This is an earlier experiment. Read the amendments below before treating its findings as current.
The question
Some fast runs sat about 1.1 km beside the lanes all hour. What held them there, and would a stronger pull toward the route help?
What was tested
Examined saved traces, then paired runs with the pull toward the lane capped for the emergency craft's own speed rather than the fleet's. Three route questions went to the owner.
What was found then
Out there the pull sat at its 5.6 m/s cap every second, while the craft's own dodging pushed it away. Nearby traffic was sparser, as a description: nearest craft 310–328 m away, against 47–51 m for runs about 400 m off. With the stronger pull the craft stayed within 200 m of its line 78–93 % of the time, no emergency craft was lost, and fleet losses were not worse (one layout each). But its trips took 2.4–2.5 times the straight-flight time, slower than the default.
What we learned later
The lost time came from slower flight, not a longer path. The speed column in the c227 table was wrong: the far runs flew about 48 m/s before the change and 21–22 m/s after.
research/design/prio_pullcap_line_close.md
The brake was the craft matching the speed of slower traffic, not the pull toward the lane.
research/design/prio_align_off.md
The owner later reframed the three route questions as corridors enforced by an airspace authority.
research/design/prio_escape_question.md
What changed
nothing (an existing switch was tested, not adopted)
Corrections retained in the record
Earlier wording: The nearest other craft is 310–330 m away out there, against 47–51 m for the slow runs inside the grid: that is the whole speed difference.
Record correction: The cycle's record scores the far course as 'like', not emptier, by its registered rule (emptier only descriptively; 310–328 m vs 47–51 m for the ~400 m runs), and does not call it the whole speed difference.
research/design/prio_lateral_record.md (L3)
Numbers and source references
- 24 / 4 / 4
- saved traces / re-flown runs / paired pull-back runsresearch/design/prio_lateral_record.md; research/design/prio_pullcap_record.md
- three
- route questions sent to the ownerresearch/design/prio_escape_question.md
- 5.6 m/s
- fleet-sized cap on the lane pull-backresearch/design/prio_lateral_record.md
- 310–328 m vs 47–51 m
- median horizontal distance to nearest ordinary craft, far runs vs ~400 m runsresearch/design/prio_lateral_record.md (L3)
- 16.8 m/s
- pull-back cap sized to the emergency craftresearch/design/prio_escape_question.md
- 78–93 %
- time within 200 m of its line with the own-speed pull-backresearch/design/prio_pullcap_record.md
- 2.4–2.5×
- lap ratio with the own-speed pull-backresearch/design/prio_pullcap_record.md
- 92 vs 114, 22 vs 28
- fleet losses with vs without, one seed eachresearch/design/prio_escape_question.md