🏅 Top traders

That was close. I count 22 Yes and 21 No.

Unless the creator prefers intent and decides to rule based on intent...

Resolution:

After market close, I will count the number of Yes and No positions holding up to (and including) 100 shares. If the number for Yes is greater or equal to the number for No, this will resolve Yes. Otherwise it will resolve No.

I think there are three factors that should be considered:

• The purely textual reading of the rules. Here is when an author's intention may become relevant - if ambiguity is present, for instance in the form of multiple definitions existing for the same word, then if the author's intention can fit into a potential textual reading, that textual reading should probably be preferred.

• The understanding of bettors in the market. Is there a significant portion of bettors who assume conditions other bettors in the market don't?

• If a reasonable person meticulously went through the market conditions, what would we expect them to interpret them as.?

I think an amalgamation of the things above is what should determine how a market is resolved.

Ultimately, I don't think an author's intentions matter at all, except for how to interpret the ambiguity in any textual reading. Besides that, all that matters is what intentions were communicated through the title and market conditions, and what intentions THE BETTORS had when betting, which is far more important. The following example demonstrates why I think that:

Say someone creates a market titled: "Will Joe Biden come out as gay in 2024?":

• A textual reading could lead to an interpretation of 'gay' being either 'homosexual' or 'happy'.

• I'd assume 99% of bettors, and probably more, would read the market as Joe Biden coming out as a homosexual.

• A reasonable person reading that title, would also come to the conclusion that the market is about Joe Biden coming out as a homosexual.

Given the conditions above, it's pretty clear the resolution of the market should be done based on Joe Biden coming out as a homosexual.

But let's now say that the author's intention was actually for 'gay' to mean 'trans'. What now? Do we scrap everything else because 'gay' is technically ambiguous, and the author's intention was for it to mean 'trans'?

That's my two cents. I admit there's a big chance my position is inconsistent and faulty - I haven't exactly made this position my life's work, and it's only relatively recently that I started thinking about this. And so I genuinely welcome people to point out areas where I might indeed be taking a position that leads to absurd outcomes, or where my logic itself is absurd.

@JamesColiar Title <> Rules. Both matter.

@MaybeNotDepends Yes, of course. In the market example for Joe Biden's gayness, I assume no description and no clarifying comments, for the sake of simplicity.

It depends, for complex or long term markets you at some point need to just reason with uncertainty and accept that it's infeasible to predict and address every eventuality. I could argue that simply including a clause in the form of "In unclear cases I will resolve to my best judgement" now means that all markets resolve "To the letter" only. I think it's more useful to understand that this is implied and was in fact always the case.

There is also utility in being able to say "Will X happen?" and forcing bettors to use their best judgement. Making a market should not only be available to people with exceptional foresight, it's also fine to force some risk of ambiguity on the bettors, as long as this is clear. In many cases, "I can argue Y fits the criteria of X for a creative definition of X" is not actually that useful or helpful, and a truly airtight market description would be unreadably long. Having said that, markets with unclear descriptions are often not worth betting on if they're not made by people with a track record of resolving markets correctly. In that sense, needing to write clear descriptive market resolution criteria is a cost of not having built up a strong reputation of proper resolution on manifold.

Anecdotally, I have made markets which I think have quite good (even excessive, in my view) resolution criteria and markets with poor resolution criteria as a test, neither seems clearly better for all cases. I do think proper market resolution is a problem that manifold has, and I think on the margin it would be better for most markets to have better thought out criteria. I also think that the current situation is pretty hopeless if you speak english as a second language and far too much work is being done by "clear" english phrases which are in fact ambiguous but have only one accepted meaning.

Specifically in the case of this market, I think it's clearly bad to resolve in a way that technically fits the criteria but no traders had considered up until that point, and it's also clearly bad to resolve markets in a way that is ambiguous and dependent solely on the market creator.

In $ prediction markets we use a mix of the two. If there is an obvious typo or logical error, we go with the general intention.

From my point of view - any market that says "Will x happen before Dec 31, 2023" but lacks a start date in the rules - means will this have happened at any time before that date. However, 99.9% of people would argue that if it is something like "Will NYC reopen (pandemic)?" that the 1919 flu pandemic does not apply (a real example from Polymarket - NYC restricted things like movie theaters during the Flu pandemic and reopened them a couple years later).

