I would tend to apply it over the pre-asic era to calibrate a scaling factor, and over the post-asic era to make an estimation. It's just a price-of-electricity estimation of a support level. I should add a Bitcoin price factor to test correlation to historical price:

(H*(T - aR) *  b + c)*B

a, b, c are the parameters to fit.  Since I was going from a one parameter model to a multiparameter model, I thought I should remove the ratio hack as no longer useful, thus eliminating the time window stealth parameter.  

Adding a hype factor: Use Google trends on "Monero" to obtain

(H*(T - aR + dG) *  b + c)*B