Recently I took part in a geopolitical forecasting study that was designed to help the US intelligence community improve upon their ability to predict the future. The study—which was organized as a tournament between thousands of forecasters—was funded by the office of the Director of National Intelligence through a government research agency known as IARPA. It’s called The Good Judgment Project and has been reported on by The New York Times, The Financial Times, the BBC, NPR, The Economist, and The Washington Post.
In the tournament, my collective forecasts on over a hundred geopolitical issues proved to be more accurate than all but one of nearly 3000 other study participants. Which is partly why I'm posting this! In my spy novels, I try to give readers an insider's perspective on a whole host of cutting-edge geopolitical issues. So if my performance in this study leads potential readers to conclude that I'm not just one more half-informed pundit talking out his you-know-what, well, then I'm a happy man.
I'm also posting information about the study because I genuinely think the methods employed and conclusions reached point the way to a future where the forecasting aspect of intelligence analysis can be radically improved and depoliticized. I am in full support of any and all efforts to get us to that future and I encourage others to be supportive as well. If you're interested in learning more about the science behind effective forecasting, I'd recommend reading Superforecasting: The Art and Science of Prediction and/or viewing this series of online lectures: Edge Master Class 2015: Philip Tetlock: A Short Course in Superforecasting.
Finally, I think the questions that were asked lend beyond-the-headlines insight into what the intelligence community was actually thinking about in 2014-15. Having had the pleasure of getting to know many of my readers, I'm guessing many of you will be as interested by the questions as I was. Here are links to them, organized by subject matter:
Note: Although my final forecasts are visible, they're not necessarily reflective of how I forecast over the course of a question.