I'll cross post what I wrote in the other forum regarding rockers. At least some of it applies to this discussion:
First is that the rocker broke exactly where they do - at the intersection of highest load and thinnest cross section. Whether the failure was from an assembly, machining or a material defect I cannot say.
That is the weakest area on a rocker and you cannot add much more material there with factory shaft diameters - one of the reasons that a T&D or Erson runs a high quality, but smaller diameter shaft. It is also the reason that I do not particularly care for thick walled bushings in most rebuilt stock or aftermarket rocker arms. The only broken stock arms I ever saw had bushings in them and the failure looked exactly like this one.
The roller tip does bring a benefit over the rounded to stock rocker. It uncouples the rotating rocker from the vertical motion of the valve - improving guide life and reducing drag & heat at that interface. One need only watch Bill Conley's SOHC high speed valve action videos to see the amount of heat generated by a sliding contact - he literally had smoke coming off the tip of the valve. The benefits here are likely to be pretty small though in the context of a mild street build.
A bearing at the fulcrum seems to be a good idea. At least it seems that the evil GM LS stuff has been running bearing fulcrum rockers in OE performance applications for several years now. I kind of like the honed ID/bushing at the fulcrum myself - seems to have fewer parts and works OK on wrist pins with few failures. Could be a toss up. Likely a great place for a thinwall bushing on a slightly reduced shaft diameter - or some sort of DLC coating on the rocker's parent material.
Steel is a better material for rocker arms - no question about that. Its a matter of mass versus physical size versus load capacity versus fatigue strength. But given good design, and good quality readily available aluminum alloys, fatigue failure is something to be concerned with in a .800 lift solid roller application running 8000+ RPM in a 500 mile race and never should be a concern in any normal street application. There are eighty gazillion street cars running down the road on aluminum rockers virtually forever, and at least a few OEM aluminum performance rocker applications that doubtless passed durability testing - so aluminum as a material is certainly good enough for this application.
The pushed out adjuster might provide the best clue here. As once explained to me by a Federal-Mogul piston engineer, aluminum alloy heat treating is more complex than initially meets the eye. Some processes will deliver a part that is initially "better", but which will degrade over time with heat. Alternate processes will not look as good initially, but will be more stable over time and can actually be a better choice for a given purpose. This is a long way of saying that the material may have looked and tested perfectly good at the point and time of rocker manufacture, but have not really been correctly treated at the extrusion supplier end. If it were possible to check it to the root cause might have been something as simple as a clogged gas jet in an oven somewhere that hurt a single stick of material...