Didn't realize misresolutions could be corrected. I bet yes on the off-chance the market was misresolved. Cest la vie.

Also my daily profit + profit graph are still wrong.

@TesseractCat It's a brand new capability, but they've been intending to do it for a while (see https://manifold.markets/jack/will-it-be-possible-to-undochange-t)

@TesseractCat The daily profit and graph are only calculated every 30 mins, I think yours has returned back to the way it was

@ian

@TesseractCat Oh yeah that’s a bit slower even iirc

This is literally a game changer

I corrected the resolution as requested by the author. The leaderbord displayed is incorrect, however. Will figure out how to fix that as well.

@ian actually it looks like the leaderboard is correct

@ian Awesome! Really glad to see this functionality has been implemented. Any news on using it to correct all the other large misresolutions, and/or giving the functionality to market creators directly?

@IsaacKing yeah, would this ultimately lead to us re-resolving some older markets to their correct form? Though I think that might lead to more troubles than anything

@ian I only see people that made money by selling before the market closed. But that Ham guy he made money by buying yes right before it was resolved

@IsaacKing The earliest misresolved market can be corrected with our current tech if it was misresolved on or after Mon Feb 06 2023 14:33:46 UTC time (1675694026838 ms after the epoch) - https://manifold.markets/jack/will-it-be-possible-to-undochange-t#jhockif5ASLC3elhxBST

@Simon1551 Our current tech only undoes the final payouts and nothing else.

I only see people that made money by selling before the market closed. But that Ham guy he made money by buying yes right before it was resolved

They bought NO first and then bought YES, so it was the same as selling the shares for a profit. The payouts for that look correct.

@Simon1551 oh actually he earned that profit by selling his no shares early and profitably, so there isn't a problem there

@jack @ian It's because he also was in the leaderboard when it was resolved incorectly I thought it was still the same profit

@ian Wait, so the market still displays everything else incorrectly, and it just changes who gets how much mana? I think that's likely to be very confusing if people will still see the displayed market resolution as wrong.

@IsaacKing The market doesn't display anything incorrectly afaict, what do you mean?

Everything looks correct. People were just confused because they saw some traders, for example AlexLiesman (the top trader on this market), had a big profit before it was fixed, and the same profit after. This is totally correct because AlexLiesman sold all their shares before the resolution, so changing the resolution didn't change their profits at all. Same thing with several other people

@IsaacKing Market on whether we'll get the ability to do this easily without admin assistance:

@ian You said "Our current tech only undoes the final payouts and nothing else."

@IsaacKing I thought you'd seen that he resolution had been undone - as you can see we can unresolve the market as well.

@ian Really cool to see it happen, I can’t imagine anyone will want to buy mana if they are having it taken by freaks intentionally mis resolving things (not what happened here, but has been happening a lot lately - even to me)

@Gen I'm not sure this fixes the intentional misresolutions, which to my understanding Manifold still permits (and, I think, should continue to permit)

@CarsonGale I believe their line is to ban fraudulent misresolutions done for profit, but to allow resolutions that stem from legitimately different beliefs.

@CarsonGale Why would you permit intentional mis-resolutions? If a market maker decides to mis resolve a big event like this, or a presidential election, and screw 1000 users out of significant bets, the site will lose users.

Yesterday a user misresolved a market that had ~5 participants to lose me 800 mana, because I disagreed with him in a different thread where I defended a market-maker making a highly contentious subjective resolution. I have >20k balance, so it's not going to send me away, but if I lost >80% of my balance because that clown wanted to spite me, I'd have certainly quit the site. If you pay $10 for 1,000M, having someone choose to steal 800 is like having $8 taken from you.

I love the site, and I like betting on small markets from new creators to encourage more people to participate, but I'm far more hesitant to do it now. I agree that market makers should have discretion in resolutions, but Manifold updated their own terms on 14 December 2022 to include protections from market makers deceiving users to enrich themselves, and I hope Manifold enforce that condition.

I have no issue with subjective resolutions, but if someone is just mad, misleading, or trying to spite someone, it completely destroys the fun of the site. Market participants can & should lie and cheat to get the resolution they want, but the market makers should not be resolving incorrectly to benefit themselves or punish people they don't like.

@IsaacKing "I believe their line is to ban fraudulent misresolutions done for profit, but to allow resolutions that stem from legitimately different beliefs."

Which is exactly as it should be 👍

@Gen "it's not a lie if you believe it"

@StevenK If there is any room for dispute, I support market makers having absolute discretion when settling on a resolution

@Gen But I also think those makers should withdraw conflict of interest in those circumstances (closing bets, donating profits, etc)

The main thing I (and I think Manifold) want to avoid is Manifold becoming an arbiter of truth. Let users be the judge rather than Manifold. Sometimes that means you can't trust unknown users to resolve correctly, which sucks, but you retain the discretion of market creators and it becomes increasingly important as the user base expands to develop a track record of faithful resolutions as a market creator. I think the "Trustworthy.ish" designation will eventually need to expand to something more developed to show a spectrum of trustworthiness vs a binary.

