This. I think there is a form of natural selection that works with incandescents. You start of with a load of random ones but gradually all the short lived ones blow and get replaced and you end up with only good ones that will last for ages.
Duff fittings or being switched on and off all the time mitigate against any kind of long life for them though. Office use where they are switched on once a day or sometimes left on all the time is perfect for maxing out the life of a good incandescent bulb.
Back in the good old days, the effect of turning on and off on lightbulb longevity was the helmet debate
[1] of the stagecraft newsgroups. It goes something like this:
Lightbulbs usually fail when you turn them on...
Therefore turning lightbulbs on is bad for them (inrush current, thermal cycling, etc)...
Therefore they'll last longer if left on continuously, rather than when cycled on and off.
The counter argument is that, manufacturing defects aside, eventual failure is due to the evaporation of material from the filament, which is a function of run time.
Usual digressions include:
The use of pre-heat (commonly theatrical dimmers set to 0% will light the filament to a dull red glow, in order to improve their response time).
Halogen cycle.
Anecdotes involving traffic lights.
Anything from Big Clive under the pesudonym 'Davie Dimmers'.
At the time, I predicted that filament lamps would be thoroughly obsolete before we saw an end to these arguments. I failed to anticipate the withering of usenet due to social media.
[1] Technically, the ability (or otherwise) to stimulate fluorescence using a tungsten source with a #181 (Congo Blue) gel is the more direct equivalent of the helmet debate, but let's pretend it's analogous to the hi-vis debate for general neatness