I’ve probably annoyed quite a few of my friends and acquaintances by insisting on inadequacy of statistics in forecasting. Statistics could be useful in presenting past in the (statistically) selected way - that’s how I argue - but that has nothing to do with the future. I’m sure I’m right but I see a question on many faces. My argument is missing something, believability obviously.

So I continue arguing with myself. Why statistics has nothing to do with the future? Statistics does not reveal how and why things happen - statistics only reveals what happened. Then, how and why are our construction based on outside assump- tions, on criteria outside statistics.

To the point, statistics belongs to the domain of data, not information. Somewhat structured data but still only data.