Show HN: Create your own prediction market in two minutes (manifold.markets)

84 points by akrolsmir ↗ HN
Hi HN! Excited to show off the project we've been working on for the last couple months. We started with an idea for a crazy twist on prediction markets: You come up with a question for traders to predict, and then decide the outcome yourself.

For example, you could create a market on “Will my date with [X] go well?” Anyone can bet on it, and the bets create a forecast on the chance your date goes well. After the date is over, you get to judge the result and reward the traders who picked the correct side.

There are so many ways for this mechanism to go wrong: the creator of the market can be dishonest in deciding the outcome, or you may just disagree with their resolution, with no recourse. Nevertheless, we pitched user-created prediction markets in a grant proposal to the blog Astral Codex Ten, and somehow we won!

Since then, it’s been an exciting two months of us hacking away to realize this idea, and now we’re happy to announce the official launch of Manifold Markets! We’ve gotten so much support from early users who are just as passionate about prediction markets as we are.

You might ask: Is there a reason to create a prediction market, and not a Twitter poll? Are outcomes really that different from a simple person-by-person vote?

We believe so. The magic driving prediction markets is accountability. When wagering a scarce currency, those proven right live on to make future bets, whereas those with less-savvy bets have their influence diminished. Prediction markets succeed because they reward accuracy, and that makes all the difference!

Our goal is to make prediction markets an order of magnitude easier for you to create and share. It should be as frictionless as a Twitter poll. As part of that philosophy, we’re launching with a play money currency, which we believe is just as fun and predictive.

We’ve already built up a passionate community of predictors and market creators, including writers like Richard Hanania and James Medlock, who have predicted everything from CDC recommendations to newsletter subscriptions to fatal shark attacks.

There’s so much unexplored space to ask questions and get valuable forecasts. For example, with conditional markets, you can create several related markets that help you make a choice based on which has the highest likelihood of success!

Let us know if you have any thoughts or suggestions; otherwise, looking forward to what markets you create!

75 comments

[ 2.9 ms ] story [ 60.3 ms ] thread
Today's the launch day for Manifold Markets, you can read our announcement here: https://manifoldmarkets.substack.com/p/above-the-fold-manifo...

We've been working for the past two months to make the easiest-to-use prediction markets! We want to make creating and trading in prediction markets as frictionless as using Twitter polls.

It's been a very exciting and fulfilling project to work on given the dedicated fans of prediction markets that have already joined our community.

Happy to hear any feedback you have on our site!

One complaint is that it requires a Google account to sign up. It does make sense why this is a feature, but it's still kind of annoying.
Thanks for the feedback! A Google account makes the sign in flow a little bit simpler, and lets us to find a reasonable avatar to attach to each user. But Twitter or email auth are things we're interested in supporting (just want to balance those against it being easy to create sockpuppet accounts)
As someone who just learned about predicition markets through ACX, I'm pretty pumped about them! Personally, I do not have a Google account, so was really disappointed to be excluded. Really hope you guys figure something out! Will be keeping an eye on this.
>so was really disappointed to be excluded

You can make a Google account for free here: https://accounts.google.com/SignUp?hl=en

If they don't have one in 2022 it doesn't look like they'll do it for you. Would advise to not take it personally (at least not about you, it's probably personal about Google)
I don't think this person has anything to do with the prediction website, just a drive-by commenter.
Thanks for the data point. Happy to hear that ACX is teaching people about prediction markets! We'll definitely look to supporting other auth methods; and until then, you don't need to sign in to view any of our hundreds of markets!
I think having a plain email signup/login would be best, especially since a lot of people are moving away from Google and going to alternatives.
(comment deleted)
fun, but "Our goal is to make prediction markets an order of magnitude easier for you to create and share" is a lie. your goal is to take a % lol
Every “doing it for the community” project makes me think this.
Glad you enjoy our site! To be frank, we do plan on monetizing by taking a percent of profits (right now, we're modeling a 1% fee on trading profits, much lower than traditional betting sites). We do think it's important to set up incentive systems for all participants to do the right thing, and this includes ourselves (the platform) to support continued promotion and feature development. However, we also all really, really believe in the usefulness of prediction markets, and hope that this turns out to be a valuable public service!
ive been listening to people trying to convince the world of prediction markets utility since the old iowa electronic market days. it would help if you had like... a single example. i find them fun, but i dont think they offer all that much beyond that, sorry
So...I take it you're...against democracy?

