I’d like to incorporate prediction markets at my company, a technology consultancy. In the early days of Manifold, the Manifold team actually built me a white-labeled version of Manifold for private internal use, but for various reasons I wasn’t able to invest the time and effort to ensure its success. More old documentation here if you’re interested.
I’d like to give it another shot, but with the public Manifold instance. How might I best structure it to be private enough to not reveal proprietary secrets while also making it easy to use and increase the likelihood of successful adoption?
For example, my initial thinking includes:
Grouping all questions via a topic and/or Dashboard
Obscuring title and description details via a link to an internal Google doc that has the actual market titles and descriptions (but this seems clunky and likely to depress adoption)
Encouraging team members to anonymize their user names to encourage more open trading (in the earlier instance, as an example, I found that I was hesitant to bet my true probability because as the leader of a team, I needed to display confidence in my team to get a project done, otherwise they’d see my lack of confidence and it would be a bit of a self-fulfilling prophecy.)
I’m not sure how to incentivize accuracy while also not incentivizing someone to sabotage a project if they have a ton of NO shares that it will be successful, for example. (This might not be an actual issue, but I’d like to proactively solve for it.)
Any and all comments about any part of this endeavor are welcomed and will be considered for bounty award.
Happy to answer questions in the comments.
Related questions
I don't think anonymizing individual accounts will help, because they can eventually be roughly de-anonymized by matching what someone says with their bets.
Instead, I think you should create shared accounts, where the company or project teams share the login information to an account, and each member is able to trade and comment under the guise of that account. So if one member over-purchases the delivery estimate for a project, another could sell some shares.
You'd want to have a collective reward for whatever you determine as the metric(s) of success. That's harder to generalize, because you'd want to make sure it aligns with the short-term and long-term goals of your company.
You could still allow individual accounts, but have to make it clear that sabotage will be punished.
I also think that as a team leader, you should abstain from betting, instead using prediction markets as an information resource that alerts you to changes in the sentiment of your team on the delivery timeline, or other metrics.
Do you have specific types of questions in mind that you expect will be most valuable?
How many people at your firm? Would it be feasible to start small (eg just your direct team)?
Maybe start with generic questions about code named projects, eg “will project alpha finish by June” or “if we hire jon, will project beta be successful”
I think that the most motivating way to use manifold for teams is to have default bets on the negative outcome. This incentives teams to bet in favour of positive outcomes (bigger reward) and motivate them to work hard to achieve the positive outcome.
Example:
Will team A have a functional Beta test of project C before start of Q3.
Default company bet of X M on No.
Possible payouts for further No bets are lower, while possible payouts for yes are higher.
Another added benefit is that the company can incentivise work of specific deadlines by betting more against a deadline, making a positive outcome more profitable for teams. If you use it in this way, you can also link it to a bonus system.
