I"m not a metalurgist, my eng background is nuclear but I recall that most all the metals we deal with have a molecular structure which is damaged by cumulative heat cycles. This is most true when those metals are operating just below their plastic or melting point. Each time a metal is subject to exceptional heat, the molecular structure is deformed, or weakened just a small amount. We're talking about EGTs of 1500-1600F flowing out an exhaust port made of siliconized Al. The continuous temp of that exh passage is working very close to the melting point of the alloy. Keep doing that, and the bits that make up the valve train/piston are going to start deforming.
And as this applies to catastrophic failure, I would say that the limits only counting toward TBO are inaccurate. Exceeding the time/power limit could definitely result in immediate failure. A hot spot in the cylinder, suffered as a result of exceeding the conservative limits could lead to detonation, and very rapid temp rise, which will see the cylinder leave the aircraft, or the rod detaching from the crank in some spectacular display. Next time I'm at the hangar, I will provide some photo evidence of the problem. I have one on my bench with a large hole, where a large hold does not belong.