This is an earlier experiment. Read the amendments below before treating its findings as current.
The question
If right of way does not help, something in the emergency craft's own steering must slow it. Which part?
What was tested
40 half-hour runs with the wing at 2× traffic speed and right of way, in four versions: no own dodging; its matching of neighbours' speed cut to a fifth; both removed; and the craft alone in an empty sky.
What was found then
Alone, the wing flew exactly its straight-flight time, so the lane and airframe cost nothing. Removing its own dodging recovered nothing. Cutting its speed-matching to a fifth recovered about 70 % of the delay (on the grid, from about 2.0 to 1.3 times the straight-flight time) at no measurable safety cost. A craft that neither matches speed nor dodges was on time, but the grid lost about a quarter more ordinary craft (106 against 84).
What changed
A switch that scales how strongly the emergency craft matches its neighbours' speed was added, off by default; nothing adopted.
Corrections retained in the record
Earlier wording: the ghost's 1.00 is bought with 27 % more ordinary losses on the grid (106 vs 84 craft)
Record correction: The record gives +26 % (×1.269 per craft-hour; 106 vs 84 craft).
research/design/prio_delay_record.md (P8)
Numbers and source references
- 40
- half-hour runsresearch/design/prio_delay_record.md
- 2×
- emergency craft speed relative to trafficresearch/design/prio_delay_record.md
- five
- seeds per versionresearch/design/prio_delay_record.md
- four
- versions compared (plus the as-flown baseline)research/design/prio_delay_record.md
- 1.000
- lap ratio alone in empty sky and with neither flocking nor dodgingresearch/design/prio_delay_record.md
- 2.006 → 2.021
- grid lap ratio with its own dodging removedresearch/design/prio_delay_record.md
- 2.006 → 1.299
- grid lap ratio with flocking cut to a fifthresearch/design/prio_delay_record.md
- 1.515 → 1.148
- lattice lap ratio with flocking cut to a fifthresearch/design/prio_delay_record.md
- 70 %
- share of the delay recovered by cutting flockingresearch/design/prio_delay_record.md (P6)
- 106 vs 84
- ordinary craft lost, craft neither flocking nor dodging vs right of way as flownresearch/design/prio_delay_record.md