Hedging can be a motive for trade, and if the bet event is correlated with one of the few market "factors" that have risk premia, then the market probability for that event will weight high risk states differently. The prices can still be informative then however, and most claims don't correlate much with any of these factors.

Expand full comment

I'm confused about your point re betting prices. If the outcome of A is correlated with other valuable things e.g. stock market prices, then betting on A can act as a hedge, distorting the price away from the underlying probability.

Expand full comment

I like Hanania and Hammond, but a few mentions from each of them is FAR from sufficient for this purpose.

Expand full comment

Yes, it is far from making the play money feel valuable and induce a sense of stake. I appreciate your skepticism!

What I meant is that it could be a self-sustaining use case. That a creator and their followers would enjoy predicting, trying to get on the leaderboard, and in the process create valuable forecasts that would encourage them to set up more markets and share them further. The experience would have to be compelling.

We're giving it a shot, in any case!

Edit:To elaborate a bit more, this is a network effects problem! Play money prediction markets are only useful with many traders that are trying to win, so the predictions are accurate.

Our strategy is to solve the network effects problem by bringing on a creator with a following.

Expand full comment

Thanks for the correction anyway.

Expand full comment

Sorry, getting some writers with many readers to talk about some market questions seems FAR from sufficient to induce the sense of stake I think is required.

Expand full comment

Thank you for the reply! I'm sure you're right that it's not easy to get play-money markets started.

One strategy we're trying is to get market creators to participate with their following. With a Substack writer or prominent tweeter, there may be enough traction between them and their audience to make the markets useful.

For example, we've talked with Richard Hanania and he's created a few markets and is looking into running a mini-tournament on China policy (I believe) with his readers. https://manifold.markets/Ri...

A second strategy is leaning into viral social media. We already saw a spike in traffic when we got Sammuel Hammond to create a market on when Build Back Better will pass by March and tweet it. https://manifold.markets/Sa...

On judging, I think we mostly agree. For our site, there's a marketplace of creators and people can choose to bet in markets where they trust the creator. It naturally serves every niche for how people want their markets resolved, while remaining simple.

We still have a long ways to go, but I'm hopeful!

Expand full comment

This is me strongly encouraging you to make more videos like this. I wish the thumbs up button on your website still worked.

Also, though there's nothing that you can really do about this since you can't change a posted YouTube video, but at 5:20 you say "Probability of A given B", but it shows p(B|A). Sorry for being the way that I am.

Expand full comment

Hundreds of play money markets have been created so far. They usually fail to create the impression that participants have enough at stake to credit winners with sufficient prestige to induce enough participation. It is logically possible for such markets to do so, if they became large and focal enough. But it turns out to be very hard to achieve that, and merely writing software then announcing a play money market is far from sufficient.

On judging I'd rather that there was just a clear brand on who would judge a question, so that participants could choose how much to discount it by that brand.

Expand full comment

Nice video! It's a shame how little prediction markets are used for decision making today. I would love to see your fire-the-CEO markets realized.

We have a new idea for how to popularize prediction markets, for which we recently won an ACX grant from Scott Alexander.

Our idea is to reduce the friction in running prediction markets as much as possible by:1. letting anyone create a question and later resolve the outcome according to their judgment, and2. using play-money.

If it's as easy to create a prediction market as a Twitter poll, wouldn't that be useful?

Scott has called our idea the "Chaotic Evil" version of prediction markets because of the freedom of the market creator to choose how they resolve it (and thus possibly defraud traders). We think it streamlines all the edge cases when you can trust the market creator to make a reasonable decision.

We'd love to know what you think about our site, and if it's something you'd find useful for answering research questions.

Our site: https://manifold.markets

Here's an example of a conditional market we made for our own company's decision-making. Note it could resolve YES, NO, or N/A: https://manifold.markets/Ma...

Here's a group of markets on Scott's 2022 predictions, with a leaderboard! https://manifold.markets/fo...

Expand full comment

Thanks for posting that educational and interesting presentation.

Expand full comment