Dance of the P-values. If you thought that using p-values as the go-to shorthand for measuring the quality of a result was a good idea then think again.

In my misspent youth I briefly worked for a company doing simulation software for submarine construction. There I learned of the "admiralty coefficient"; a way to summarize the power ("how fast, how strong") of a given vessel design in one single value. And of course wildly different designs could get the same coefficient. A single-dimensional value just isn't able to capture even the most critical design points. P-values really are no better for classifying experimental results.
Shared publiclyView activity