This is an earlier experiment. Read the amendments below before treating its findings as current.
The question
The spacing rule has several parts. Which one sinks the low layer?
What was tested
New test switches (off by default) moved where the push away from the ground starts, or cut the avoidance push beyond a set distance: every direction, softly, vertically only or sideways only. Two layouts.
What was found then
Moving the ground push's starting height did not stop the sinking. Cutting the far avoidance push in every direction lifted the layer 3.6 to 10.6 m; this cut also changes when a separate speed-holding rule applies. Vertical-only or sideways-only cuts lifted it under 2 m. The downward push came from the layer above, but the sinking did not follow it in proportion. The vertical-only cut left the craft's downward acceleration unchanged, read then as 'the removed push never reached the aircraft'. Lifted versions lost fewer craft but came closer in their worst near-misses. Cause open; line closed.
What we learned later
'The push never reached the aircraft' was withdrawn: the cut never lowered the total downward command, and it did show up as more upward motion and less net sinking.
research/design/prio_sag_realized_record.md
What changed
Test switches for the ground-push height and the avoidance cuts, plus counters, were added, all off; nothing adopted.
Numbers and source references
- +3.6 / +4.5 m
- lift from a hard cut at 1.71 spacing standardsresearch/design/prio_sag_line_close.md
- +10.0 / +10.6 m
- lift from the half-size rule's push profileresearch/design/prio_sag_line_close.md
- −0.2 / +0.7, +0.8 / +1.6 m
- lift from vertical-only and sideways-only cutsresearch/design/prio_sag_line_close.md
- 21 % / 28 %
- downward command removed by the vertical-only cut, with no liftresearch/design/prio_sag_line_close.md
- 20–32 vs 46–58
- craft lost, lifted versions vs own ruleresearch/design/prio_sag_line_close.md