I like the varied and different tools. I'm a bit worried about the minimum funding bar not being hit.
Charlie's Forecasting Tools
Project summary
A host of trader tools to enhance the forecasting ecosystem. I will work to build and maintain working versions of the following tools based on interest:
Manifold Tools so far. Mostly just an encrypted API key-storage handler, and a front-end mockup for the recurring markets tool. API key handler is 85% complete. Recurring markets front-end is 75% complete, and the backend that will handle the scheduling is 10% complete.
Proposed Manifold-Centric Tools:
Managram Subscriptions* (25hrs)
Allows you to set up recurring mana payments
Would encourage more uses for mana and mana-funded microproducts to be developed
*UPDATE: Managrams are going away. I will reach out to Manifold about whether they would be willing to make something like Manifold Subscriptions possible.
Recurring Markets: (15hrs)
A web app view to allow market creators to easily create and maintain recurring markets on Manifold
I've already started working on this. (See image above)
Manifold Arbitrage Bot* (20hrs)
Tag this bot in the comments of one market with the link to another market to set up automatic arbitraging between the two for your account.
*UPDATE: @wasabipesto has wisely pointed out that being required to publicize one's arbitrages is not ideal, so this likely operate privately.
Committee-Resolved Markets (and other mechanisms for someone other than the creator to resolve a market) (20hrs)
A tool for markets to be resolved by multiple users, consensus, majority, etc.
Can give various permissions to certain users (who aren't a mod) to resolve your markets.
Manifold Commitment Market Manager: (35hrs) (or 20hrs if Recurring Markets is ranked above this project)
An extension of recurring markets, this will allow users to set personal goals that they want to achieve. Easily have recurring commitment markets created on their behalf based on their goals.
The weekly commitments will ideally dynamically adjust based on the user's progress and try to get users the best deal on their commitments (i.e. set smart limit orders that are likely to get filled).
I will also explore infrastructure for easy reporting of goal progress (which attracts more traders), like SMS or WhatsApp.
This idea builds on my learnings building Emile.chat, a chatbot for weekly bedtime goals that adjusts to your progress
Super Secret Private Manifold Markets (20hrs)
Manifold has discontinued private markets, but what if they could be private out in the open?
This idea is to create a tool whereby private groups and markets can be created in a way that are readable on the original site, but encrypted gibberish on Manifold
The markets could be unlisted on Manifold, but still make use of their many trading tools.
I could possibly figure out how to make a Chrome extension that would allow you to see the private markets you're privy to right on Manifold itself.
AI Resolution Criteria Creator (15hrs)
I have several successful markets that I would not have had the confidence to run had it not been for a strong, exhaustive AI-written resolution criteria.
This will be a tool either built as a Chrome extension or a separate site that will auto-create strong resolution criteria based on a Manifold question text.
AI Market Creator and Resolver (25hrs) (37hrs with add-on)
A bit more controversial, I believe there is a place for wholly AI-managed markets.
If this tool is funded, I will build a tool that allows you to manage a market creating and resolving agent that uses tools like SearchGPT to resolve markets (all of this disclosed in the market description, of course)
Add-on: connect this to Twitter or Manifold for even more seamless transitions from internet debates to relevant prediction markets
Manifold Alerts (@manifoldmaven enhancement) (10hrs)
various improvements to my @manifoldmaven twitter bot
*Manifold Loan Infrastructure (30hrs)
A tool for managing lending and borrowing mana to other Manifold users
*UPDATE: Managrams are going away, which was the primary way loans were conducted. This project seems DOA unless Manifold mentions that they think it would still be doable and are interested.
Manifold Tipping (17hrs)
A chrome extension that introduces an interface that makes tipping easier and displays all tips a comment has received
Proposed Non-Manifold Forecasting Tools:
Predictor's Digest (30hrs)
A news site for the biggest changes in the most important prediction markets
Essentially a dedicated website for the types of updates that my twitter bot @manifoldmaven tweets out, with AI summaries of that actually new behind the prediction market change.
