And just to be clear, the experiment should reflect the proposition that with preordering available, those who do not are worse off because the chance of not getting their meal choice goes up as the number of passengers on that flight goes down.
It's about probability on a flight, one flight different numbers of passengers, not different number of flights with the same cumulative number of passengers.
It's the 10 and 100 passenger flight, not the 1 and 10 flights with a total of 100 passengers.