A group calling itself the Intelligence Advanced Research Projects Agency (IARPA) conducted a test project last year of inviting folks to make forecasts on major trends around the world. Their forecasts beat the forecast of a control group by 60%.
Now, they are skimming the best forecasters from the project to see if they can get better.
I doubt they can and said I would put money on it. Another commenter, “DK”, has taken me up on my offer. We have traded emails and established this bet:
I bet $20 that using the same weighting algorithm that the super-forecasters do not outperform the control group by 60% or more. If they outperform by less than 60% or do not outperform, I win. If they outperform by 60% or more, DK wins. If we find out the algorithm changed, either of us can use that to call off the bet.
DK offered $200 originally. I turned it down for these reasons:
I do realize that a) I could be wrong, b) the super forecasters could get lucky, c) Tetlock could change the weighting-algorithm without telling us to rig his results and/or d) the control group contains some crazy guesses (and may even be selected, unknowingly, to contain crazy guesses).
I offered $20 because I have a corollary to Caplan’s oath: Never bet more than you are willing to lose.
I’m not sure if my ‘weighting algorithm’ condition means anything or matters.
My memory is bad, so I’m counting on DK to update me when the second year of results are released. And, I’m rather confident that he or she will. I’d be willing to bet on that.
- An email from Philip Tetlock (marginalrevolution.com)