Superforecasting: The Art and Science of Prediction

by Philip E. Tetlock

“unlikely,” but as we have seen that introduces dangerous ambiguity, which is why scientists prefer numbers. And those numbers should be as finely subdivided as forecasters can manage.