Just a little historical anecdote:  When we at Bendix were developing the very first US production fuel injection system (remember the Cosworth Vega?) we sized the injectors for about 80% on-time at maximum manifold pressure, minimum air temperature and full power.  This is a compromise, of course - a low-flow injector is desireable to have the most accuracy at idle when the pulse width is very short.  But if the injector approaches 100%, there will be a pulse width less than 100% when the injector will stay fully open, creating a step in fuel flow.  This is a normal occurence during cold engine operation when the enrichment algorithm commands more flow, but it turns out that a max engine speed there is little or no need for cold enrichment regardless of temperature, so the effect is pretty much ignored.  The idea of sizing the injectors for 80% duty cycle had nothing to do with racing applications, but it applies there, too.  The obvious problem encountered when undersizing the injectors is that once 100 duty cycle is reached the flow is limited - the engine will lean out if the rpm or air flow goes higher.  Not a good thing.

Inceidentally, there is no concern for injector relaibility or durability.  It is a simple solenoid valve and running at 100% duty cycle would theoretically make it last longer - it is the total number of open-close cycles that wears the injector.  As injectors wear the flow generally increases and most modern injectors are rated for a billion cycles with a flow increase of less than 3%.  The coil is cooled by the fuel itself, so overheating is not a concern.

Gary Casey