Follow-up market for 2023:

Another benefit - being able to change resolutions would also make our lives better in many markets where the market is almost certain but not completely certain to resolve a certain way. For example:

• Markets on who will win an election are typically resolved when a decision desk calls it, but these calls are typically incorrect about 0.1% of the time - which means statistically it's fairly likely for at least one election market to end up resolved wrong.

• Markets like Will Liz Truss still be Prime Minister of the UK on Christmas Day 2022? technically have to remain open until the end date, because theoretically Truss could become PM again (it turned out to be a surprisingly real possibility with Johnson).

Being able to resolve them earlier and then correct the resolution in the rare case it's needed means the markets operate more efficiently - less locked up capital, more accurate prices due to less discounting, etc

I think in this market I will also count as sufficient being able to change the resolution to a different one after resolution. (Although being able to return it to open would be more ideal, being able to modify the resolution would still achieve the most important effects.)

I think I will also count it if the author is able to request Manifold admins to undo or fix the resolution, even if it requires admin action and there isn't just a button to do it.

Any objections?

I mean, I can go on discord right now and ask Manifold admins to undo resolution of one of my markets. Would that be sufficient or do you have some further criteria that would need to be met?

@citrinitas The admins have no way to do it today. If they build a way to do it, then this resolves YES.

Another misclick: https://manifold.markets/Gigacasting/will-a-nuclear-weapon-be-detonated-5b3361923e29. (This was a huge market)

SirSalty said: "I think we probably are planning to code a mechanism that allows us to retroactively change market resolutions for situations like this. But it might not be in the next few weeks."

Also see:

Can I suggest a better description?

This market resolves to NO. If after that happens I am able to resolve it to YES, I will do so.

@IsaacKing Clever! I think there's a difficulty though - I can't just resolve it no right now because this market resolves YES if I can undo even if only for a limited time. And I can't exactly wait until 12/31 to do the resolve no thing, because what if undo is implemented and then reverted because they found a bug - that ought to resolve YES according to the current description. Still, I like your description

https://manifold.markets/RuthGraceWong/will-the-ftc-interfere-with-adobes#UGaGZD9A17EcwDUorYhR

Another example where this would have been immensely helpful - even a 10 second undo would have been sufficient because I noticed the typo immediately after seeing that the payout numbers made no sense. https://manifold.markets/jack/new-york-home-price-index-for-june

partially to encourage implementation

True undo seems like it would have an awful lot of edge cases. But just waiting five minutes before the market actually resolves sounds very doable.

Good point; @jack would you count a 'fake' undo, e.g. say clicking a Resolve button doesn't actually resolve the market until five minutes later, and during those five minutes it says "This market will soon resolve to YES; click here to undo"

Yep, that would count. One of the simplest suggestions I saw was "wait 10 seconds after clicking the button before actually resolving, with an undo button", which is one example I was referring to in the market description. A longer period would of course be better for catching mistakes but might feel like too long of a delay.

A delayed resolution like that should ping all the traders immediately, so they have 5 minutes to express any concerns.

And it would obviously need to suspend any trading in the mean time.

This is a follow-up to