The Prediction Horizon is the notion that prediction quality is much lower than intuitively assumed, both in quantitaive and qualitative realms. Superforecasting by Philip Tetlock discusses the fact that even the top performing forecasters, who outerformed CIA analysts, couldn't predict much past about 400 days. Robin Hanson has discussed the atrocious track record of experts making qualitative predictions in their various fields. Happiness research highlights how bad we are, on average, at predicting even our own future preferences. This lends weight to the concept of option value. If we don't know what will happen or how we will respond we should like to keep our options as broad as possible.