*UPDATE: An MVP of this could build off of my @manifoldmaven Twitter bot and be a daily or weekly newsletter digest.
Personal Prediction Tracker* (17hrs)
Most things that we might like to be better at predicting are not necessarily something we would want to create a prediction market for, but we can still learn from tracking personal predictions.
This would allow for one to very easily note down a quick prediction and probability for later evaluation
I'm envisioning folks able to make many personal predictions a day to quickly see what aspects of their life they are well calibrated about and about which aspects they tend to be over or under-optimistic.
*UPDATE: as @wasabipesto kindly pointed out, Fatebook exists and is awesome! If this project is voted highly, I will dedicate my time to a few ideas for Fatebook add-ons that would round out my ideas for a personal prediction manager:
Even more frictionless input and review:
Create a prediction with a single blob of text (AI pulls out date, probability, and prediction)
A quick-resolve feature for mobile via swiping
A quick update feature to quickly scroll and update your predictions
Fatebook on WhatsApp??
More qualitative views and insights from past Fatebook predictions
Auto-tagging
Kalshi API Manager (50hrs)
Tools for Kalshi users to set up more advanced betting strategies
Project Predictor* (~~15~~hrs) (5hrs)
A project management tool with a built-in forecasting tool for getting better at predicting project timelines
I already have a self-built project management tool. This would just involve adding timelines to this:
*UPDATE: Thanks to Fatebook existing, I think it will be a lot easier to implement this and have downgraded the timeline for this to be completed to 5hrs!
This would also be built on top of the personal prediction tracker.
Additional projects added based on interest
What are this project's goals? How will you achieve them?
The goals are to build as many of these products as possible based on mine and the community's interest.
How will this funding be used?
I will work for one hour for every $50 of funding. I will work on the top voted project that has full funding (based on estimated time for completion) until it is completed (or until time spent exceeds 50% of the estimated time budget), then move to the second-most voted tool. I will continue until 15% of the funding is drawn down.
I will use the last 15% of funding to conduct maintenance and make improvements on all completed projects over the course of the next year.
The final estimate for finding the hours dedicated to each project came out to around $17,000.
Project priority will be determined by weighted ranked-choice voting, with each voter getting weight in proportion to the amount of matching funding brought in. (I am open to other, better ways of prioritization. For example, I think mini kickstarters for each tool would be a better way of doing things... )
I will work until all funded hours are accounted for. In the unlikely event that all projects are completed before this funding runs out, I will keep it as bonus for finishing early or pay myself to work on more similar projects or maintenance.
I do not expect to be able to spend more than 10 hours/week on this.
Who is on your team? What's your track record on similar projects?
It's just me, a chatbot builder/founder turned full stack developer (with help from Claude). Previous builds include Emile.chat, a chatbot that's been helping a small number of folks get to bed earlier since 2016.
Last year I created @manifoldmaven, a Twitter bot that tweets when a popular market moves by more than 35%.
I've also built a few arbitrage bots with the Manifold API, including a bot that bet Manifold markets toward PredictIt's odds during the 2022 Mid-terms to devastating effect when it turned out PredictIt was way off, and another bot that bet on Manifold markets of Brazilian soccer matches based on a sports odds API.
I have begun building full stack Firebase react web apps with help from Claude, and have been successful building tools for my own use with full login, database, and Firebase functions.
What are the most likely causes and outcomes if this project fails?
The more this project is funded, the more tools on this list can be built. This project will be a complete failure if no projects on the list are funded enough to give me the time to work on them to completion. In that event, I will still hold a project priority vote.
Another failure mode is underestimation of the length of time it will take to complete each project. If I've reached the time budget for a project and have not finished it, I will work 30% more time for free before moving on to the next project.
A third form of failure would be if some projects can be worked on within my time budget, but none of the ones that funders have voted to prioritize. This should be something funders understand when choosing to only partially fund a large project.
What other funding are you or your project getting?
None, but I may work on any of these projects for free in my own time before funding is secured or after funding is depleted. I may consider spinning each tool idea into a micro-kickstarter for extra funding.
