Especially if you can pick just the forecasts that are not already wrong, and forget about the rest.

According to this article in nature, there is rather a tendency to use worst case scenarios that are no longer probable.

https://www.nature.com/articles/d41586-020-00177-3