Today's realization

Feb 22, 2010 23:53

Game-able prediction markets can devolve into nearly the equivalent of completion bonds.

Organization A, with the power to dominate a market, desires a software (or any other) project completed. Organization A short-sells many shares of the "claim" that said project is completed by a certain date, thus bringing down the price of said claim.

Programmer (or other worker/set of workers) B decides she wants to make a lot of money by completing said project. Programmer B buys many shares of said claim, cheap, and gets to work completing project. Depending on how much Organization A spent on short-selling said claim, this could accommodate multiple competing groups.

If said project is completed on time (whether by Programmer B or a competitor), Programmer B makes lots of money by way of prediction paying off. If said project is not completed on time, Organization A makes money by way of prediction paying off.

Strengths? Weaknesses?

ETA: This scenario is specifically for projects with results that are important to Organization A, such as open source software projects or services/goods that are necessary for Organization A that aren't yet available on the market. It's not a substitute for contracts where Organization A expects to have ownership of the project's result.
Previous post Next post
Up