Prediction Markets

General Information

  • Prediction Markets are Wisdom of Crowds in action
  • Amazing results in lots of applications:
    • Wikipedia
    • Social Bookmarking
    • → Internet in general

What we learned so far

  • Basic theory: Condorcet Jury Theorem
  • Wikipedia: „where the average chance of a member of a voting group making a correct decision is greater than fifty percent the chance of the group as a whole making the correct decision will increase with the addition of more members to the group“
  • self selected traders volunteer to put their knowledge into a prediction
  • Certain value (real or play money) is connected to personal prediction
  • One classic approach: ask a group of experts and take the average
  • Manipulation? Have a look at the general stock exchange or even Wikipedia!

Types of Prediction Markets

  1. winner-takes-it-all
    • defining one specified event
    • pays off a specified amount if the event occurs
  2. index
    • pays off for every single percentage point
  3. spread
    • pays off even if the percentage is higher than expectation

Requirements

  • Design issue ⇒ how do buyers and sellers meet and exchange?
  • Contracts must be defined very clearly (Event occurred just because of a fusion with other contracts)
  • Play or real money?
    • by creating a „game feeling“ people are encouraged as well, although just play money is at stake
    • Play money: only good betters can be successful in the market (equality and fair competition!)
  • Why do people participate in MMORPG, browsergames in general?
  • Motivation of the traders is very important (e.g. real money showed: even thin markets (20-60 people) have very good results)
  • Only work well if there is enough useful intelligence
  • When public information is selective, inaccurate or misleading ⇒ PM won't work

Power of Prediction Markets / Conclusion

  • Provide incentives (dt.: Anreize) for truthful revelation (dt.: Offenbarung, Aufdeckung)
  • Provide incentives for research and information discovery
  • Provide an algorithm for aggregating diverse opinions

Links: