Skip to main content
MANIFOLD
What are the most interesting facts/opinions/questions about prediction markets?
9
Ṁ800Ṁ550
resolved Apr 6
19%17%
Automated market makers make combinatorial prediction markets possible.
18%16%
You can use a pair of conditional prediction markets to make a decision!
16%15%
Prediction markets with sufficient liquidity are very robust to manipulation.
8%7%
You can generally trust people to resolve correctly.
8%7%
Prediction markets, if setup properly may be used to judge quality of experts (though neither Metaculus nor Manifold succeeded here)
8%7%
There are tentative indications that real-money prediction markets can outperform world-leading salaried analysts in prediction accuracy
8%7%
You can use prediction markets to funnel your money into those future worlds where money is more useful to you.
7%6%
For CFMM, if you place 50 YES and 200 NO on the market-making table instead of 100 YES and 100 NO, someone who knows the answer is YES can only take 50 mana from you, which can be useful when you were trying to sell NO to people. You can do this by initializing the market at a lower probability than your true estimate, then buying YES in your own market.
5%5%
If you put trades in a queue for an hour, you can lower liquidity when many trades go in the same direction, to pay less for suddenly common knowledge!
4%3%
In markets like this the best tactic is to snipe a big bet on an outlier at the last second
4%Other
2%
You can't generally trust people to resolve correctly.
0.0%
If you're poor and nobody believes you about a 100 markets that all resolve this midnight, you can make a single market about whether *all* your hunches are correct and still end up rich by tomorrow!
4%
If you're poor and know how 10 markets will resolve this midnight but nobody believes you, you can make a single market about whether you're right on all counts and still end up rich by tomorrow!
This market resolves to whatever the market probabilities are at market close. Does that make this prediction market impossibly circular? Maybe! It's an experiment. My hypothesis is that it's a bit of a popularity contest but is still interesting and meaningful. Bet on your favorite response and if other traders don't disagree, you'll make money! Mar 28, 3:06pm: Thanks to @Undox for pointing out a sniping strategy for these kinds of beauty contest markets. To fix that exploit let me hereby commit to resolving the market at a random time a couple days before market close or to whatever the probabilities were before any last-second sniping. Close date updated to 2022-04-07 11:59 pm Apr 6, 12:33pm: Also I have to actually agree with the fact to include it in the resolution.
Market context
Get
Ṁ1,000
to start trading!

🏅 Top traders

#TraderTotal profit
1Ṁ83
2Ṁ22
3Ṁ7
4Ṁ6
5Ṁ2

People are also trading

Sort by:
You can weigh answers with a lot of shares lower in your PROB resolution, to give each answer the mana it deserves.. Each share is multiplied by the probability you resolve it to, and then the pool is split equally. (Yeah that's pretty silly.)
Also a true fact. But we'll get there!
In that case I am yet to see a sufficiently liquid market on MM.
Oops, good point. I guess the convention can be to resolve at a random time a bit before the market closes?
True fact! It's very hard for people to come up with objective resolution criteria but you can trust them to resolve in good faith as best they can. (Maybe eventually Manifold will need a reputation system of some kind but I think that's a bridge they can cross when they come to it.)
It's nice how it's quantified though. You know how much someone would have to spend to manipulate the price!
My sense is that it's less manipulable than usual. For example, see misconception #2 in this write-up (which also links to Hanson's original paper): https://forum.effectivealtruism.org/posts/ijohdoDbPvdeXMpiz/summary-and-takeaways-hanson-s-shall-we-vote-on-values-but
Interesting, but I'm not entirely sure it's *true*. I don't know if the incentives on markets for the more probably counterfactual market will be aligned for predicting accurately. It seems ripe for more market manipulation than usual.
Aka futarchy, aka decision markets, invented by Robin Hanson. You pick a success metric and then you make markets that predict that success metric -- one market for each possible world corresponding to a possible decision you could make. The possible decisions should be mutually exclusive. Then you commit to making whichever decision the markets predict will maximize the success metric. The market corresponding to that decision pays out and the other markets are all voided.
"sufficient" is doing A LOT here, as if manipulation is happening then they are by definition not sufficiently liquid. And so far on all I have seen cases of extreme silliness.

People are also trading