Originally Posted by
s0ssos
I would actual prefer airlines say actual flying time, because there is really no way to account for delays. There are many significant delays, and they measure in hours. How do you split that up into numbers for every day, when they don't happen every day on the same flight? Do you take, say 3 hours that happens once a week and divide it by 7 and add it?
How do you account for ATC, airport congestion? The past isn't a good predictor, as some airports just get busier and busier (like Newark. There was a good article about how it was a sleepy airport back in the day, now is massively overloaded)-forget where it was.
I think you have in your mind that airlines pad the block time for a flight so they don't get dinged for late arrivals. There are far more things to consider.
Remember, if a flight is 30 minutes early, the plane just sits there longer. Also, I'm sure the crew doesn't get penalized for a flight being early, so if you pad you have to pay for a crew for a parked aircraft.
Obviously it makes sense to account for minimal delays, long taxis, and the like, and I'm sure DL has an algorithm to determine how much time to block for a flight given lots of factors. But the padding is there to account for common factors, not to bolster data.