Most of Mann's tests, particularly the barrel shortening experiments that indeed turned bullets into "bubble gum", were done with weak alloy and black powder. You can create, or avoid this situation at launch by manipulating alloy, alloy temper, and powder burn rate selection. I know, obvious. But, the precise manipulation of alloy constituents and how bullet moulds and powder brands and types are chosen still has an effect on the finer points of group dispersion even when magnificent failures are not taking place.
Something else Mann discovered was no matter how fastidiously perfect he tried to make his static fit, his best groups were still made by pre-conditioning a dynamic alignment and fit of the bullet....in other words he had to give the bullet some wiggle room and then bump it to fit at the right moment to have the best results. I've found that for the most part, that holds very true for smokeless powder and harder alloys, particularly at very high velocity. The exact timing of "bumping" to fit has to happen just as the bullet finds center. The powder pressure must not exceed alloy strength until the bullet has centered in the throat, and the alloy strength and pressure curve must allow some self-centering without damage to the bullet's concentricity while it is presented to the throat.
Something else Mann worked out which I find MOST fascinating is an actual calculation for torsional stress upon the land engagement surfaces of the bullet. A few of us worked on that for several years to no avail and then I stumbled across it one day when reviewing The Bullet's Flight. I plugged in some of my own data from successful and unsuccessful loads and found land stress to be aroung 10-12,000 p.s.i. on a bullet in the 27-30,000 p.s.i. ultimate compressive strength range, at peak pressure of the load, which I estimated to be in the 42-45,000 p.s.i. range. What I took away from that is land deformation failures affecting group dispersion at high velocity and pressure might be occurring, but if so likely are not due to chamber pressure alone. Perhaps a combination of frictional heat and abrasion along with that pressure could cause issue, if even that is the cause of high-velocity failure points. I quickly discovered that the only thing necessary to destroy groups in my .308s at 23-2400 fps was to add some tin. Same thing in a slightly slower twist .30 XCB at several hundred f.p.s. more velocity. Another fellow currently has found that powder coating and Hi Tek epoxy coatings are allowing respectable if not magnificent groups of 6mm bevel base bullets at nearly 3,000 f.p.s. when otherwise the limit was round 2,000, and that in .30 caliber 2,300 fps was possible with reasonable groupings when without coatings the limit was in the teens. All with the same alloy. I've speculated for years that the phenomenon of cast bullet groups going to pot at high velocity, once one has the other basics of fit, alloy, and powder have been reasonably mastered, ultimately comes down to land engraves being damaged. Normal twists need some sort of jacket at high velocity, be it paper, copper, or even a couple thousandths of slippery, tough cross-linked Polyester paint. So Freebullet's speculations about softening after shooting, even like I mentioned even if only taking place in the barrel for a few milliseconds where it cannot be measured, might very much deserve some more consideration.