There has been some criticism of the Polarized Fractal Efficiency formula. It can be shown that its inconsistent, i.e. that it doesn't take into account scale, which in this case biases the calculation on lower time frames. This can give false signals for example when you shift between time frames, so scale is important.

I have made a dimensionless PFE that passes the "C-test", which was advocated by William Eckhardt. I.e. it doesn't have scale bias like the original PFE.