Back to the original question of the reason why bottleneck cases are problematic.  My speculation was that the bottleneck cause a condition similar to a shaped charge used in anti-tank rounds.  Look up ‘shaped charge’ on the internet to see what I’m talking about.  That idea came from my EOD training back in the early 1970’s.  The bottleneck case concentrates the gases and produces higher heat and pressures primarily in the neck and throat areas and possibly damage to the base of the bullet.  That would bake or damage areas of the neck, throat, and bullet base during firing.  Repeated firings would tend to build up residue/damage in those areas.  I have absolutely no proof of that theory, just speculation.  Supposedly the military during the development of a more suitable cartridge than the 50-70 in the early 1870’s stopped using bottlenecked cartridges as they produced more pressures and disproportionally lower increases in velocity.  They ended up with the straight case 45-70 as a result.

I think I finally found the reference for the Government conclusion that bottleneck cases produce more pressure with the same loadings than straight cases.

METALLIC CARTRIDGES (REGULATION AND EXPERIMENTAL) AS MANUFACTURED AND TESTED AT THE FRANKFORD ARSENAL, PHILADELPHIA PA.

Prepared under the direction of the Chief of Ordnance by Major T. J. Treadwell, Ordnance Department, Commanding Frankford Arsenal, Washington GPO, 1875.

Plate LXV shows that ten shots of a .45 caliber, 400 gr bullet with 70 gr musket powder in a bottleneck case, gave a mean pressure of 18,500 psi.  The same bullet and powder charge in a straight case gave a 16,300 psi mean.

A .42 caliber with 370 gr bullet and 65 gr musket powder, gave a 17,150 psi mean for 10 shots.  A straight case with the same loading gave a mean of 16,250 psi.

The pressure gauge used was the Rodman, which was set up with a built-up Springfield Trapdoor receiver, standard barrel which had a hole in the chamber and a piston that pushed a knife-edge against a copper disc.

The knife edge was transferred to a "weighing machine" and weight was increased until the same depth of cut was made in another copper disc, and the calculations made for psi.

Plate LXIV comments that 25 shots of these calibers gave a mean of 2300 psi more pressure in a shell crimped over the bullet compared to an uncrimped loading.  Also a 30 ft/sec increase in velocity.

I cannot help making the comment that maybe this testing should be repeated with modern piezoelectric detectors.  It seems a lot of judgment calls need to be made with the Rodman gauge, but maybe people were that good back then.  But the pressure difference between bottleneck and straight cases being approximately that of crimped vs uncrimped would seem to leave open the question of how significant this "pressure increase" really is, given the absence of distribution around the mean; not to mention the possible variations built into the testing apparatus.