S10 · stats trap simulation

The rescue package that rescued 87% of weak flights

People, data and outputs in this scenario are simulated. Each stage is recorded when you commit; the reveal comes after.

5stages
30 minabout
0stages committed

The brief

A revenue-management analyst (simulated) presents a slide for the commercial committee: "Our rescue package (targeted promotions and wider fare-class availability) works. We apply it to the weakest flights, and they recover."

The numbers are computed from flight_day in the course database (synthetic): 2023, 19 flights, weekly load factor (passengers over seats, Monday to Sunday) for each flight; 969 consecutive week pairs. The package is a hypothetical overlay for the exercise: the team says it uses it, but did not log which flights got it. Everything else is data.

Stage 1 of 5

The analyst's table:

Group	Flight-weeks	Load factor in week t	Load factor in week t+1	Change
Weakest 10% of flight-weeks (below 68.6%)	98	64.8%	72.7%	+7.9 points

"86.7% of them improved; the package lifted weak flights by almost eight points."

What is the main problem with the claim?

The 72.7% is still below the network average of 78.3%, so the package did not fully succeed in rescuing the weak flights and needs a stronger version.There is no comparison group: flights picked for an unusually weak week would likely rebound whether or not a package was applied.Eight points is too small to matter commercially, so the package is not worth its cost to the commercial team, even if it worked.