I mean, it is basically the same dynamic. Right?

People vote based on a prediction of what they think will be a "good" outcome for themselves, the ones they care about, and probably society as whole (for some folks anyway), and they [spend money] to do so. Erm..or that's the theory...let's ignore all the misinfo, short-term emotive-not-substantial-issue optimization, special interests...etc

Basically, your "single example" is the governing system of most of the 'Western liberal' world.

What do you think about that?

[spend money]: People are spending real money to vote (not just in supporting donating to a candidate) but by how they expect their choices will cost them or reward them financially (and of course there are other 'non fiscal' costs and benefits associated to different choices). I guess you could argue that's different but to me it seems it's pretty much the same.

I don't think that's a lie or a good example of one. You can both have a desire to make something easier for people, and to profit from that. Each does not automatically exclude the other.

Isn't that sort of like saying, "You say you love that man but you don't really, you just want to marry him so he gives you money."? Which can be true...yet you say you know it's a lie, but you don't know that.

Their completion doesn't take a cut; you only have to pay blockchain gas fees.
Sorry I don't really understand that. What does it mean?
Cool, I might use this, thanks!
Awesome; let us know what you think! We're always interested in hearing from new users, especially if there's anything that is confusing that we can fix.
What is the currency used for the bets? Can I cash out back to USD? Why not just use actual money or even stable coins?

The idea is cool, but I’m never gonna put money into something where there’s ZERO documentation for what happens to my money once it’s in your pocket.

There's no documentation about that for pretty much any SaaS out there. Any of they could literally disappear with your money and shut the shop down or change its name to continue defrauding others...

Edit: you can downvote, but it's true...

Well most SaaS products provide value as you pay, or even before you pay. You're not expecting to get money back once you give it to them
This is objectively untrue. Show me a respectable platform for payments, trading, gambling...etc that do not have documented policies regarding user funds.
If you don't trust the site author, why would you trust something they've written in their own site?

My idea of trustworthy documentation on this sense would be an escrow or something like it.

They give you 1k it’s just internet bux, alpha test not real money
It's funny money to begin with. Even if you were based in Gibralter there would extensive licensing to get approved to run a gambling site.
Great question! Right now, Manifold Dollars (M$) is just internet currency. The reasons for doing this were:

- Much faster for testing out new market systems

- Lower barrier to entry for new users signing up

- Less regulatory risk for everyone involved

But we're actively looking at both the crypto space (for stablecoins), actual fiat like USD, and alternative withdrawal mechanisms (e.g. sending your Manifold balance to 501c3 charities). Attaching real value to these bets is very much something we're interested in!

I love the idea of playing for charity. I’m pretty unmotivated by play money but I would absolutely play for charity if I thought that there was a good chance that my choice of charity would be much better than the average persons.
Um, aren't prediction markets backed by real money flat out illegal in the US, except for a few specific exceptions like political races? I hear there's a prediction market that allows you to use real money in the EU but not in the US for this reason.
I'm not sure I understand. I pay you guys money, and get points that do nothing except for allow me to gamble for more points?

Just to be clear I'm not trying to shit on your product. In fact, if your platform allowed me to participate with real money I would 100% use it.

What do I get for buying access to the service?

A lot of online gaming and mobile gaming works like that. I would say to me this is more interesting (but still doesnt pass my threshold to buy it)
Yup, right now "free-to-play game currency" is a good way to think about Manifold dollars. We'd love to support a real-money equivalent at some point though; stay tuned for updates on this front (https://manifoldmarkets.substack.com/)
No money this is alpha. They supply 1k for you to play with
They give you 1k in funny money. There is no payment this is alpha
This is great! Placed a couple of wagers and created a market.

Have you read Shockwave Rider? John Brunner envisioned a network that placed odds on everything, just like this.

Not quite 'just like this'. Brunner wasn't talking about prediction markets, but essentially Delphi method panels (https://en.wikipedia.org/wiki/Delphi_method) which would provide odds for betting (hence the name 'Delphi pool'), but continually rather than merely being convened for a specific issue. Think of it like Las Vegas bookies: there's no 'market', there's a centralized bookie with its own hired guns & expert bettors setting the odds, which may, at its discretion or financial needs, change those odds based on bets its users make.
Prediction markets : The Delphi method :: Classical liberalism : Liberalism
> you could create a market on “Will my date with [X] go well?”

Why do I get the feeling one of the underlying key variables here would be how likely [X] is to like prediction markets?

