Sunday, March 6, 2011

On Prediction

Just to follow up on my apparently less than clear point regarding models and predictions I made last week, I want to point out that this should not prevent trying to plan for future events. It just means that past events should not be analyzed for cause and effect so as to be predictive. It’s nice to pretend through hindsight and applying effects to certain causes that we actually knew that what would happen did but were nevertheless caught flat footed. Doesn’t say much for expert analysis, does it? But, in the case of the Arab uprisings, it is critical to have some plan that takes into account not only scenarios that we can think of, but of the effects of the improbable.

In other words, if we stick to the old way of analysis, then I guarantee that the results will be completely different and the experts will be asked by the pundits – but why did this happen and shouldn’t we have known? The answer to that question from the again proven wrong and highly paid experts will be that it was unanticipated because they had never encountered it before. When, of course, they had.

No comments: