I think Clark's data is pretty useful in determining what the real danger area is when reloading. We all hear all these horror stories about guns blowing up, and then the person that blew up the gun always claims that it must have been due to a small error. Then Clark goes and does something insane, like putting .357 magnum overcharge loads through a .38, and proves that a small error usually won't blow you up, despite the horror stories.
Looking at all the crazy overloads Clark and others have done, it looks like in many cases, not even a double charge is sufficient to blow a gun up. With lots of powders and calibers, it's physically impossible to jam enough powder into the case to cause a catastrophic failure.
The main reasons to stick to published load data is not to keep from blowing yourself up, which usually requires an insane overload, but to keep your gun from battering itself to death, keep the bore from getting prematurely eroded, and keep your hands from getting bruised. There are some exceptions, like Hodgdon Universal and Winchester Super Field, but you can usually find info about what powders genuinely are unpredictable.
If anything, Clark's data is useful because it establishes how safe or unsafe a powder is in a certain cartridge. You can look at Clark's antics and go "oh, Bullseye looks like a very safe powder to reload .25 ACP with, because it's physically impossible to blow up a quality, all-steel gun with it, even if I somehow accidentally put in such an insane amount of powder that it requires double compression to get it all in there, and resizing the case again after the bullet is seated to make it chamber in my gun" or "gee, the Colt 1908 Pocket Hammerless .380 is a pretty good gun to reload for. Even if I accidentally use a .357 magnum overload in it, it won't blow into pieces."
Looking at all the crazy overloads Clark and others have done, it looks like in many cases, not even a double charge is sufficient to blow a gun up. With lots of powders and calibers, it's physically impossible to jam enough powder into the case to cause a catastrophic failure.
The main reasons to stick to published load data is not to keep from blowing yourself up, which usually requires an insane overload, but to keep your gun from battering itself to death, keep the bore from getting prematurely eroded, and keep your hands from getting bruised. There are some exceptions, like Hodgdon Universal and Winchester Super Field, but you can usually find info about what powders genuinely are unpredictable.
If anything, Clark's data is useful because it establishes how safe or unsafe a powder is in a certain cartridge. You can look at Clark's antics and go "oh, Bullseye looks like a very safe powder to reload .25 ACP with, because it's physically impossible to blow up a quality, all-steel gun with it, even if I somehow accidentally put in such an insane amount of powder that it requires double compression to get it all in there, and resizing the case again after the bullet is seated to make it chamber in my gun" or "gee, the Colt 1908 Pocket Hammerless .380 is a pretty good gun to reload for. Even if I accidentally use a .357 magnum overload in it, it won't blow into pieces."