In this week of predictions, with the Nobel prizes just announced (one more quasi-Nobel prize to go – Economics tomorrow), I thought it’d be interesting to look at the ways predictions are made and what the reigning methods are.
For the Nobel prize itself, we saw some of the predictions come true, while others turned out to be off the mark of course, as always. It’s a tricky business making predictions, especially about the future, as Niels Bohr is said to have quipped. This goes especially for an event such as the Nobel prizes, where there are no longer-term trends to help guide our thinking. This year, UK bookies thought Murakami would get the literature prize (I still don’t understand why he’s even considered – he’s imaginative in terms of stories, but really weak on characterization), and The New Republic just published a list of how often they are wrong.
For the science prizes, it might be slightly easier, since there tends to be more consensus on what is Nobel-worthy, even thought the exact year of the prize is hard to predict. This year, Higgs did get his Nobel, but many of the others were surprises. Thomson Reuters’ ScienceWatch makes a set of predictions for the Nobel Prize winners based on the number of citations scientists and their articles get. This year, their nominees did include Higgs of course, but not the other winners.
Looking at the wider world, an interesting question is if big data is changing the way forecasts and predictions are usually made. I’m currently reading Expert Political Judgement, by Philip Tetlock and The Signal and the Noise by Nate Silver. Tetlock’s main argument is that in order to be a good forecaster, you need to be a fox rather than a hedgehog, in terminology borrowed from Isiah Berlin. The nimble minds who accumulate information from several different sources and who are not afraid to update their forecasts that perform better than the hedgehogs with strong convictions and the tendency to interpret the world according to your worldview. The book was written a few years ago, and the world has arguably become even less hedgehog-friendly over this time. The kremlinologists that Tetlock studied would struggle to keep up with the pace of the changes in e.g. the Middle East. The fact that we have more and more data would, not change our ability to make better forecasts, like Silver says, it just makes us more able to create models that support our hedgehog views. However, I think Silver would argue that, these days, you need to be a fox who is even more knowledgeable about the subject at hand, along the lines of what Matt Yglesias said in his review of the book in Slate.
However, if you’re a hedgehog, don’t despair! There is always Long Bets, where you can place bets decades into the future, significantly decreasing the likelihood that you’d be off a few years on the timing.