I don’t know about this... Forecasts in general can definitely be wrong. Assigning a 99% chance to Clinton winning, for example, was a modeling error. If you don’t have the outcome in your prediction interval, from a modeling standpoint you very likely did something wrong.https://twitter.com/zeynep/status/1323649467015376896 …
-
-
There are some recommendations for understanding/quantifying model error in this paper https://pdfs.semanticscholar.org/18f8/ba2806833f3a1460380f56d31325ea592c53.pdf …
-
Most of these strategies depend on many data points. How do we validate a model against a single data point? Even if outside the CI’s, that’s expected for some rare cases. Easier to do for the house races, maybe

- Show replies
New conversation -
-
-
Consider two possibilities: 1. An event that you gave a 1% chance to happens. 2. You modeled the situation incorrectly (an event that itself has some unknown probability, which we might strongly suspect is >1).
-
What this means is if a sufficiently rare event happens, it's more likely that your model is wrong than that you were unlucky.
End of conversation
New conversation -
Loading seems to be taking a while.
Twitter may be over capacity or experiencing a momentary hiccup. Try again or visit Twitter Status for more information.