Two days ago myself and @Osnat KM along with many others released the production-grade version of a completely open-source, MIT-licensed prediction market platform, SocialPredict, which can be found on Github here.

There are a couple of core tenants to this project I wanted to hightlight:

• Easy to use and deploy, meaning small organizations, classrooms with very little or no budget can use it.
• Isolated, meaning you can simply run a tournament or prediction market on a temporary basis, add users ad-hoc as an administrator, and then throw it away.

Now that being said, there are extreme limitations to this software, which is typically expected in open source. Comments are not enabled (which some may see as a positive), there are some problems with how positions are calculated leaving the last bettor in a lurch, and many other issues listed here.

But we have a direction and momentum, having been used by Kenyon College and St. Lawrence University to teach a class in 2024, and having gotten letters of endorsement from UVM and Stanford.

Many questions and have been asked about prediction markets on the Effective Altruism forum in prior years. Our hope is that making them more accessible and affordable to run for essentially anyone, can help bring more clarity.

You can read more about our journey as we continue to build this out here.