aka Manifold Prop Bets aka WvM-as-a-service

Goal: Create an always-on, fast, fun betting experience to help drive user engagement

Proposal

• Every n minutes, select one of the top 50 Trending markets and create a new market on whether the probability of that market will be above or below P% at the end of n minutes, where P is the current market probability of that market. This is the “quick market”
  • Eg if the selected market is “Will a Neuralink device be implanted in a human brain by the end of 2023” 19%
  • The quick market will be “Will the probability be above 19% at 2:15 pm?”
    • [i] randomized close time is actually really easy and makes it slightly less non-predictive. Not sure if that’s good or bad
• The quick market should be displayed prominently on the homepage
  • With a custom UI that includes the underlying market
  • and maybe a countdown timer
• After n minutes, the market is resolved automatically and a new one is created

Other thoughts

• If done right, this could be an engagement gold mine: it will ensure that there’s always something for the casual user to do at any given moment in time
• This proposal is basically the House centralizing and sucking the oxygen out of all the other nonpredictive/gambling markets, which will have many good effects:
  • All gambling content going in one place, making it easier for uninterested users to ignore
  • No trader bonuses going to creators of gambling markets
  • We can unlist all of these markets to not dilute league / leaderboard calculations or clog up Trending markets
  • We can feel more empowered to unlist all other users’ nonpredictive markets knowing that we won’t take an engagement hit
• The time length n should be chosen such that is as small as possible while ensuring decent trader participation
  • Could be dynamically chosen so that’s 5 mins during higher traffic periods of the day versus 4 hours or something at night…
• In addition to markets on the whether the price is higher/lower we could have a whole list of possible questions involving market dynamics
  • will the price move by x%?
  • Will there be more than m traders?
  • etc. Basically all the non-predictive questions our users have come up with
• We might consider
  • a simpler mechanism like vanilla parimutuel or DPM
    • Since the probability for these markets naturally hovers around 50%, parimutuel is a reasonable choice
    • The self-boostrapping liquidity dynamics of PM, DPM are a good fit in this context and could mean the house doesn’t need to subsidize the market much at all
    • We also don’t care about price efficiency at all here…
  • and an even simpler / more normie friendly betting interface
  • …Or not. Just using a normal market is easy and consistent
• Even if we decide we don’t like this proposal, we should try to solve the meta problem of creating more content with a faster feedback loop to drive usage and retention