Search This Blog

Tuesday, 3 December 2013

Recent Evidence on 'Forecasting'

Forecasting is important in every aspect of life and in business in particular. Grossly simplifying, the last two decades' academic research came to the general conclusion that crowds do better than individuals, the so-called "wisdom of crowds (WOC)" hypothesis. One of the resulting 'innovations' is the idea of prediction markets where people trade securities whose payoffs are tied to specific outcomes/events. Market prices provide 'superior' forecasts for the likelihood of the events in question. With social media broadly available to large and distributed populations, prediction markets are thriving...
    Recent research conducted by IARPA seems to indicate that the WOC insight might be challenged - at least to some degree. Clearly, prediction markets do much better than the average expert participating in them. But it also seems to be the case that the top 2% of forecasters can beat the market by a relatively large margin. First, it seems that the elite forecasters do seem to do better systematically over time, so their performance is not just luck. Moreover, if you team them up, then together, they can beat prediction markets by 20-35%.
    Prediction markets are cool forecasting tools but their weakness is that they give away the forecast, so in business where one wants to generate superior insights it is hard to discover proprietary information this way. But what if firms ran forecasting tournaments to discover the few experts that can provide sustainable advantage for them? Of course, the question then becomes: at what price will these experts share their views with the firm!

1 comment: