This is an earlier experiment. Read the amendments below before treating its findings as current.
The question
The winning rules were flown again on the eight test cases fixed at the stage's start. The adoption rule required at least 6 of 8 to qualify.
What was tested
40 half-hour runs with the rules: speed-matching off, own dodging at a fifth, and a 10-second wide reserved corridor ahead. Two cases repeated the previous step exactly, so six were new evidence.
What was found then
The emergency craft was on time on every case (typical trip 1.00–1.05 times the straight-flight time) and never lost. Where they could be scored, ordinary losses were no higher than with the same fast craft without priority. But only 5 of 8 cases qualified: one had a single collision in 1 500 craft-hours against none, and two had no losses in any version, so could not be scored. Close-approach time between ordinary pairs rose 6–8 %, within noise.
What we learned later
The scoring script had called the one-collision case 'infinitely worse'; the stage's rule says such a case cannot be scored. The count stayed 5 of 8.
research/design/prio_scen2_record.md (P15)
The absolute grid results in this stage describe a fixed-wing emergency craft that dives far below its planned 60 m height; comparisons between versions still stand.
research/design/prio_stage_close.md §0b
What changed
nothing
Numbers and source references
- 40
- half-hour runsresearch/design/prio_scen2_record.md
- five
- seeds per cellresearch/design/prio_scen2_record.md
- six
- cells with new evidence (two were exact repeats)research/design/prio_scen2_record.md
- 1.00–1.05×
- emergency craft median lap ratio on the familyresearch/design/prio_stage_close.md §2
- 5 of 8
- cells passing (target 6 of 8)research/design/prio_scen2_record.md (P15)
- 68 vs 80
- ordinary lost, package vs no emergency craft, grid wing 1.5×research/design/prio_stage_close.md §2b(i)
- 1
- ordinary-ordinary collision on the grid point 2× cellresearch/design/prio_scen2_record.md (P15)
- 1 500
- craft-hours on that cellresearch/design/prio_scen2_record.md (P15)
- 0.0002–0.0009
- emergency craft's share of close encounters on the grid wingresearch/design/prio_scen2_record.md (P18)
- 6–8 %
- rise in summed ordinary-pair close-approach time over no emergency craftresearch/design/prio_scen2_record.md (P18)