I think what is fishy is the assumption that the 0.5% loss rate per year for new units would not change over time as the units got older. Hazard rates for produ

I think what is fishy is the assumption that the 0.5% loss rate per year for new units would not change over time as the units got older. Hazard rates for products are not constant like that -- there is a period when the failure rate increases rapidly after the design life of the unit.