Well, yes, I said it wasn’t a good correlation!

But then you point out other contributors to various failure modes? The topic I’m addressing is the crack initiation, growth, and eventual thru-wall leak. That is primarily due to stress cycles at an overstressed location, not time at temperature. Time and temperature are not the primary contributors. If so, then the new radiator design will fail in similar time, right, regardless of its redesigned stress riser? Several of my Lexus radiators on the top hot side discolor, and although I’ve never tested the material, I’ll readily concede that some sort of material degradation occurs over the years/miles. But that does not cause this particular type of failure. I think I recall your pics (I saw a few), and recall observing some internal surface cracks in the “brown” plastic. Again, I consider that “normal degradation” as none of those were leaks, right? The entire top side material is the same and at essentially the same time/temperature environment – thus those are not contributors even though the material continues to degrade.
Additionally, I could still take your time/temperature belief as primary contributors and demonstrate how poorly that corresponds to mileage. Similar comparison, except substitute speed for time. An average speed of 50 mph (morning commuter jumps on interstate for most of their drive) will achieve 3200 hours in 160k miles, whereas an average speed of 25 mph (stop and go drives to school, errands, city driving) will achieve the same 3200 hour mark in 80k miles. Thus, poor correlation to mileage regardless of stress cycle or time at temperature. {which is why low mileage vehicles aren't necessarily any better}
I agree with what you call “the biggest problem”, but that doesn’t mean it’s a big problem. To me, the big problem is that I can guarantee a leak will occur in a relative few cycles if nothing else is done – which means just replace the radiator now with the redesign (several will advocate the “better safe than sorry” stance). I chose to try a $15 dollar/15 minute patch with a belief that I understand the failure mode, and that this repair attempt could mitigate. I don’t need to eliminate the stresses…I just need to reduce the stress to less that what is required to propagate the existing crack. You are right in that “a more dramatic failure” mode could be created, but I seriously doubt that will happen. I think partial separation of the patch will occur, allowing seepage to escape and be observed. Please note how my patch is not that wide…that too was intentional. Continued crack propagation will eventually peek out the side and leakage could be observed. There is no benefit to having the patch be wider (but there is benefit to having the patch maximized in the front to back direction).
Perhaps, to mitigate your concern, place a piece of string perpendicular to the crack orientation, at the crack’s center, and lay it toward the front of the vehicle, before applying the exotic goo. The small compromise in strength is negligible, but now you have a leak path to monitor?