that. But just focused on the shortlist feature on the homepage so it gives additional exposure to good stories that are not already on there. You can support it by following this link and maybe push it up on the wish list so that the admins can seriously consider this as an option:


https://feedback.royalroad.com/i653-hot-fictions-featured-on-home-page-should-just-show-stories-that-are-not-already-on-the-home-page

Alternatively, if you disagree with it or think this idea can be improved on, you can also mention that and suggest alternatives in the comments section of that suggestion.

Re: Voting feedback

#4

Zonzog Wrote: It is redundant, unless first page is omitted

How exactly will omitting the first page help?


If the first page is similar, I would presume the second page is similar as well. Or am I completely wrong there? Moreover, if people were to strive for the second page, we'd get all kinds of shenanigans. I can already see the "please stop voting" messages above and below chapters.

If the second page is similar and there is some need for more/different novels on the front page, we could also simply expand the existing system and add more novels, no? Or maybe add in a new category that mixes different stats. Popular + trending for example. And I don't doubt there are many more possibilities like that with the many different types of stats this site collects. I just don't see how a voting system would add much to a front page over using statistics.

A voting system should be useful regardless of the front page, not because of it. 

Re: Voting feedback

#5

Zonzog Wrote: After a few days it was implemented we can clearly see that it is filled mostly with the same novels that are on the first "Top rated" page. It is redundant, unless first page is omitted, since it does not help you find something good you would otherwise miss.

Then, the top stories authors could say "Please don't vote for my fiction!", and artificially lower the amount of votes to land on the second page, and still be promoted. It's a flawed idea 

Re: Voting feedback

#6
I've suspected for a while that after a few rounds of voting, with the associated reset (as I understand it), the top stories' readers will get tired of mashing the vote button, at which point lower level authors can run campaigns for votes and maybe make it higher on the list. Mind you, by lower level I really mean mid-high tier, since you'd still need the high raw numbers of people reading your stuff in order to get enough votes.

Do we want mid tier level authors to run campaigns for votes? That's the real question to consider. If no, then voting only encourages bad behavior. Imo, anyways. 

Re: Voting feedback

#7
I don't think it's redundant. 

I mean look, the truth is, there are a number of stories that absolutely dominate the site in terms of readership. This means that the first page of nearly ANY metric you come up with is likely to be highly correlated. But these different metrics do serve a purpose. 

"Weekly popular", for example, is likely to favor stories that update more frequently, with shorter chapters. "Best rated" will skew towards stories that have more engaged readers, who are more likely to rate/review. This voting system, I believe, will be similar to "Weekly popular", except that it will give more of a chance to stories that don't update as frequently, since the votes expire every two weeks. 

The problem is that the readership distribution on this site (and likely every similar site) is exponential: that is, a (relatively) few stories absolutely dominate all others in terms of readers. This essentially brute forces every metric. Does "Weekly popular" favor stories that update more frequently? Well, it doesn't matter how much it favors those stories, if an infrequently updated story has 100x the readerbase of a frequently dominated one. Same goes for any metric. So the first few pages of any system tend to be dominated by the same stories. Where you see a lot more variance between the lists is, I suspect, further down, perhaps past page 5 or so. 

I think it would be interesting to see the difference between the metrics, excluding, say, the top 100 on each list. And that's not meaningles, either; my story gets a non-trivial number of hits from being on page 13-20 of some of these lists. 

Re: Voting feedback

#9

CloverCloverClover Wrote: I don't think it's redundant. 

I mean look, the truth is, there are a number of stories that absolutely dominate the site in terms of readership. This means that the first page of nearly ANY metric you come up with is likely to be highly correlated. But these different metrics do serve a purpose. 

"Weekly popular", for example, is likely to favor stories that update more frequently, with shorter chapters. "Best rated" will skew towards stories that have more engaged readers, who are more likely to rate/review. This voting system, I believe, will be similar to "Weekly popular", except that it will give more of a chance to stories that don't update as frequently, since the votes expire every two weeks. 

The problem is that the readership distribution on this site (and likely every similar site) is exponential: that is, a (relatively) few stories absolutely dominate all others in terms of readers. This essentially brute forces every metric. Does "Weekly popular" favor stories that update more frequently? Well, it doesn't matter how much it favors those stories, if an infrequently updated story has 100x the readerbase of a frequently dominated one. Same goes for any metric. So the first few pages of any system tend to be dominated by the same stories. Where you see a lot more variance between the lists is, I suspect, further down, perhaps past page 5 or so. 

I think it would be interesting to see the difference between the metrics, excluding, say, the top 100 on each list. And that's not meaningles, either; my story gets a non-trivial number of hits from being on page 13-20 of some of these lists.

I agree with most of the issue you point out, I however fail to see how the current voting system solves any of those issues.


Moreover, there are plenty of ways you could create an alternate front page list without a voting system. Think of things like:
weighted ratings, a ratings threshold, a list of random novels from the top 100/500/1000, you can highlight specific genres (for example the top 10 rated horror novels this week, romance next week, ...) and you can combine different statistics and weight them in any way and shape you want.

I also think some of the issues regarding the exponentially large readership of larger novels, can partly be mitigated by a rating threshold, so only votes of subscribers are counted -- or maybe even readers with X reputation or more. Those people are often more involved and are much more likely to use the search function and read more niche novels. I'd guess that a list based on those statistics would at least take away some of the issue. Following that you can obviously still add more rules and qualifiers, until you get a list that seems reasonable. 

In my opinion a voting system as currently implemented, simply shouldn't have front page allocation as its main goal. Anything on the front page can be managed just as well or better with existing data, plus maybe some simply programming and additional rules. A voting system should add something relevant besides front page allocation -- e.g. some additional feedback for authors or additional utility for readers. 
If however the voting data becomes polluted due to competition for the front page, what use does the data still have? And if begging for votes becomes annoying, how much would a system like that be worth for readers to begin with? 

Skipping the top X would make things even worse. Than people will want more votes one week and fewer votes next week. At that point front page allocation will be the only goal of the voting system, because all other data will be useless. Meanwhile you can achieve nearly the same front page result with existing statistics (views, ratings, growth , ...). Why wouldn't you be able to skip the top 100 in for example a ratings list? Or maybe combine a ratings + activity list and then skip the top 100, or [insert a 1000 other possibilities]. 
Moreover, those existing statistics and data are much harder to manipulate and thus won't lead to negative behavior (e.g. begging for votes). 

The only voting system that would arguably have some merit, is some sort of communities or editors choice. Obviously a list like that would need a threshold (reputation for example), be only for writers(writers choice), or have some other threshold or qualification. Hell, that would even incentivize community interaction.
Though even there I'd doubt it would work well in the long run, unless it was maintained and regulated pretty strictly. Because a system like that is bound to turn into a popularity contest.

If I look at other comparable sites, I mostly see no effect or negative effects from aditional voting systems. So overall I highly doubt this will add anything of value.