This is an earlier experiment. Read the amendments below before treating its findings as current. This source entry is marked unverified.
The question
In wind, lanes sat hundreds of metres downwind of their tracks.
What was tested
Two ways to angle into the wind were compared on several test sets: steering on the drift each craft measures, and steering on the forecast steady wind only. A lane rule for the one crossing pair the forecast version made worse was queued.
What was found then
Steering on measured drift chased gusts and cost collisions (one sweep: 107 to 123), so it was rejected. The forecast version brought the express lane back from 741–879 m off track to 69–195 m, with about the same impacts, and was put on hold. The lane rule had only partial results (75 to 79) when testing paused.
What we learned later
Never resumed after the 2026-09-08 pause: no commit or record of either experiment exists, and the simulator has no steering into the wind (it mentions a sideways drift angle only as the passive effect of wind).
equilibrium/viz/sim_core.js:3754; git log (no commit)
What changed
Nothing at the time (the forecast version on hold, not adopted).
Numbers and source references
- 107 → 123
- cross sweep, measured-drift version (Flight Plan only)research/flight_plan.html (Rev D-cont)
- 741–879 m
- express lane off track before, wind 5 (Flight Plan only)research/flight_plan.html
- 69–195 m
- after, forecast version (Flight Plan only)research/flight_plan.html
- 75 → 79
- lane-level partial result, two seeds (Flight Plan only)research/flight_plan.html