TECH TALK X
Bias, Information, Noise: The Central Forces of Forecasting

Decision-makers often inquire forecasters’ best guesses about topics such as whether an investment will yield a target return, an unusual tumor warrants surgery, or an adversary is violating an arms-control treaty. Unfortunately, forecasters typically work under less-than-optimal conditions: they have too little information, or their data may not be perfectly reliable. Many forces can interfere with the process of translating information or data into a prediction — among them are bias, information, and noise. Research has long focused on the effects of bias on the accuracy of predictions. More recently, our work finds that noise is a much bigger factor than expected. In this talk, we will describe the roles of bias, information, and noise in accurate forecasting, explore three simple interventions to improve these three factors and develop new insights on the secrets behind the so-called “super forecasters”, i.e., individuals who are consistently more accurate in making predictions than general public and experts.