This is part of a series of markets about things I could do in 2024. See also: /Mira/which-of-mira-s-cool-ideas-will-mir
Mirafold
Imagine Manifold, but with leverage, no currency inflation, options trading, and portfolio margin. You pay to enter tournaments and the best predictors get prizes.
To prevent news trading, scores are calculated using an integral of predictive accuracy. The guy that placed a bet 10 months ago gets a 100x larger share than the person betting the day before. And the news traders get nothing, because the score is calculate up to the moment the resolution is known. (Soft news might get a retroactive "reduction in uncertainty" penalty)
Each set of high-quality markets is in an isolated group with its own currency. So you can compare predictive accuracy within specific tournaments, but you can't "spend the money" elsewhere without trading it(which has its own prediction interpretation).
There's a sandbox group with free money printing similar to Manifold for people that just want to ask questions without thinking about money being meaningful. You can make your own groups and set how currency is issued and control what sorts of exotic features(options, margin, leverage, etc.) are enabled. New features can be rolled out into the "beta" group.
To seed the site, we steal all the Manifold markets into the sandbox and then poach their users by running ads on their own site offering bounties to try Manifold V2. (not the final name)
Also it won't lag so much.
Market Mechanics
Trigger condition: Deploying any application website to a domain name. Planning, writing prototype code, a static placeholder website, etc. will not be sufficient as minimal effort for this market.
Resolves NA if the trigger condition is not met. Otherwise, resolves to the logarithmic interpolation between nearest bounding answers(or to 1 if I start this but literally nobody else uses it). Units are "number of separate accounts". I'll add new options at multiples of 10 as needed before resolving.
Example: 500 people sign up. The nearest options are 100 and 1000. Then this market resolves (ln(500) - ln(100)) / (ln(1000) - ln(100)) = 70% 1000 and 30% 100.
Related questions
It’s difficult to find a balance between news traders autowinning and the meta being to all in something on day1 (as your 100x multiplier might encourage)
Here is @SG‘s working proposal for the 2024 elections cycle tournament. You might be interested: https://manifoldmarkets.notion.site/Trading-tournament-payouts-840344fe7e214250a207cf3e6e8646c7
@Soli It's actually a lot of work getting the details for all of these right. Like several days of effort. So conditional on me putting that effort in, it's more likely to be something I do on my own.
For the integral scoring: Currency is approximately a "weight on your opinion that you can choose". The price is also a signal, but the amount you're willing to bet at each point on the liquidity curve also says a lot and is a different signal from the price. If somebody predicts only on long-term markets and you wait until the end to realize a profit, they would have a very low weight despite being a good predictor. So they need to be able to unweight their opinion to reallocate it elsewhere. But if there's scoring penalties based on news, if you've been betting against the market, there's a lot of work to figure out the right rule.
For all the financial tools: I don't expect it to be anything fundamentally new or difficult, but there's questions like: "If somebody buys an option to reduce their risk, and then uses the freed margin to take a leveraged position, and then the option is close to expiring, when do you force-liquidate their position?" and the software needs an answer for every variation of that. Someone also has to be very careful implementing these: I would not trust most devs to properly test their code here.
Groups with currencies: This one has come up before. I don't think they're interested.
@Mira Yes it does sound like a-lot of complicated but fullfilling work. I think there are easier products to build that would make you money and challenge you. This one seems to be a-lot of work and realitvely low reward to me.
Would be awesome if they just pay you a bit of money and you help them figure out the most important details then you hop on the next thing.