I'm fine with banning explicit misresolution for profit. But if some market creator has a market like this, misresolves, and then explains afterwards that they think the Eagles actually won in spirit and that's what they wanted to measure, then I would be contra forcing that resolution change without the consent of the market creator. Maybe mark as an untrustworthy creator or something and stop boosting their markets, but at the end of the day leave it to users.

@CarsonGale So you’re saying if the maker here said that the eagles won, it shouldn’t have been re-resolved? This site would lose all users if fraud was supported like that (it explicitly isnt, in the terms as I showed). It’s so easy for bad actors to move their mana to other accounts that I don’t even care about bans, but markets which are inarguably fraudulently resolved for a makers benefit should be reverted every time. The only way to prevent it happening is if they know it’s pointless. Otherwise, one guy burns 10 people, and 2 of those burn 10 more each to recover their losses. I’ve seen that exact story play out a few times already.

Manifold deciding to N/A a market because of user reports IS letting “users be the judge”

@CarsonGale I would very strongly prefer Manifold becoming an arbiter of truth, or consulting an external arbiter of truth, to the status quo where the arbiter of truth is unreliable random users. Ideally, an arbiter of truth would consist of multiple people, each carefully vetted for reasonableness and reliability in following resolution instructions, with no knowledge of or interaction with bettors and maybe each other. Then there'd be an interface of "you send the arbiters your public resolution instructions and any other information that the question description says you'll send them, and they return a YES or NO, maybe with a short explanation, probably after a majority vote". Subjective resolution by the question author can be a special case that has to be explicitly announced up front. This may sound like a lot of effort, but I think there would be huge efficiency gains from specialization, maybe the full process would only have to be used a fraction of the time, and so on.

It's a slippery slope to begin manually adjusting users' markets. There is a spectrum from "flagrant and obvious fraud" to "honest disagreement" and I'd rather Manifold stay further away from toeing the line vs closer re: manual resolve changes.

I think the eventual path forward regardless is to develop a reputation system that rewards trustworthy market creators. Such a system would have natural resistance to intentionally misresolved markets (given reputation costs). And if Manifold establishes an early policy against interference except in the most egregious of cases, it enables them to avoid the inevitable culture war debates that will accompany hotly-contested markets where one side thinks their preferred resolution is "fact" and the other "fiction".

@CarsonGale Of course if there is any room for disagreement, it shouldn't be manually overturned.

If the question is about the winner of the superbowl, and has clear resolution criteria, as is in the description here, then this market should have been overturned regardless of the maker requesting it. Otherwise, you're going to lose new users and encourage bad behaviour from users you don't want.

Market participants should not be expected to be honest or reliable, they should campaign for mana and nothing else, but the makers must be honest or the system doesn't work.

@Gen I agree that this specific case seems clear and decisive. But a precedent of giving centralized power on cases like these may lead to non-consensual resolution changes in future markets that are less certain. I think I just have less trust in institutions re: power creep than you do.

I'm comfortable that this approach might hinder growth and may lose users that lose money on misresolved markets. It's relatively easy to gain a reputation for resolving markets accurately - do this consistently and you will establish a track record so that traders know to trust you.

The system as I envision it would work just fine - but earned trust between creators & traders would be paramount.

@CarsonGale Yeah, I'm mostly just assmad that some guy stole 800M from me and I want it undone. If it's discouraging for me, I can only imagine how it would be for someone with significantly less mana to fall back on. Several times makers have stolen and then just laughed about it afterwards. It really spits in the face of the idea of buying mana when someone can openly fraudulently resolve their market to take it from you.

I do agree generally with the idea of trusting the community, and I still plan to continue making large bets on noob makers to encourage new market makers. I'd be much happier though if we at least got back the scammer badges, otherwise you just feel helpless knowing reports have almost no real function.

@CarsonGale Reputation systems won't work because most creators don't have a large enough sample of resolved markets. In the best case someone will have resolved maybe 100 markets correctly, so there's no strong evidence against a 1% rate of them misresolving markets, and then some particular markets will be 10x as vulnerable to misresolution. Markets with controversial resolutions, where different traders have different expectations about how they'll resolve, are disproportionately the ones that attract a lot of betting and attention. Other creators will have a shorter track record, so paying too much attention to creator track records requires traders to avoid betting on a large fraction of topics, unless there are several markets per topic, which has its own problems. Reputation systems also require trusting Manifold to track reputation, which seems to create all the same issues as trusting Manifold to resolve markets.

Yeah that really sucks @Gen, sorry to hear that happened. I haven't personally had that happen to me in any meaningful quantities but maybe if I did I would have a different sentiment.

And did they get rid of the scammer badge? I quite liked those - thought it was helpful in identifying creators with a history of perceived misresolution without actually blocking them from the site / forcing a resolution change.

@StevenK that surprises me that you don't think 100 accurately resolved markets is a good indication of reliability. It takes a good amount of effort to make 100 markets, and you would lose that goodwill with just one sustained misresolution.

I think on many topics - particularly those with debatable resolution criteria - you do actually want multiple markets with different creators, since they provide a different perspective to the question resolution.

Seems like there are relatively-objective metrics that Manifold could use to identify high-trustworthiness creators.