Why rating tank armor and penetration into RHAe is wrong

People on the internet and even some authors of military textbooks and magazines often pretend that armor and penetration can be accurately measured in milimetres of equivalent steel thickness. They use a so-called rolled homogenous armor equivalency (RHAe) to rate the performance of special armor in comparison to rolled steel (RHA o RHS). 
The logic is simple: When a projectile penetrates X armount of steel armor, but only X-Y after penetrating a certain type of special armor, then Y is the protection provided by this armor. To a certain degree RHAe has been used in science, but to a much more limited extend than people on the internet and many authors of books focused on tanks pretend.

There are many different examples which however show that this is not the case. Relatively simple spaced armor used to provide considerable protection against earlier types of AP(FSDS) ammunition, but has been rendered useless by more modern APFSDS designs. Kontakt-5 used to be highly effective against APFSDS, but modern APFSDS design and improved metalugry make it obsolete. ERA and NERA are highly effective against shaped charge warheads, but warheads with precursor charges and tandem charges perform very well against most types of (N)ERA.
But also the exact shape and design of the used ammunition will affect the efficiency of modern special armor. The Soviets experimented with DU as material for HEAT ammunition and even adopted one type of DU-HEAT, despite this ammunition not performing better against RHS; against special armor however the DU-HEAT was considerable better.

This also affects the supposedly superior penetration of DU compared to tungsten penetrators. According to US studies, penetrators made from tungsten-iron-nickel (W-Fe-Ni) alloys had between 8 and 10%¹ less penetration depth into rolled homogenous steel targets than DU penetrators.

An example about how big the difference between penetration into special armor and RHS is given in a scientific study of different penetrator designs.² The penetration of conventional monoblock penetrators and jacketed penetrators into a block of steel and a block of steel protected by spaced armor is compared.
Both penetrators have an aspect ratio of 25 to 1 and are fired with the same propellant. The jacketed penetrator is a tad faster, due to the steel jacket fitted to the tungsten penetrator weighing less than the tungsten of the unitary penetrator.
Against normal RHS, the jacketed penetrator had a 12% lower penetration compared to the unitary one - against the target protected by spaced armor however, it had a 17% higher penetration depth! So while any "RHAe values" would put the penetration of an unitary penetrator above that of a jacketed one, the reality shows that against more complex armor systems the jacketed penetrator might be a lot better.

---

1 - The values for the lower penetration of tungsten vs DU into semi-infinite RHS targets comes from "Tungsten Alloy Properties Relevant to Kinetic Energy Penetrator Performance" from Downing et al; but improved tungsten alloys have mimicked DU's self-sharpening ability
2 - J. Stubberfield, N. J. Lynch & I. Wallis, "Comparison of unitary and jacketed rod penetration into semi-infinte and oblique plate targets at system equivalent velocities", International Symposium on Ballistics