Originally Posted by
Howste
His experiment certainly is based on probabilities. As the number of samples goes up, the more likely the true probability will be represented in the results.
You're proposing to do a one-time experiment and then use anecdotal evidence to settle the question. Not exactly a statistically valid proposal.
I propose to do as many experiments as you like. Arguing central limit theorem is not on point.
The probability of being unhappy in meal choices on a flight is independent of the number of passengers, ceteris paribus.
The number of unhappy passengers increases as the number of passengers on the flight goes up.
The probability is invariant.
There is no anecdote involved. Its simple math.
Now, if you want to play the raw numbers game, I'm in. You fly 1 trip, I fly 1 trip - you with 100 and me with 10. I pay you $10 for each unhappy passenger and you pay me $10 for each unhappy passenger.
Lather rinse repeat - as many times as you want.
I win, you lose.