I’ve got a bunch of ideas for stuff I could do to help Manifold out. This is an umbrella project for all of those.
(I’m also planning on doing a Manifold Live Show with Joshua!)
What are this project's goals and how will you achieve them?
Great question. This market covers what I will do; this market predicts which ones will lead to this project being funded.
To summarise my main ideas:
Open-sourcing @NcyBot: I have a relatively effective arbitrage bot (although his current league standing is kinda bad because I messed up). I think it’s the only arbitrage bot that can currently trade on multichoice markets. I think Manifold would be improved by making it open-source, both by enabling others to arbitrage any markets they like but also by providing a starting point for an all-purpose Manifold trading bot that anyone can use. I’d like it to also be able to take a user-given list of markets and estimated probabilities and determine the best way to allocate their balance.
Automated Fire-the-CEO markets: Robin Hanson says that if he had a million dollars, he’d fund a group of markets predicting the stock price of each Fortune 500 company conditional on the current CEO staying on, and on the CEO stepping down (as well as a binary market on whether the CEO will step down). The idea is to make it obvious when a CEO should be fired, eventually giving shareholders enough info to sue when a company board ignores the markets’ advice: “A revolution in CEO accountability would then be complete, all for only a million.”
This seems simple enough to automate on Manifold. All you’d need is a bit of code (which I’m confident I can handle), a reliable source for stock prices + the Fortune 500 + current CEOs (which I’ve briefly looked for but not found to my satisfaction), and about M$150000 per quarter (which is probably why no one’s done it yet).
This is a high-risk high-reward project: it could be the killer app for prediction markets with its own heavily-visited spinoff site, or it could be 1500 markets with very little activity.
Making a market to predict what will improve my mood over the coming months and writing about it: I’ve been regularly measuring my mood on a scale from 2 to -2 for the past month, with the average tending between 0 and 1. This maps very well onto a binary market, allowing me to use one of Manifold’s new binary multi-markets to predict what my mood will be over a given time period, conditional on doing any number of activities during that time period. In theory, answers that would improve my life should have higher probabilities. Hopefully, this market will make my life better and more interesting in a way that will make for a good blog post/sequence.
If this works well enough, I think it’d make a great spinoff website a la Manifold Love: maybe Manifold Life, where experienced traders trade on what will help you be a better person and live a better life.
(There’s another option on the markets linked above where I might use a different metric, because I’m not sure that moods are the best thing to optimise. Ideas for metrics are appreciated.)
Other very rough ideas:
Manifold arbitrage web app: instead of (or as well as) open-sourcing my arbitrage bot, I could build a web app allowing anyone to select markets that should be equal and pressing a button to arbitrage them. I could try building this directly into Manifold.
Standardisation of common resolution criteria ambiguities, to give a Schelling point in arguments over what a vague market description means (which would also let people type shorter descriptions, if they can be understood to be operating by standard guidelines outlined elsewhere).
A feature on Manifold that gives ‘free’ (loaned) bets on any market that has gone long enough without a trade, to incentivise keeping all markets fresh.
How will this funding be used?
Depends on what I do, obviously. Some very rough cost estimates:
$720: Open-sourcing @NcyBot
$300: Fire-the-CEO markets
$100: Mood market & blog post
$500: Arbitrage web app
@NcyBot has made about M$8000 in about 4 months, so at M$2000=$20 a month and a 3-year repayment period I’d happily open-source him for $720.
Automated Fire-the-CEO markets will cost a fair bit - to go by Hanson’s specification, it would cost M$300 per company if I used a binary + a (old style) numeric market (well, two numeric markets with one N/A’d), which would cost M$150000=$1500 per quarter for every company in the Fortune 500, not accounting for trader bonuses. (I could also use a multiple choice market, which would probably cost a similar amount.) I’m convinced that with enough activity this would be a great mock-up of Hanson’s basic idea, but 1000 markets is a lot, and the only realistic way you’d get much activity is with bots (which wouldn’t give trader bonuses). If I were to try this, I’d probably start with just the Fortune 100, which would cost M$30000=$300.
If Manifold grant me M$30000 (or M$150000), I’ll just mock this up and see how it goes.
Funding expensive-but-effective lifestyle interventions would likely help demonstrate the efficacy of my lifestyle market idea.
Funders can comment to earmark their money for a specific project if they’d like.
Who is on your team and what's your track record on similar projects?
This is an umbrella project for everything I pursue on my own related to Manifold in the period covered by the Community Fund, so the team is just me. I have a very strong track record of doing whatever I feel like and seeing if it works.
Okay seriously. I’m pretty experienced at making markets on Manifold, I’m a pretty good writer, and I’m a good enough coder that I built https://manifold.markets/ancient back when that was relevant.
What are the most likely causes and outcomes if this project fails? (premortem)
Again, it obviously depends on what I pursue.
My arbitrage bot might fail to get any users or contributors, and might not have much effect due to the presence of several other arbitrage bots on the site already (including itself!)
If Fire-the-CEO markets fail, it’s because the markets didn’t get many traders, possibly meaning that the markets conditional on the CEO stepping down get little movement, becoming totally uninformative. Subsidisation and boosting would help a bit on this front, but I might have to face the fact that most people who want to daytrade stocks won’t do it on Manifold...
My mood might not be very affected by most interventions, and there might not be enough ambiguity in whether I’ll go through with any given intervention or not to have a noticeable effect
What other funding are you or your project getting?
I have no other funding source for these projects in particular, unless you count trading bonuses on their corresponding markets.