Uncertain predictions

By Robert Hickson 19/02/2015

“Active open-minded thinking” and “massive ignoramus”. Those are views on what makes a good forecaster according to two  “super-forecasters” from Philip Tetlock’s and IARPA’s Good Judgment Project.  It also helps if you forecast as part of a team with other good forecasters.

But some things are more predictable than others.

Michael Burnam-Fink, in a paper in Futures, cites a observation made by Frederick Pohl in Arc 1.2  that, in science fiction stories at least, there has been better success in correctly predicting developments related to engineering where there are already well developed plans and research projects (space, weapons systems, etc), than in the biological and psychological fields, where hypotheses and theories rather than engineering solutions dominate.

No surprises there, its easier to make accurate predictions when there is more certainty. But those enthralled by new developments often forget that.

This is illustrated by MYOB’s recent Future of Business Report – New Zealand in 2040, where Simon Raik-Allen describes with complete certainty how we will all be living, or doing business. No ifs, buts, or maybes.

Including some uncertainty into your predictions is also good practice.

Rules for predicting geopolitical events have been proposed. Though I consider them as simply guidance about human behavior rather than predictive tools. These “rules” are not mutually exclusive, they can be contradictory, and they don’t describe clear outcomes.

But, as Nesta notes, making predictions is the easy part, doing something about them is much harder.

Being better prepared for the future doesn’t mean simply getting better at predicting, it also means being better prepared to make decisions in the face of uncertainties.