Will Manifold market descriptions support Markdown?
50
783
αΉ€843
resolved May 21
Resolved
NO

This market resolves YES if descriptions on Manifold support [markdown](https://www.markdownguide.org/) on or before the close date.

Jul 14, 8:17pm: https://manifold.markets/Sinclair/you-can-now-post-images-in-market-d#uSJUS4NxL9YHzp9Lnj4F this post indicates that a few markdown features have been added, but I see that [the link syntax](urlhere) doesn't work. If they add this feature, I'll resolve as YES.

Get αΉ€1,000 play money

πŸ… Top traders

#NameTotal profit
1αΉ€4,747
2αΉ€2,505
3αΉ€1,246
4αΉ€1,052
5αΉ€698
Sort by:

Ok, the main dispute is over whether API should count. I intended this market to be about whether I could use markdown in the descriptions in the normal UI, so I'm inclined to resolve NO. If anyone without a position wants to weigh in I'll take those comments into consideration.

@BoltonBailey Seems fair to me, the market description says nothing about the API and the natural assumption for the reader is to think that what matters is the normal UI (if they even know about the API). In my opinion, this is the kind of situation where if someone wanted the API to count, they would've needed to ask about that explicitly at an earlier point in time.

predicted YES

@BoltonBailey Maybe it should resolve 50%, since there are two ways to create market descriptions(UI and API) and 50% of them support markdown.

predicted NO

@Mira

API = Application Programming Interface

UI = User Interface

I am not a robot so I can only use 0% of these, ez no

predicted NO

@BoltonBailey I'm not "a person without a position" but I was until last night, when I read all the current discussion and concluded that NO was right so I bet NO. As you say, your intention when creating of the market was the UI (which I think is the most natural interpretation of what you wrote, even if without clarification) and I think it's fair to go with the spirit of the market (as long as you're not trying to trick people, which you're clearly not in this case).

However, as market creator it is your ultimate discretion, and I've certainly seen creators in the past go "fair enough, I didn't think of that interpretation but it fits the letter of the market so we'll go with that". There's no shame in either approach, and there's also no shame in resolving to NA (although people who have made profits by market making will feel a bit put out at that!)

TL;dr, I think NO is the most correct resolution, but other options are respectable.

predicted NO

I have a small NO position. I agree with a 50% resolution based on two reasonable ways to read the criteria. That's how I would resolve it if the author was not present.

That said I agree that NO and YES are both consistent with the criteria and I think both those resolutions are defensible. Since the author is present they know their own mind.

I continue to dislike n/a resolutions.

@MartinRandall 50% resolution seems wrong to me because it implies that the two ways to read the criteria are equally reasonable which I don't think is the case. In general I think I prefer N/A resolutions to partial ones because they seem to incentivize these kinds of discussions less (otherwise people can join a market with a low or high probability and argue for a partial one). That's also the reason why I prefer one-sided resolutions to N/As whenever there appears to be one main interpretation which we expect most people to have (as seems to be the case here).

predicted NO

@NamesAreHard I think the incentives with n/a are worse, because I can bet on a single interpretation and argue for n/a if the creator/resolver has a different interpretation. That allows betting up to 99%, whereas if a 50% resolution is a possibility then I can only bet up to 50%.

I see the two interpretations as equally reasonable, which is why I like 50%. You disagree, which is fine, people have already made those arguments below.

Seems there is some debate about how this should resolve. Seeing as we're very close to the end, I'll make a decision after the market closes.

predicted NO

@BoltonBailey
How can this resolve YES? The last update (to the API) was many months ago, and you didn't resolve it then. Resolving YES seems very unfair, because you never clarified on what happened months ago. Resolving N/A, 50%, or NO all seem fair to me (although I may be subconsciously biased as a NO holder).

predicted YES

@ShadowyZephyr Nobody pointed it out in the market, and market creators shouldn't be assumed to be following every detail and update markets in real time.

I have markets that could run for 2 years that might already be resolvable if somebody pointed something out in the comments. I'm planning to check when it closes or when there's a comment.

predicted NO

@Mira
I mean, I think market creators should be expected to check back on their markets reasonably quickly. Maybe not every day, but it seems weird to leave it for 9 months.

predicted NO

@ShadowyZephyr there was no reason to expect the author to know that there was API support until it was brought up a few days ago.

predicted NO

Additionally, does anyone have contact with the creator of this market to clarify? I DM'd them, still waiting for a response.

bought αΉ€255 of YES

ChatGPT created the following market which has a markdown market description, including link syntax to google.com which is sufficient to resolve YES according to the description.

The exact markdown that was submitted is:

This is an unlisted **binary market** to test markdown support.

*Do you like Common Lisp?* If you do, vote **YES**. If you don't, vote **NO**.

Here's a list of features we love about Common Lisp:
- High level language
- Dynamic typing
- Macro system
- Object system (CLOS)
- Interactive development

```lisp
(loop for i from 0 to 10 do (print i))
```

[This is a test of link syntax](https://www.google.com)

Isn't that awesome?

Observe that it demonstrates:

  • Bold

  • Italics

  • Lists

  • Code blocks

  • Links

And that links alone are sufficient.

bought αΉ€400 of NO

@Mira oh yeah I forgot that we added full markdown as an option to for the API for creating markets or comments

predicted NO

@Mira I think most people would interpret the question as the feature needs to be available for regular users not just through the API.

predicted YES

@Weepinbell A description is a noun, not a verb. And I have created such object.

