I had a partial case head separation using a 1.8oz buffer

[DeNinny]

here it is in four easy images

It's clear in this image that the head is not supported by chamber wall at the failure point. There is a gap under the bevel between the end of the lugs and the rear of the chamber. Even if this pic is not to scale, there's going to be a gap with a small amount of unsupported wall. And it will vary between barrels.

Based on this, there doesn't even need to be an out of battery scenario. The case could have failed in battery. And again, the case that will more likely fail will be the one with thinner walls or base (all other things equal).