Mark Ingraham and his alts are only one user for purposes of this market, and alts of trustworthy ish users do not count towards the "average user" for obvious reasons.
I think the answer is clear, as should be my purpose in making this market.
I will resolve YES if manifold confirm, NO if manifold directly deny, and N/A if they never comment and I get bored of this market.
I hope if manifold have to think about this for too long, they will consider reframing the badge as "valued contributor" or "power user" rather than "trustworthy.ish". It sends the wrong message to new users and gives these users far too much sway over markets.
There are certainly some users worthy of the trustworthy badge, (Jack, Conflux, etc.) but overall I see it abused or invoked for selfish reasons more often than used correctly. Of course this is biased because people willing to abuse/invoke the badge will do so more often than the actual trustworthy people who almost act as though it's not there at all.
1,000People are also trading
@Conflux Do you mean more instances of rule breaking or more time to break rules once?
My expectation is that most casual users don’t engage seriously enough to do anything intentionally nefarious and don’t “accidentally” break rules with enough frequency or intensity get caught. Probably even more so in the past when rules were rarely enforced or well defined
I am really interested in the diversity of opinion here tho! Thanks for your input 😄
@Gen Wait, are you talking about more rules broken overall? I was talking about more rules broken per hour/site action, or something
@Conflux Sorry for the slow response - I was thinking more in terms of if someone breaks a rule or multiple rules to the point where they are being recorded/monitored/banned, that person is a "known rulebreaker" on a list somewhere. Or even just people who had received warnings for breaking rules.
My expectation is that as a percentage, 10% of normal (real, active) users might be in this group whereas 20% of trustworthy.ish users might be.
I don't really have any good criteria or way to resolve though, so I will allow any manifold admin who wants to give their take at any point in the future full power of interpretation and resolution outcome. I'd prefer it if they could give their criteria before they resolve so the market can adjust though. I'm hoping they have at least some stats that they can use to put together an estimate, even if it's as simple as: 11% of normal users are in X rulebreaker category, I think more than 11% of trustworthyish are in that category, resolves YES. Doesn't have to be super precise if they can't do it
@Conflux it wasn't really what I had in mind, but if a manifold staff saw and interpreted it that way, they can resolve it based on that idea (if that makes sense).
@Conflux Yeah it really isn't. If Manifold said they had specific data to base it off, I would happily reformat it around the available data. I'm pretty sure if I make it ultra-specific it will be impossible to get a response and the market will just end up getting resolved N/A, so for now it's basically up to staff if they ever wish to comment to run/resolve the market however they want (including N/A or PROB if they don't want to make a yes/no call)
Alternatively, if you think there is a more measurable way to run it while still ensuring high likelihood of a proper resolution, I'd be happy for you (or someone else) to re-run it more appropriately.
@Gen I just don't know what question you're asking. Do trustworthy-ish users break more rules overall? Maybe, because they spend more time on the site. Do trustworthy-ish users break more rules when controlling for that? I'd be pretty surprised.
@Conflux Oh sorry, my ideal for this market would be less about instances of rule breaking (e.g. Isaacs farm wouldn't be 10,000 rules broken, just 1) and more about the "rules broken per user".
I would expect that the "rules broken per user" would be higher on average for a trustworthy.ish user than a regular user.
As an alternative, and something I think manifold may actually be able to produce an answer for, would be users "actioned" for rule breaking. For example, if there are 2000 real or at one time active users, 125 actions (6.25%) then that could be compared to the rate at which lets say 50 trustworthy users with 10 actions (20%).
I'm guessing on the numbers to flesh out the example, but I imagine it would look something like that.
@Gen I just feel like you're trying to use "rules broken per user" as a proxy for "likelihood of breaking rules", which seems flawed to me when trustworthy-ish users probably spend much more time on the site than the average user?
Sorry if I'm coming off as annoying here, I just feel like that's a significant confounding variable which you're glossing over.
