You do not have Javascript enabled. Some elements of this website may not work correctly.

When we make predictions, we often examine the specific situation in front of us. This is called “inside view” thinking. For example, if we want to predict whether we’ll get to work on time tomorrow, we might think about the schedule we plan to follow on that particular morning.

However, another prediction method, "outside view” thinking, is sometimes more reliable. We can improve our prediction by thinking about all the other days we’ve gone to work, and how often we were late. We call this the “base rate” of lateness; the chance of being late to work on an average day.

If we are late to work, it could be due to many different factors: oversleeping, missing the bus, forgetting our keys, etc. Calculating the odds for each individual factor would be difficult, but using the outside view to find our base rate of lateness lets us avoid this, since that rate combines the odds for all of the factors. If we are late for any reason half the time, we can start with a prior credence that we have about a 50% chance of being late tomorrow, for any reason.

After we’ve used the outside view to find a prior credence, we can update our prediction using the inside view, by considering specific details about tomorrow. If we have an important meeting in the morning, we might be more likely to arrive on time; if a blizzard looms on the horizon, we might be less likely.

The outside and inside view both deserve our attention, but before we update away from the outside view, we should have a specific reason for doing so. If tomorrow will be typical in every way, we should predict a 50% chance of timeliness. The history of past events is often more reliable than our own judgment.

Further reading

Kahneman, Daniel. November 2011. "Beware the 'Inside View'".

Kawamura, Kohei, and Vasileios Vlaseros. 31 July 2014. "Expert Information and Majority Decisions".

Tetlock, Philip, and Dan Gardner. 29 September 2015. "Superforecasting: The Art and Science of Prediction".