David Glidden
3 months ago
Love the creativity and lots of swings at bat for tools that I'm likely to find useful.
Charlie Molthrop
3 months ago
@Dglid Thanks for the contribution! Feel free to rank your projects here in the comments if you have a preference. (I can also pass you my discord if you prefer to DM them).
wasabipesto
3 months ago
An interesting set of ideas! Some questions:
I believe Manifold is getting rid of the managram system as part of the prize point pivot, so unfortunately I don’t think managram subscriptions will be very useful.
For an automated arbitrage bot I think it would probably be better to submit this separately from the comments section to avoid revealing your linked positions. If someone had that information they could take advantage of it by e.g. forcing you to trade into their limit orders. **If this was in your web app you could overlay the two target markets and keep a list of active arbitrages, or even arbitrage against a different site’s market.
The secret market idea is interesting, but if I wanted to do that right now I’d just make one called “Secret Market” and communicate the true purpose of the market to whoever I wanted to trade with separately, through direct messages or a private google doc. What advantage would this extension have?
The manifoldmaven account is neat! I would love to see something like that with a little more signal:noise. The latest 3 posts are all about the RFK news, and the same Trump twitter market is on there three times. There’s also a few that are only caused by transient spikes which were quickly reverted. I think there’s also some low-hanging fruit to finding the root cause of a swing, usually a news article posted recently in the market.
Additionally, I wanted to see what the “Get these alerts in your inbox” link would do, but it just takes me to a place to pay $20/mo. If I were to pay this would I get all of these alerts in my email inbox? SMS inbox? Would I be able to customize the swing percentage or timeframe? More information would be good!
Unfortunately I don’t think I would pay any amount of money for this information, since learning about news instantly is nice but has basically no value to me over learning about the news whenever I get around to reading it.
Reading further down I think the Predictor's Digest is much closer to what I would want out of this, a full newsletter linked to prediction markets. I’m unsure of the AI aspect but would be interested to see.
I think loan infrastructure for Manifold would be very nice and has lots of potential, but with mana exchange going away I think loans and tips will be no more.
For the personal prediction tracker, how would this compare to something like Fatebook? I have used Fatebook in the past and it has a very nice interface for this sort of thing.
Charlie Molthrop
3 months ago
@wasabipesto
Thank you for all these thoughtful comments!
1. That's right, I'd forgotten that they'll be getting rid of Managrams. I wonder if they would still be open to a more controlled version of mana exchange subject to their approval. I'll have note that.
2. Ah you're right that it might be best to keep arbitrage private. I've seen folks list arb markets in the comment of another market, but that isn't exactly a rational trading move.
3. Hmm the core of the idea of secret markets would be some way to more easily decrypt the hidden details of a private market via a third party site or a Chrome extension that adds the privileged details to the Manifold market site. I agree that you could have a document that organizes things, but the idea here would be that no one would have to remember what "Question 2" or "Option A" means across a bunch of different markets.
4. Thanks for the compliment! I whipped it up all in one night so it's a bit rough around the edges. Regarding the alerts, I set it up to work both on SMS and via email, but truthfully, I haven't tried getting it working on SMS, though hypothetically it should be set up. I have it working on email for me, and I can honestly say it's actually how I get a decent percentage of my breaking news!
5. Yes, regarding Predictor's Digest, I had originally envisioned it as a New York Times front page but populated with prediction markets. I seem to have seen another attempt at this somewhere on Twitter. That said, the newsletter also makes a ton of sense (maybe even more so). I think in that case the AI can ideally jump in and provide a quick blurb at its best guess of why the probability shifted.
6. Wow! I had never heard of Fatebook until you mentioned it, and it's pretty much exactly what I envisioned! And they have an API! I might extend it to cover a few of the user experiences I envisioned for my version.
Thanks again for all these thoughtful notes. My fresh new Fatebook prediction only gives a 15% chance of this project getting any funding at all, but I've had all these ideas kicking around for a while. Worth posting if only for the feedback. Thanks again! I'll circle back to the project soon to modify a few of the tool descriptions.