What about no.

I absolutely cannot hate this project more.

You don't have to use it. That's how I feel about Facebook and yet it still exists.
And I can hate its existence.

Facebook has an effect on the society you live in and on YOU despite you not using it.

One thing this makes me think of is "Oracles" that could potentially decide a market. Someone could write an oracle, and open source the code, and MM could then run the oracle for the lifetime of the market.

The oracle would be responsible for resolving the bet and would turn an ambiguous statement into a well defined one. i.e. What you are betting on, ultimately, is how the linked oracle will resolve the bet, and everyone can see the code.

The first bet I see is about Russia invading Ukraine. We could have an oracle that checks Wikipedia's list of Russian military engagements every day and if there is a new one in Ukraine and it stays for five days that might be a sign. Another sign might be scraping New York Times headlines for a list of terms you might expect to see in an invasion. Maybe there's a website tracking Russian troop deployments, or Russian flagged planes and ships and those could become signals. Then, the oracle collects and judges based on its signals.

Yeah, oracles are a common solution to the question of "how should the prediction market get resolved". We think there are some usability problems with oracles in practice, which is why most prediction markets (e.g. PredictIt, Kalshi) still end up relying on platform resolution.

That said, we'd be very happy if anyone was interested in setting up an oracle on our Manifold! "User-resolved" includes the possibility of "user publicly commits to resolving all markets via an automated, predefined oracle".

Oracles are cool, but I humbly submit user-resolved markets as a cleaner and better mechanism.

As akrolsmir says, someone could commit to resolving according to an Oracle.

But mainly, it allows for a diversity of approaches to serve different user and subject-matter needs. Some markets can be resolved subjectively, others need to rigorously specify each case. Plus, the real world is often ambiguous. You need someone you trust to make the right judgment call.

User resolved simply won't work if people care about winning their currency. And people will care if they ever use real money or become popular.

Imagine - I create a market "Donald Trump will be speaker of the House in 2022" and people bet on that. I, through sock puppet accounts if necessary, bet that he will as much as possible. Then, I just resolve the bet so that I win.

Cheaters will have all the currency and cheating will make the predictions useless or worse than useless.

It sounds plausible, but our experience so far is that cheating is very rare. Out of hundreds of markets, we had our first clear incorrect resolution yesterday.

But the argument is not simply that people won't cheat often. The argument is that people will choose to bet in markets where the creator is trusted. Even if the majority of creators cheated, users would react appropriately and flock to the ones who don't.

It's just like how you purchase goods and services. They often have a chance to keep your money without providing what they promised. But in practice, they can't run a long term business that way.

By the way, even in our play-money currency, the market creators earn a fee on the traders' profits.

I bet this specific example will be tricky, invasions/incursions now a days seem to be usually denied by the aggressor in various ways, i.e. there might actually be a debate, 'Has Russia actually invaded Ukraine yet?'
>> turn an ambiguous statement into a well defined one

A well-designed series of forms would take away the friction of having some dude come in later and tell a moron what they meant to say.

Unfortunately, a form would also deter morons from creating contracts.

Ah. But what we have here is a would-be casino operator with a short-term view, who would love to have their site full of contracts, even if they're free-written by morons.

God-for-fucking-bid we deter some idiots from starting contracts by making them write down logical rules. Let's get as many idiots as we can!

It ain't well thought out. It's not designed to be able to deal with the public. If it makes any impression at all in the media, it will just cast more negative shade on prediction markets, because the people behind it didn't think for a second about how humans fucking work.

Actually, they're a prime example of dumbass greed themselves, since they did it as a result of a stupid idea winning a grant to build...a stupid site. Wonderful.

GO MAKE SOME BETTER GENETICALLY MODIFIED RICE, YOU OVERPAID LAZY FUCKS.

> We could have an oracle that checks Wikipedia's list of Russian military engagements every day and if there is a new one in Ukraine and it stays for five days

See, that's the issue. Wikipedia already has the Russian-Ukrainian war, because Russia has been occupying Crimea for years.

Wikipedia as a source of truth when money or power is on the line is also difficult. That's why topics frequently get locked.

The NYT would work as a source, but keyword matching could be iffy.

A nominally third party site (again, if real money is on the table) could easily be a fake set up by someone heavily invested in one of the two answers.

That shouldn't be used to say "ground truth isn't available", I'm just not sure it's machine readable and scalable yet. It's possibly I'm missing something obvious.

