Something weird happens when you start drawing time-series graphs a lot.

You start to assume that all continuous experiences have a predictive value.

  • If the business has grown at a yearly 12% rate for the last 5 years, it’s probably going to do the same this year.

  • If my last 3 relationships ended with a fight, the next one will too.

  • The stock markets average a 6% increase in value every year, so next year is probably quite safe.

  • I start a gym membership every January and quit it every May. So it’s a given that I’ll do that next year.

Of course, when examining these defaults, it’s surprisingly easy to see how broken they are. Patterns break. Sometimes it’s as simple as deciding to break them, and other times it’s driven by external factors. But the past is never actually very predictive of the future in the time-series consistent way.