B
Bill in Co.
Excessive crossposting removed, per someone's good suggestion. 
I don't think they explicitly dispute what I've said for incandescent lamps.
I guess the difference is, you don't believe the filament is under any such
mechanical stress and strain (due to the large current inrush, and its
sudden thermal expansion), and I disagree; i.e., that it DOES contribute to
metal fatigue. Remember the Properties of Materials and Reliability
Engineering courses?

M.I.5¾ said:Then you tell the Underwriter's Laboratory that their results are
completely
wrong.
I don't think they explicitly dispute what I've said for incandescent lamps.
Practical usage also shows that there is no life reduction with
frequent switching in spite of what intuition might tell you.
Incidentally,
metal fatigue has nothing to do with it. Metal fatigue is a phenomenon
that
occurs when metal is subject to mechanical load stress not thermal
cycling.
Tungsten, like steel (but unlike aluminium), has a wide load stress range
over which it will never fatigue.
I guess the difference is, you don't believe the filament is under any such
mechanical stress and strain (due to the large current inrush, and its
sudden thermal expansion), and I disagree; i.e., that it DOES contribute to
metal fatigue. Remember the Properties of Materials and Reliability
Engineering courses?