Could be interesting for the real estate sector
At first glance I thought this was a self-hosted thing, which would be really great. Still a fun idea though, best of luck. Someone should do an open source version of this, if it's not already,
Would love to hear about your interest in self-hosting! We've consider supporting that use case for e.g. software teams who want to set up markets on sensitive data like revenue forecasts or user growth numbers. If you'd like to chat more privately, reach out anytime to me at akrolsmir@gmail.com!
Yeah it probably makes more sense as an enterprise product. I'm curious if that's already a thing, I haven't looked into the space - corporate/internal prediction markets. I don't have much of a use for it personally but I like your thinking.
The twin facts that the market maker gets to decide the outcome, and there being no binding requirements for how something closes, makes it an exercise in absurdity. By leaving the decision up to them, you've made it impossible to measure the variables independently, so you'll never even be able to draw any useful opinion data out of it. It's not that hard to write a trading platform. I thought of an open prediction platform like this in 2011. The really hard part would be to write a framework in business logic and code that lets people design and resolve prediction contracts in a clear, enforceable, stable way. I'd personally be embarrassed if I wasted time building something like this just to say bluntly that there are a hundred ways it could be taken advantage of. If that's the case, you've thought just enough about the real problem to know that you can't solve it, and you want points for coding the easy part? Garbage.
I definitely understand this frame of mind. On the face of it, it seems quite absurd that anyone can do anything, e.g. insider trade, or use sock puppet accounts to manipulate information. In practice, we've seen it work okay so far!

One key realization is that individual market creators on Manifold are given the same powers that existing prediction market platform or casino bookies are trusted with today. A trader is always placing some amount of trust in the bookie to not manipulate the own market, or pricing that risk into the transaction. We hope to make this risk clear and up-front, and to reward market creators who repeatedly adjudicate markets in a fair way.

(comment deleted)
You've actually got it backwards.

User-resolved markets is the brilliant part of our platform, but most people are not able to figure that out.

(comment deleted)
This sounds like (hopefully a better executed) PredictionBook. [1] Definitely worth doing, as PredictionBook's UI leaves a lot to be desired, but I was confused at first seeing it described as a totally new concept.

Where can I follow updates on it so that I can join when there's Email signup? Will you be posting to the Substack for future updates too?

[1] https://predictionbook.com/

Yup, you can join our Substack to hear about updates (https://manifoldmarkets.substack.com/), or come chat with us on Discord (https://discord.com/invite/eHQBNBqXuh)

PredictionBook was definitely one of our inspirations! They are somewhat out of date with respect to UX; and we've also spent some time innovating on a new Automated Market Maker so that our markets work well for differing levels of trader volume. We hope to bring PredictionBook's core ideas to a much wider audience~

Ours is also market-based, rather than simply recording your prediction.
I’ve wondered, instead of betting on “truth” why not bet on the outcome of a webscraping query, where the code is open source?

Example, I publish some code that scrapes the outcome of an upcoming sports game. It returns a True/False payload.

Bettors can examine my source code and see the exact time when it will be run. They can then place their bet on the code itself.

Sure the code could point to a bad actor and someone could pull the rug, but you could also take the consensus result from a few trusted sources, e.g. Google sports, ESPN, etc.

Yup, this is the idea behind blockchain oracles. At the end of the day, you're still making a judgement call behind who you want to trust to represent "truth" e.g. when deciding that Google sports is more trustworthy than ESPN. We think this approach makes sense for certain kinds of markets, especially those that occur more regularly and are well-formed.
In trading, there are penalties for insider trading. In sports betting, teams get disqualified for match fixing. So there is protection against adverse selection. What about here? If the bet is "Will Apple air an ad at the Super Bowl," am I violating the rules by working at Apple? You really need users to trust that everyone's on an even playing field.
The whole point of a prediction market is to try to uncover hidden information like this.

I remember this idea was floated after the 9/11 attacks. A spike in a terror attack odds would be an early warning system as someone with inside information on a terror attack tried to profit from it. This idea was shot down somewhat quickly though.

It is really a different animal than the stock market in a sense although the stock market still operates on information vs noise trades.

How similar is this to Kalshi?
They have a lot in common! Kalshi and Manifold Markets both offer prediction markets: places where you can come bet on future outcomes.

Kalshi accepts real USD, and all questions are listed and resolved by the Kalshi team; Manifold works with play money (for now), and allows anyone to list and resolve markets.