Inaccurate Example of Metric #1
Description
Th article says:
This particular metric is what Fisher chose, but it's certainly not the only possible one. A different metric could optimize for different results (perhaps you care a little more about minimizing scatter so you multiply the denominator by a large number to give it more weight).
However, if you just multiply the denominator by a constant, it changes all the scores the same way, so it won't change the relative "best". See:
https://twitter.com/D_M_Gregory/status/1069965514120851457
Maybe come up with a better example?