“Trends” analyses are everywhere, but verifiable forecasts are few and far between. Almost no one keeps a scorecard. But we do. For the past year, the science & tech future has done moderately well at matching our expectations. We give it a “B”.
After an overview and live demo of the SciCast prediction market, I will demonstrate improvements made in the last 6 month and discuss new results regarding forecaster accuracy & calibration, as well as some technological advances under the hood. I will end by describing our current accuracy contest, and present useful forecasting tips for anyone wanting to improve their chances at one of the top-prize $2250 Amazon gift cards.
This talk is aimed at introducing a broader audience to the SciCast system, and should appeal to potential forecasters, question writers, and forecasting researchers. SciCast is an online prediction market for science & technology. SciCasters invest points to make forecasts. When the outcomes are known, forecasters that raised the chance of the actual outcome will gain points. Over time, better forecasters dominate, improving system accuracy. Unlike other markets, SciCast allows forecasters to link questions, so (for example) photovoltaic $/W can depend on the progress of multi-junction arrays, and Paypal’s chances of using Bitcoin depend on whether Google Wallet does.
Charles R. Twardy, Ph.D. leads the SciCast forecasting project at George Mason University. Dr. Twardy received a Dual Ph.D. in History & Philosophy of Science and Cognitive Science from Indiana University; he works on evidence and inference with a special interest in causal models & Bayesian networks. Dr. Twardy also works on Bayesian search theory, especially the analysis and prediction of lost person behavior.
He has worked on argument mapping, information-theoretic trajectory clustering an sensor-selection, and Bayesian models for counter-IED work, source credibility, image recognition, environmental decision-making, and epidemiology.
1:00 pm - 2:00 pm
C4I Center ENGR 4705