How Crowners SHOULD Work

Figured I might as well bring this up here since this has become a point of contention recently, for reasons I'm about to explain.

About a month ago, there was a crowner in the TRS thread for Esoteric Happy Ending that determined what its new name would be. The two most popular options were Allegedly Happy Ending and Allegedly Optimistic Ending. Allegedly Optimistic Ending recieved over twice as many upvotes (28 compared to 13) but Allegedly Happy Ending won due to it having a higher upvote-to-downvote ratio (6:50 to 1 compared to 5:60 to 1). This is because, per How Crowners Work, the option with the higher ratio takes priority over the one with the most amount of upvotes. Up until now, this wasn't an issue, but now, there have been ppl who disagree with the crowner result and feel that Allegedly Optimistic Ending should have won due to it not only having more upvotes, but also for it being a more appropriate name, among other things not coming to my head at the moment.

So how do you guys think crowners SHOULD work? This thread is to discuss that, as well as any potential changes regarding crowner policy.

Edited by JHD0919 on Mar 29th 2024 at 7:23:44 AM

This is Idol Tap. (My Troper Wall)

For example

• Option A: 2 upvotes, 1 downvote (2:1)
• Option B: 20 upvotes, 11 downvotes (1.81:1)

By current logic, Option A wins.

I feel it'd make more sense that the option with more total votes is more valuable, Option A may as well have been added the last minute.

TroperWall / WikiMagic Cleanup

To be fair, a ten vote minimum is required. So option A would be invalid for falling under the minimum. That being said, I still absolutely agree, and the ten vote minimum is just sort of kicking the can down the road a bit.

In addition to the idea of votes counting more, there's also the option of requiring a run-off crowner when the positive choices are relatively close. Plus, I THINK it's currently policy that any new option added to a crowner is supposed to extent the minimum time by three days, to give people the chance to see it.

• A: 10 upvotes and 5 downvotes
• B: 100 upvotes and 51 downotes

good point that new option would extend the crowner, but the idea remains that older options have more discussion or people bothering to check in. Though maybe there's logic in reverse of this.

Edited by Amonimus on Mar 29th 2024 at 4:28:27 PM

TroperWall / WikiMagic Cleanup

This issue of adding new items last minute had actually caused a lot of confusion in TRS before the process became more standardized. We were never really sure how long to wait, or how to count the votes.

As for the ratio stuff, I agree it's rather confusing.

Currently Working On: Incorruptible Pure Pureness

It does seem weird that ratio comes into it first. A minimum ratio as a cutoff point? Sure. A ratio as a deciding factor in the event of equal positive votes? Sure. The ratio as the primary determination over the actual amount of votes? Probably not a good idea.

Avatar Source

The effect of the ratio is that the winner of the crowner is the least controversial option. "Allegedly optimistic" had more upvotes but also substantially more downvotes; the wiki consensus was fine with either one but more people had objected to it. If these people really don't like "allegedly happy" winning they could have downvoted it.

Now, that's not to say there aren't issues, and this ratio-first approach may be more fitting for technical and rules changes affecting the whole wiki than TRS or IP decisions, especially those where there are multiple options but a single winner. For one thing, I don't know how late in the process "allegedly happy" was proposed. For another, since both options had strongly positive ratios you could make the argument that the option with more support behind it should carry the day even if it does also have more downvotes, since a handful of downvotes can have a disproportionate effect on the ratio and not everyone may bother to downvote every rival to their preferred option - especially if they might be okay with one option but think another one is clearly better. The fact that crowner options are, by default, ordered by upvotes-minus-downvotes, which is also more prominently displayed than the ratio, may also mislead some people as to what's "in the lead" if they aren't familiar with how crowners work.

Using the ratio to select the least controversial seems like it should only be employed as a rule of thumb, not the primary determiner. Especially when we're talking about strongly positive ratios like this example. Here, we have a ratio of 28:5 and 13:2. So, 5.6:1 and 6.5:1. If there had been one more negative vote on the latter, then it would be 4.33:1 instead. One less vote on the former, and it would be 7:1. As you can see, in practice, a single negative vote when the denominator is very small can completely change the outcome. That's both unduly swingy and means that slight dislike is given a lot more weight.

Honestly, above a cut-off threshold, I don't think ratio should be applied at all except as a tiebreaker. Ratio subjects you to picking an option with an order's of magnitude difference; you could have an option with over 100 votes on some wiki project, and one with just over 10—but if the one that barely got attention has a slightly better ratio, it's somehow better? Even if more people voted against the other proposal than even acknowledged this one's existence—that's hardly the least controversial choice. It means you get situations like this, where both options are obviously non-controversial on their merits, because they're very positive, but the one that more people seem to like is discarded. And it's very swingy if the denominator is low, which is... unideal.

But if ratio-based determination was to be kept, I think you'd have to remove the option to add to crowners after they're open. It just adds too many factors to a value that's much more easily changed than raw totals.

Avatar Source

I think I'd echo most of what MorganWick said. If we're aiming for consensus, ratio is a measure that minimises controversy.

And I'd agree that when Crowner options are mutually exclusive, it's problematic to add new options once the vote is running.