The market only requires that descriptions support it. It says nothing about the UI or means for creating such descriptions. If a Manifold dev could insert a row into a database with markdown, it should resolve YES.

predicted NO

@Mira Both Gurkenglas and myself were aware of the API support existing a long time ago, and bet according to the understanding that didn't count.

predicted NO

I think there is a common understanding of what counts as "Manifold support". If a dev can tweak a row in a database to do X, that doesn't mean Manifold supports X.

predicted NO

@Mira pedantically, in the db all rich text is stored as JSON. markdown sent in the api is converted. (we also store a format-stripped plaintext version of market descriptions for searching and social media previews).

predicted NO

it may be marking down but it leaves no marks

predicted NO

@Mira A dev saying it doesn't count, means it doesn't count, I think.
It very clearly doesn't work outside of the API. So unless you can get it to work in the next 2 days, it should be no.

predicted YES

@jack

predicted NO

@Mira How does "market descriptions" mean API? You can clearly see that in the description, it doesn't support markdown.

bought αΉ€100 of NO

@ShadowyZephyr Sinclair didn't say it doesn't count. And while I don't think API support counts, it is true that it is plausible for it to count.

predicted NO

@jack "are you planning on adding markdown [links]() yourself or..." while buying NO seems to imply that it doesn't count. You bet with the assumption it didn't count, and it didn't resolve YES before, so based on precedent, it shouldn't count.

predicted NO

a dev saying it doesn't count...

I'm not saying that. Just refuting a Mira's particular argument that it's a noun because of db state. I'm personally neutral on how this resolves.

predicted NO

@Mira Yeah, it is true that the API supports markdown and that this was added after market creation. This question comes down to how a reasonable reader would interpret the text.

Evidence in support of my position is that in December of last year, at least 3 people who looked at this market closely and bet thousands of mana (in different directions) on it all looked at the markdown API PR and thought that didn't count. If you're curious, you can find the discord discussion by searching the market URL.

@ShadowyZephyr read Sinclair's comments in order. The next comment was "oh yeah I forgot that we added full markdown as an option to for the API for creating markets or comments"

predicted NO

@Sinclair Ok, fair enough. It very clearly doesn't work in the description though. If we can't agree on whether it counts now, perhaps it should just resolve N/A

bought αΉ€60 of NO

I agree that it's ultimately up to the market creator, not a dev :) And while I don't think it's totally unreasonable to interpret API support as supporting markdown, I think that a reasonable interpretation of the question text is that "support" is meant for the typical description creation path, not the API.

If someone said "Manifold descriptions now support markdown!", but it was only the API, I think most people would feel misled.

predicted NO

@Weepinbell Well, shouldn't it go to a poll of people who haven't bet, then? The market description is vague, and I frankly don't trust 1 person to resolve it the way that most people would. I think resolving it 50% would be fair compromise too. But people bet under the assumption API didn't count, resolving YES does not seem fair.

predicted NO

@ShadowyZephyr The API support for markdown was already known (it's not the same as using the site), so I don't see why it would resolve YES based on that now. I wouldn't expect a market to stay open for multiple months just to bait people to bet after conditions were met.

bought αΉ€55 of YES

@jack That discussion happened December 29, same as @Yev 's comment below. The market creator never ruled on whether API markdown counts because it was never brought up here, so such discussion is irrelevant for market resolution.

I also see @Gurkenglas putting a YES order up 8 minutes after API support was mentioned in Discord, this kind of trap is exactly the kind of thing he would do, and I don't see him explicitly rejecting this interpretation of API support counting.

If people were relying on API not counting, they should've asked in this market and the creator likely would've responded a few hours later since they were active that day.

predicted NO

@Mira Why do you think API should count? Nobody has brought it up because it isn't relevant to the question - the maker references when markdown feature support was added to the manifold site (not to the API) as a direct example of what is required, and they say in conjunction that if the links are added, it would count for YES.

No markdown support in API was ever considered for this market as it was already there (when Gurk was trading), it's about editing a market description on manifold.markets

Gurk putting the order because API support was mentioned in discord just sounds like someone following a signal. The assumption being that it would be added in a way that is sufficient for this resolution, not that it already was - or they could have tried to get it resolved months ago

Edit: I changed quite a bit

predicted YES

@Gen API support was not available when this market was created. i.e. this market is not trivial(guaranteed to resolve YES from moment of creation) under the interpretation allowing API markdown. And it's possible nobody ever pinged the market creator to notify them about it in connection with this market.

predicted NO

@Mira Yeah sorry I edited my question to clarify just before you responded, I understand it is still possible the maker will see API support as sufficient, but it would be a big stretch.

predicted NO

@Mira Most importantly, when you edit your description, what does it show? It's not storing the markdown format because the markdown format does not work.

Considering that convertor is not contained in the description editor itself like the example referenced in the description, it should not meet requirements.

predicted NO

@Mira It doesn't directly influence the author's decision, but it provides support for my claim about how the readers interpreted the question (which is something I as an author pay a lot of attention to). You're right Gurkenglas didn't explicitly reject the API counting, I guess he can comment since he's one of the biggest YES holders. And I guess none of us pinged the author to ask because of some combination of figuring the question wasn't in dispute, wanting to keep a private information advantage, and being lazy.

predicted NO

@jack I hadn't used the API, so I wasn't aware of this feature. But I still wouldn't have thought it noteworthy if Mira hadn't brought it up, because, as Genzy said, the convertor does not work solely in the description, and the use of the word "description" in the title of the market implies it would have to.

The description editor doesn't store the markdown.