I agree that 99% of rod failures have to do with high rpm, and a lot of the time it's right as the throttle is lifted (the engine is compression braking or freewheeling). That is not to say that rods don't bend or break under power. Rods (aka pillars or columns) are typically better in compression than in tension. However, when you induce angles to the loads, then you start introducing some bending moments as well. Those bending moments can be on more than 1 axis, hence why H-beam rods are popular as they resist those types of loads much better than say a round cross section.
I find it interesting that people have problems bending wrist pins. I can understand what may be causing this, but even in hydrolocked engines where the rod is bent/broken due to a compression load much greater than it was ever designed for I've found it not that common for the pin to be bent. Given that the piston is aluminum I'd expect the pin bore to elongate due to work hardening the aluminum before the pin bending. Of course the material and thickness of the pin will have a large impact on what actually happens, plus it has to be considered that the pin is in single shear while the piston is in double shear.
From the descriptions of what people are doing when they have these issues it almost sounds more like a tuning problem given that the reason stated for this phenomena is peak cylinder pressure. There is a larger chance of things not being happy when that happens close to TDC. It almost sounds like a little compromise on torque by retarding the timing a bit would help this. Of course I'll admit that I have a lot of theory background, but my real-world is lacking...a LOT.
I just did a basic search for bend wrist pins (for all kinds of applications), and it really doesn't seem that common. Thickness of the wall seems to be the main culprit, then material, then pin length. There are a LOT of applications that run the .866 (22mm) pin size. From the limited search I did it seems about 600whp is around where people start to see problems with the stock .866/.150 thick pins in DSM's. The one upgrade I found was an increase in pin wall thickness to .225. It looks like an upgrade to 9310 steel is also very common, though not cheap (probably the alternate upgrade being offered by out vendors), the one site I finally found a price for pins alone were just over $100 for a set.
It also sounds like beyond the reduced fatigue due to the constant deformation of the pin that lubrication plays a large factor in pin failure. Once the pin is scored, it is doomed. That stress riser will kill the pin.
One more factor I found, and honestly didn't take into account due to I was thinking about the bore deforming, was that when a pin bends, it is also putting stress on the piston and can lead to a broken piston as well. That would be pretty nasty.
Anyway, the pin diameter doesn't seem to be the issue. It's the wall thickness and material.