I used to be into rules literalism more. But I was persuaded to change as 1) the law recognizes intent in contracts and 2) 99% of people are in favor of using general intention combined with the rules.

It isn't always clear where to draw the line - but you need to find that general area.

If the rules are clear, and the market title is simply badly written (Ex. will any governments be dissolved and the rules say "involuntarily dissolved" - a Manifold example) - I strongly recommend going with the rules.

If the rules are completely ambiguous, I'd go with intention.

If the rules lean one way, and the title leans another way - but both are unclear - I'd go with the rules. This is where it gets controversial and traders will only read the rules and lose a lot of money.

If the rules say "A or B" and traders misinterpret this to mean "A and B", I'd go with the literal/logical interpretation of "or" (this happens more than you can imagine).

I think for me the most persuasive situation where markets should resolve to the spirit and not the letter is when the underlying assumed reality changes in an unexpected way.

The rebranding of Twitter is a good example. There are tons of Twitter-related markets where Twitter itself isn't the point. Like "Will Person X tweet three times in 2023" or "will somebody with more than Y Twitter followers say such-and-such" or whatever. It would be catastrophic if every single one of these markets became worthless just because Twitter doesn't exist anymore and instead it's called X.

(Oh gods, I just realised that using X as [arbitrary numeric variable] is now incredibly confusing in this context. I refuse to change it, though. Musk ain't the boss of me!)

Yes, the ideal situation would probably be for all of those market creators to say something like "Twitter or its successor" or to clarify what happens in the event of a renaming. But creators literally cannot anticipate literally every possible change that would change their resolutions on a technicality.

There are some situations where a fundamental and unexpected change in the underlying reality should affect market resolution. It's up to the creator to decide (possibly in consultation with bettors) whether the change is something that the spirit of their market should care about.

@Fion My intent (hehe) was to exclude problems like "rebranding of Twitter", by using the phrase "When in doubt".

@Primer I think I'm right in saying that there has been some doubt expressed in a "will Donald trump tweet in 2023" market for this very reason.

@Fion If Trump tweets and that market resolves "No" with the given explanation being that Trump in fact "xeeted" I will lose my mind.

Not sure I approve of explicitly allowing alts like this, but hey, I'll take advantage of my three votes if you allow it!

@Fion Not sure whether it was a good idea myself. If I could do it again, I'd just drop the line about Alts. The market would be more controversial and if it was close enough, I could maybe resolve it however I like, to the letter or to my intent 😂

@Primer ooh er, gets a bit paradoxical. If it's a narrow YES win, you'll have to resolve to the letter and include my alts and it'll resolve NO. But then you'll be able to resolve according to your intent and exclude them and it'll resolve YES!

What I really prefer is for creators to make sure the letter of the market is identical to their intent.

@evergreenemily Well yeah, that’s what everyone prefers. 😉 The premise is “when in doubt”

@evergreenemily I don't think this is possible, I don't think any language, but certainly not english, can be used to communicate ideas losslessly.

@Sailfish Certainly not losslessly, but I think "criteria thorough enough that there's no ambiguity in how the market resolves" is very achievable.

@evergreenemily Depending on the market I don't agree, there are markets on very nebulous concepts which are poorly defined or undefinable that are nevertheless useful. I challenge anyone to write a market description they don't think can contain any ambiguity at any point in the future about AI. For now I think it works because everyone is mostly on the same page, acting in good faith, and is willing to collaboratively work through any disagreement.

Maybe more explicitly, I think it's always possible to argue that any X is not in category Y, because there isn't any oracle that will spit out "True" or "False" if queried. I think this is just a fundamental limitation. I would absolutely agree that ambiguity should be minimized, but if your goal is explicitly zero ambiguity, you're stuck with writing out a literally endless list of definitions and distinctions and caveats.

For instance, consider any market that resolves by any time. If they're not specifying exactly what calendar they're using then that's "ambiguous". It's up to the market creator to say "No, that's silly, we're obviously using the Gregorian calendar and that was clear without it being made explicit"