One should think, after all this time has passed, that one of the prominent testing labs, would have, tested, afci's in a lab setting where all sorts of typical arcing fault setups would be employed. We long have read posts that detail exactly what to look for in troubleshooting afci's such as grounds touching neutrals, overdriven staples and on and on. Take those situations and build a testing lab out of possible things that could start up arcing faults. Shoot a few nails in someplace close to wiring. Install a few outlets with the terminal screws loose. Don't tighten some wire nuts. All these should have been attempted in a lab and the facts and data published. That it hasn't been attempted yet is sure fire proof that the devices are phony and they well know so.