This is an earlier experiment. Read the amendments below before treating its findings as current.
The question
If the rules cause the crawl, which one: heading copying, keeping distance, or plain crowding?
What was tested
Wing and ideal craft (instant response) at 1,200 craft, one run each with heading copying off, spacing doubled, or only 300 craft, read along each lane.
What was found then
With heading copying off, the ideal craft flies straight (3–4° spread instead of 23–26°) at nearly full cruising speed and loses nothing; no close calls at all were recorded then. The wing also straightens but loses 154 craft instead of 100, so for it the rule also protects. Doubled spacing did not reduce the weave; 300 craft trimmed it by a fifth. For ideal, multirotor and eVTOL craft the weave centres on crossings: heading error peaks at 22–28° and has eased to 8–12° 300 m past, among craft with crossing traffic within 60 m. The wing's middle lanes show no such pattern.
What we learned later
Close-call time without copying was 0.3 %, near zero rather than zero. Doubling the spacing also doubled how far copying reaches, so that test changed two things at once.
research/envelope.md:5451
Withdrawn: a first description of 'a twenty-degree sawtooth on every lane, fading over 600–900 m' overstated the picture along the lane.
research/envelope.md:5396
Withdrawn: 'each excursion lasts 250 m' and 'nothing carries on past the crossing'; whether the disturbance carries on past it stays open.
research/envelope.md:5340
The explanation 'lanes are 100–200 m wide' used the wing's width, not the ideal craft's ~80 m. Why neighbours are near so early became the next question.
research/envelope.md:5279
What changed
Nothing.
Numbers and source references
- 1,200 / 300
- craft in the baseline and reduced-traffic runsresearch/envelope.md:5451
- 2×
- spacing in the doubled-spacing runresearch/envelope.md:5451
- 3.4–3.9°
- ideal craft's heading scatter with copying off (from 23–26°)research/envelope.md:5451
- 99 %
- ideal craft's speed as share of cruise with copying offresearch/envelope.md:5451
- 100 → 154
- wing craft lost per run, copying on vs offresearch/envelope.md:5451
- a fifth
- scatter reduction at 300 craft (ratio 0.80–0.81)research/envelope.md:5451
- 100–200 m
- onset of the heading rise before a crossingresearch/envelope.md:5340
- 22–28°
- peak heading error at a crossing (ideal, multirotor, eVTOL)research/envelope.md:5340
- 8–12° / 300 m
- heading error 300 m past a crossingresearch/envelope.md:5340
- about 600 m
- width of the disturbed zone (±300 m)research/envelope.md:5340
- 60 m
- reach used for 'crossing traffic nearby'research/envelope.md:5279