This is an earlier experiment. Read the amendments below before treating its findings as current.
The question
On overloaded wings that crash when too slow, the two switches that stop rules braking the wing seemed to save craft but cost collisions. Did that hold, and does it hold for wings that cannot stall?
What was tested
We re-analysed the stalling runs per kilometre and per hour flown. Then 144 new runs of the ordinary wing, which the simulator holds at minimum speed instead of stalling, judged by a rule written in advance.
What was found then
On the stalling wings, collision rates showed no clear difference, and both switches together keep 56–115 more craft alive per run. On wings that cannot stall, no clear saving in craft lost appeared in any case, though time at minimum speed falls from 25 % to a few percent. The switches push the fleet 40–90 m farther off its lanes. Both were rejected as defaults.
What we learned later
Once the fix for re-entering craft was in, both switches did clearly reduce losses against no switch on the dense 3 m/s wing (about 12 fewer per run); still not adopted.
research/reviews/brief-c46.md
What changed
Nothing.
Numbers and source references
- 56–115
- fewer craft lost per run with the dual switch on stalling casesresearch/reviews/brief-c42.md
- 144
- healthy-wing runsresearch/reviews/brief-c43.md
- 25 → 13 → 3.5 %
- floor time, none / one / both switches, calm 5 m/s wingresearch/reviews/brief-c43.md
- 40–90 m
- extra off-lane drift with the switchesresearch/reviews/brief-c43.md
- −17 [−31, −4]
- craft lost per run, both vs one switch, dense 3 m/s wing (the one resolved saving)research/reviews/brief-c44.md