Burned valves are caused by machining errors at time of installation, the defects take time to appear, usually 600 hours +600/-200 and any mixture abuse is likely to speed up the process due heat and pressure, but the defect is there from the start and not pilot induced.
What evidence do you have for this? You would have to carefully inspect valves prior to installation and run hundreds of engines to prove that no valves without machining errors burned.
Or perhaps if you inspected every burned valve and found pre-existing machining errors in all.
I doubt that either has been done. This is the sort of blanket statement that turned me from a LOP believer 10 years ago when I first read John Deakin's columns to an APS sceptic now.