Many think the "strap" or thin bottom of the rocker fails because it is simply too thin. If  you look at the change in "actual" leverage as the adjuster ball extends away from the rockerat full lift., you can see what I mean. As the rocker gets to full lift , the pushrod is putting a tremendous "rear load" on the threaded portion of the rocker and the load ( read load induced FLEX) goes to the weakest point an begins to work harden and eventually crack the "strap" of the rocker. Aluminum or ductile iron it still does it . reducing the "backwards load" by keeping the majority of the adjuster in the rocker itself reduces the load significantly. A slight inward angle on the adjuster could help "IF" it didn't cause the relief in the rocker body to be deeper. Going to a "cup adjuster" and ball ball pushrod helps with that issue as the cupped adjuster doesn't move the same as a cupped pushrod end does. This is why Jessel , T&D and POP , have gone that way. I helped FPP figure that out  about 35 years ago during their transition from the Harland Sharps to their own ( now POP).
    Randy