Some years ago the Alaska Dept. of Fish and Game began a study, since abandoned, of penetration using a variety of powder charges and projectiles. The data still exists here and there but is not readily available. There are plans to resume the study at some time in the future. Meanwhile, the preliminary data does hint as some interesting results.
The test media consisted of a plywood frame containing the test media as follows:
innertube stretched over the opening, to simulate skin, followed by 1 inch of water-soaked telephone books, 1/4 inch plywood to simulate breaking through a rib or scapula, with the remainder of the media consisting of more water soaked telephone books.
Generally, heavy conical projectiles penetrated more deeply than did round-balls but r.b. certainly penetrated well enough to have resulted in fatal wounds to heavy-bodied big game animals. Pistol bullets fired in sabots performed abysmally. So badly that they are not recommended for use on heavy-bodied Alaska big game animals.
The best penetration measured in this test was from a hollow-base, solid-point "buffalo bullet" fired from a modern in-line rifle loaded with 3 Pyrodex pellets.
Surprisingly, black powder cartridge rifles outperformed muzzleloaders of the same caliber and similar load. For example, a .45-70 loaded with 70 grains of Goex Cartridge powder firing a 500 grain conical penetrated entirely through the test media and could not be recovered for examination. The same was true of a .50-90 and .50-110 Sharps.
Round balls performed best at moderate velocities. When driven at velocities much higher than 1200 fps penetration actually decreased, rather than increased. For test guns for which a load had been worked up for optimal accuracy, the most accurate load also seemed to the load that in that gun produced the best penetration.
Ignition system had no measureable impact on penetration. It didn't matter if the gun was a flintlock, side-hammer percussion or modern in-line.
As noted by others, rate of twist has a considerable bearing on rifle accuracy. Round balls like a slow rate of twist while conicals want a fast rate. 1:48 is somewhere in the middle. During the 1970s and early 80s most factory-produced guns settled for that middle ground hoping the compromise would result in a rifle that would shoot either conicals or round-balls well. Unfortunately, most of the guns don't shoot either one worth a tinker's dam.
I generally recommend my customers decide what their primary purpose is for a new rifle, and buy a gun with a rate of twist appropriate to that purpose. If the costumer is a hunter with no interest in living-history then a rifle that shoots conicals well (a fast rate of twist)and if the customer is a living-history enthusiast I recommend a slower twist, round-ball gun.
BTW, there is no reason for anyone to hesitate to hunt large bodied big game with a round ball. They perform adequately at all reasonable muzzleloader ranges.
Also something to consider is that big bullets will not kill an animal any faster or more humanely than small ones. Animals with a slow metabolic rate take a while to die, regardless of what they're hit with. One old Sourdough noted that it takes a moose 3 1/2 minutes to die regardless of whether he's shot with a .22 or a bazooka. Kodiak brown bears take a bit longer.
Swanny