Edited by Mrph1 on Mar 31st 2024 at 11:42:01 AM

In a practical sense, I don't think it minimises controversy when the amount of downvotes is given vastly more weight than anything else as soon as you've crossed the ten vote threshold.

If the ratio is above 2:1, it's probably not controversial. If it's above 3:1, it almost certainly isn't. If you're looking at ratios of 4:1 or higher, that's just telling you that there's so little objection that any additional one will have massive impacts on the exact number, and it's definitely time to ignore it for picking a winner.

Edited by RainehDaze on Mar 31st 2024 at 2:35:26 PM

Avatar Source

If folks really wanted Allegedly Optimistic over Allegedly Happy, wouldn't there have been more downvotes on Happy from the Optimistic upvoters? So the result ratios say while there were fewer people for Happy, there were also fewer people not for Happy.

Instead of changing the rules for all crowners, maybe allow a runoff crowner for the contested options of an individual crowner if there's enough of an outcry over the results

I think you're making the assumption there that people were placing votes remembering that crowners work on ratio (which, obviously, they didn't or we wouldn't have this conversation), and it's therefore more important to downvote anything you think is worse than to upvote things you consider acceptable. Which is a completely bonkers system where you think about it, and rewards spite more than anything.

Avatar Source

I mean, I think it's a fair point. If I'm neutral on something I don't bother to vote, but also in that case I have to accept that something I'm not as excited on might be the winner. If you absolutely don't want the other stuff to win though, downvoting is the only real option.

Currently Working On: Incorruptible Pure Pureness

Maybe, but where else have you ever seen votes weighted so that the more important thing isn't "what gets the most support?" But "what gets the least negativity?" It's a weird way to look at it—though I stand by the fact that unless you have a very large culture change so that people are mass-downvoting, it's absolutely not a reliable way of picking things. Both in the senses already brought up that a 10:2 option would get selected over a 100:21, outright dismissing the majority of votes; and that for small denominators a single negative vote either way makes a huge difference, which makes the entire result very swingy.

Avatar Source

I think the issue is that there's essentially 3 ways to vote on something: yes, no, and abstain. There are also two different ways that crowners work: whether the options are mutually exclusive or not.

The controversy seems to only be about when options are mutually exclusive, in which case it'd be a lot simpler if you couldn't add options without resetting the vote. Then change the voting to be a simple "select the options you like".

This would essentially remove the concept of downvoting for these types of crowners, making the number of votes an option has the same as its vote ratio. The 2:1 consensus rule would just mean a vote needs at least a 2/3 vote ratio to have consensus.

Crowners where multiple options can coexist would stay the same since no one has a problem with them.

Being flexible with which method is used when might help. Am cautious about run-off crowners since average user may not want to vote a second time or miss it.

Edited by Amonimus on Apr 1st 2024 at 5:24:37 PM

TroperWall / WikiMagic Cleanup

Worth noting that crowners originated as an attempt to select a single Crowning Moment Of Awesome, hence the name. In other words, they were conceived and built with the idea of selecting a single winner. The idea that this system should be completely replaced for single-winner decisions but left intact for multi-winner ones seems rather odd to me.

The key, to me, is that the most prominent number and the one results have always been sorted by is upvotes minus downvotes. That was the standard by which each "crowning moment" would be selected. I seem to recall that at one point, crowners for TRS and IP decisions only needed to meet a minimum threshold ratio and then the highest plus-minus score meeting the threshold won. I may be misremembering that, but maybe it would be a good idea to adopt or go back to that?

There are many voting methods:

• plurality (pick one)
• highest average
• total ratio
• ranked multi-choice
• borda count
• Condorcet table
• several dozen of others

Sometimes candidates win with less amount of total votes because they had "more votes by those who did not vote for other options".

In short, it's all subjective and if people vocally disagree between two top options it warrants an extension of a discussion. Generally the situations when people start complaining they prefer the second ratio option are rare, but worth asking why everyone have picked one and not the other.

Edited by Amonimus on Apr 2nd 2024 at 2:43:46 PM

TroperWall / WikiMagic Cleanup

Yeah, only having the ratio as a threshold seems more sensible. I guess it might have been changed, if it ever was, because otherwise newly added submissions have no chance. But allowing for things to add to the crowner later and specifically facilitating a way for those to win with less attention or scrutiny just feels backwards as a concept.

I think this is the only voting system I've ever seen where a single round of voting, with no elimination, could somehow select the candidate with the least votes. Fine for picking dinner in a friend group, not much else.

Edited by RainehDaze on Apr 2nd 2024 at 2:54:24 PM

Avatar Source

I had a crowner option I had previously downvoted on that previously had 9 upvotes and 3 downvotes, with a 3-to-1 ratio. I changed my vote to an upvote this morning, and now the option, with 10 upvotes and 2 downvotes, has a 5-to-1 ratio.

Raineh is right - it IS very swingy.

This is Idol Tap. (My Troper Wall)

Bumping so we can talk about this proposal.

This is Idol Tap. (My Troper Wall)

From my perspective, that's mostly just closing a loophole and bringing the IP moment crowners into line with the 2:1 rule used everywhere else.

Agreed.