@Conflux I can see how it would appear flawed, but the badge doesn’t say “user who uses the site a lot” it says “trustworthy.ish”. I think it signals to new users that these badges will not do anything wrong or break any rules, when that isn’t the case. There’s like 5 trustworthy users I actually trust more than a random user for market resolutions or reliably sourced comments.
I see trustworthy users making ridiculous claims and baiting new users into throwing mana away, so I was hoping this market would prove useful to show “btw, you shouldn’t automatically trust this guy despite the badge”
What you’re saying makes sense, they’re more likely to break rules because of total volume of interactions. I’d think that despite the volume of interactions, those which break rules shouldn’t end up with the badge (or more likely, rename the badge …).
Do your concerns fade if the question were specifically about intentional rule breaking? I think it would have a similar outcome
It’s really late so I might not be making much sense, but I really appreciate the engagement (I want this market to make sense too) and I definitely don’t find it annoying 😌
@Gen Ok, I trust trustworthy-ish users substantially more than the average user. But that’s more because I trust users who I recognize and have a record of good market resolutions in the past, which is correlated with having the badge.
I feel like this is more a question of either whether the badge is appropriately named, or if the Trustworthy.ish badge gives power such that it corrupts some recipients.
I made a market that can be tested with respect to the first question. I have no idea how you might realistically test the other question.
I would love to see a badge system that allows users to anonymously buy and give "tokens" to other users for a few categories and the result of that (if overwhelmingly negative or positive) to be publicly displayed. e.g. someone is in the top 5% it could display a badge for either "trustworthy", "untrustworthy", "controversial" (mix of positive and negative tokens), "fun"/"unfun", "popular" (lots of spread-out tokens). Most people should still probably not have public badges unless tokens are both high in number and in specific bounds of the ratio.
It's obviously not good if you can gang up and nuke people (or give enough fraud tokens to make someone trustworthy), but it should balance out if each person can only give another user one token at a time and giving them a new token "burns" your previous vote. If users buy tokens, it burns mana while also encouraging votes on people's actions.
To avoid abuse/alts, make a rule that anyone caught abusing the system in anyway will be given X negative tokens (say, 100 untrustworthy) and it should really discourage it. It also adds a new metric for punishment overall by manifold - give them negative tokens!
Thoughts? I'm sure there are 100 better ways to do it.
The exact math you expect to happen will sway things wildly here. Most non-trustworthyish users that break rules will do so once and then leave; Isaac stayed. This also assumes that every single rule break is recorded, and manifold is willing to go to the effort to answer this question.
Overall very ill-specified :(
@Shelvacu Yeah, I know :( . I don't really want it to be, but I won't set unrealistic expectations for manifold data collection. If manifold staff have a specific record keeping process (e.g. they have estimate figures on unchecked rule-breaking, warnings, or bans (including how many are just mark Ingraham) I'm happy to use any of those stats relative to the total population for an easy determination.
I don't know how many people are trustworthy.ish but if I estimate 40 people with 5 known rule-breakers (there are probably more that manifold could speak about) then that is an extremely high bar (12.5%) and I don't see there possibly being 125 known rulebreakers if Mark Ingrahams accounts/infractions are all aggregated.
This market outcome will be totally decided by manifold themselves if they ever want to do so (or I will resolve it N/A after a long time). I haven't talked to any staff about it so I have no info more than anyone else.
@Mira Probably not if they immediately quit, but if they play for like a week or two and then quit (e.g. 10 logins), I think they should count. Honestly don't know how I would measure/consider, so I will just say that manifold staff (if they ever comment) can interpret it as they wish.
I don't know how many active users are rulebreakers, but I know of at least 5 blatant/caught trustworthy.ish users who were rule breaking. I assume there are at least 20x valid users as there are trustworthyi.sh users so I would expect there to need to be 100 known rulebreakers minimum outside of the trustworthy.ish group for the question to not be true. A lot of guesswork and assumptions involved though, I really don't know.