I'm not certain that a "hard and fast" rule can be asserted to this question. Oh, BAD pun.......sorry.
Too many people have experiences that almost diametrically conflict with one another's in this sphere to make categorical statements. I am pretty sure that if ALL the variables could be isolated and controlled in a ballistics lab, we might be able to make better estimates--but whether a scientifically-sound and proven assertion is possible.......not so sure.
An example to illustrate........9mm Luger and 40 S&W loads using cast boolits of 92/6/2 alloy. Doing the "Brinnell formula" on this 13 BHn alloy (13 x 1422) shows that we can expect the alloy to slump and begin deformation at 18,486 PSI, and assume that formless and likely inaccurate shape referred to in a prior post. The loading data for the calibers show pressures running from 25K to 33K PSI in some of the loads used in these calibers.
But that doesn't happen, apparently. I have recovered dozens of boolits fired from my pistols into soft sand that bear impact indications to noses, but the drive bands--lube grooves--and most of the sidewalls appear to be largely unaffected. This same result occurs using faster powders (WW-231)--medium powders (Unique or Herco) or slower fuels like Blue Dot and AA-7.
(Pure speculation from this point on, and by a social sciences major I might add). I believe there is quite a bit of "cheat" in the internal ballistic equation when it comes to cast boolits and the influence the pressure impulse has on the boolit's base and overall form. I agree wholeheartedly that cast boolits are "plastic", are fragile, and gain considerable strength from their surroundings--those surroundings being the grooves and bore that can support them during their accelleration downbore. The closer the reloader can get that boolit to the dimensions of the barrel environment prior to firing--and the better that close dimensional relationship can be maintained (no boolit diameteric reductions by overtight case mouths and enthusiastic taper-crimping), the less variability in the equation results. There should be a reduction in or absence of leading--and an uptick in downrange accuracy from said efforts.
My point, summarized--size boolits to fit closely, DON'T SPOIL THEM WITH THE TOOLING, support them downbore, and good results should be produced.
Mi dos centavos. Free advice, and possible worth it, too.