Hoss, I have no doubt the team that designed the Cabri know what they are doing.
Miners law allows you to combine a number of different loadcases to determine the overall fatigue damage, or to determine the life of a component. The problem is working out where each loadcase is on the Stress-Number_of_cycles curve. For stress calcs the traditional method is using the Goodman's diagram, or variations of it, to combine average and cyclic stress into an equivalent cyclic only stress to check against S-N curve. The problem is that this method is not accurate in life prediction, as there is huge scatter in coupon S-N testing.
A more recent development comes from the understanding that even below the elastic limit there is some residual strain after each load application. Basically the material is not perfectly elastic, but always recovers alongs Young's modulus. Consider the 0.2% plastic strain proof stress generally regarded as material yield, if you reach this then unload the component you will have a residual 2000 microstrain. Lets say taking the component to half Ultimate Tensile Strength, then unloading it leaves a residual strain of 1 microstrain. This means you can do this 2000 times before the component reaches a strain where it yields on the next application.
If any of the design engineers that occationally appear from the aether vaccum can offer good practical guidance here, i'm interested BTW. It seems to me that the only chance of producing reliable numbers is to use the fatigue software built into many FE codes. It's a specialised area.
If anyone else is interested in fatigue, here is some general info:
http://en.wikipedia.org/wiki/Fatigue_%28material%29