Creating a “Social Process” to Answer Technology and Futurism Questions
Dr. Robin Hanson and Dr. Charles Twardy of George Mason University lead the SciCast project, which has become the largest crowdsourced science and technology forecasting site on the planet.
The site has over 10,000 science and technology experts and enthusiasts who have made more than 80,000 predictions on hundreds of topics including technology, mathematics, the sciences, consumer and geopolitical topics.
According to Robin Hanson, SciCast is intended to be an “idea futures market” where predictions about the future are made and voted upon. Hansen’s background is in artificial intelligence (AI) and Bayesian statistics. Charles Twady has a background in forecasting and believes that crowd-based predictions from groups of people outperform individuals and simple averages.
Fancy Crowds vs. Simple Crowds
Hanson has been analyzing data on the success of crowds at predicting future events and has found “fancy crowds” to outperform simple crowds by 80% in making hundreds of real-life forecasts. Fancy crowds are marketplaces, such as stock exchanges, while simple crowds provide a simple average baseline.
Because SciCast users are making real life predictions in real time Hanson, Twardy and their team are able to identify predictions that have come true and the site keeps statistics on all predictions.
According to Charles Wardy, a veteran in the forecasting field, the best forecasters on SciCast and sites which predated it, make thousands of forecast per month. Wardy also identified psychological correlates of good forecasters which include: “general intelligence, need for cognition, subjective numeracy, cognitive reflectance, domain knowledge, getting probability training, and working in groups.”
There is also the possibility that the SicCast market could open up to investors who would invest in forecasters thereby, very likely, increasing the accuracy of the market.
The following question and stats are from a real live question on SciCast.org.
Topics Being Forecast
SciCast has the following categories:
Business of Science
Global Change
Information Systems
Computational Sciences
Energy
Engineered Technologies
Biology & Medicine
Social Sciences
Transportation
Space Sciences
IEEE Spectrum 50th Anniversary
Chemistry
Mathematics
Physics
Agriculture
We will be looking into SciCast in more detail in coming articles on IndustryTap.
Related articles on IndustryTap:
- New “Killer App” Reduces Crime by Forecasting Risk Times, Locations
- What’s Next for Drones? Hurricane Forecasting to Save Lives.
- China Using Supercomputer to Design Smart Cities
References and related content:
