S.B. My thoughts are that it is very hard to replicate a freak occurrence which only happens one time in a million. There may be odd factors which contribute, residue inside the case, bullet lube, powder clinging to the case walls in a very specific pattern, powder dust, powder under the primer, weak primmer, extra strong primmer, temperature, etc. things which have to come together in a rare and specific pattern. The very process of testing in a precisely controlled environment may eliminate a random factor which contributes to the detonation. If it could be replicated in the lab then guns would blow up every time with certain loads and we all would know not to use those loads and therefore, no need of testing. I worked a few years in a government testing lab and I can assure you lab results are never exactly repeatable and I doubt that any of them tested a million loads. Your analogy of the Warren Commission is indeed very applicable, start with a conclusion and work back from there to "prove" it.
SHOOTALL, a soft case head would be a good guess, except for the manner in which that happened. I first loaded five rounds each with 47, 48, 49 and 50 grains of H-414. All indications were of very mild pressure, just as the Lyman manual had predicted. The best group was with 48 grains so I returned to the bench and loaded the same 20 cases all with 48 grains of H-414. So the case which failed had previously been fired with a similar load which showed no signs of abnormality.
Back in the fifties one could buy surplus 4831 powder very cheap and people often used it in applications for which it was not very suitable, such as reduced loads in the 30-06 and rifles were wrecked with reduced loads even though a case full and compressed produced mild pressures and the reduced load had been fired hundreds of times with no problem. That also could not be replicated in the lab so it was put down to reloader error, must have used the wrong powder. But it happened to some people who didn't even own any other powder.
