This is an earlier experiment. Read the amendments below before treating its findings as current.

The question

Held level, the craft was still lost. What hit it, and from where?

What was tested

Measured line-keeping, what the corridor heard and pushed, and lane re-entries near the craft. A re-entry is a simulation artefact: a craft reaching its lane's end is put back at the lane start.

What was found then

With the rules the craft's typical end-of-trip offset was 25–75 m at 1.5× and 2× (without priority, 137–238 m). The push records showed the radio picture was not late. The colliding craft had just been put back at its lane start 60–200 m ahead, 1–4 seconds before impact. Of 28 distinct losses on file, 17 were such re-entries. Re-entries under the corridor were common, but fewer than 1 in 40 led to a loss within 8 seconds. All four 3× losses were re-entries at the west edge; the diving 2× radio losses were not.

What we learned later

LATER / CYCLE 212

'With the corridor on, the craft meets more re-entries' held on four test cases, reversed on one and was flat on two; the loss rate is an observed fraction from five runs.

research/design/prio_reentry_record.md

LATER / CYCLE 251

'Not a corridor effect' was withdrawn at the stage-5 review: re-placed traffic does not show the corridor was irrelevant.

research/design/prio_stage4_close.md (Established 2)

What changed

Read-only counters added; the planned rule tied to the corridor was withdrawn before it was built.

Numbers and source references
60 / 20 / 80
runs in the sub-stepsresearch/reviews/researcher-c209.md; research/design/prio_3x_1s_record.md; research/design/prio_reentry_record.md
137–238 m
median offset from line without priorityresearch/design/prio_stage4_close.md (Established 1)
25–75 m
package's median lap-end offset at 1.5× and 2×research/design/prio_stage4_close.md (Established 1)
60–200 m
distance ahead at which the partner re-enteredresearch/design/prio_axis_design.md
1–4 s
warning before impactresearch/design/prio_axis_design.md
17 of 28
unique losses on file that were other craft re-entering their laneresearch/design/prio_axis_design.md
2–56
lane re-entries under the emergency craft per runresearch/design/prio_reentry_record.md
under 1 in 40
re-entries followed by a loss within 8 sresearch/design/prio_reentry_record.md
4 of 4
3× losses that were west-edge re-entriesresearch/design/prio_reentry_record.md
3 vs 0
3× / 1 s cell on fresh seeds, package vs without priorityresearch/design/prio_3x_1s_record.md
0 of 5
diving 2× radio losses that were re-entriesresearch/design/prio_reentry_record.md
Original structured record ↗