This is an earlier experiment. Read the amendments below before treating its findings as current.
The question
All delay results so far were in calm air, and a 5 m/s wind from 250° had earlier lost fewer craft than calm. Does the delay cost carry over into wind?
What was tested
84 half-hour runs on six more new layouts: wind with perfect information and with 0.5, 1 and 2 s delay, plus calm on the same layouts.
What was found then
In wind the 1 s link lost 2.7× more craft than perfect information, on every layout, and 2 s more than doubled that again. The drop at 0.5 s was too small or inconsistent to pass the tests fixed in advance. Calm repeated the previous step. With the same link, wind lost about a third of calm's craft and spent about a third as long in close calls. Whether wind changes the link's relative cost is unsettled (4.9× in calm, 2.7× in wind).
What we learned later
'About a third' holds only with the link on; with perfect information wind lost about half of calm's craft (0.56×). And a similar share of close calls ending in a loss does not by itself show each close call is equally deadly.
research/design/s5_lat_wind.md:45 (c257 reading limits)
Wind's benefit turned out to depend on where it blows from: this was wind from 250°, and wind from 70° did not give the same benefit consistently.
research/design/s5_wind_dir.md:20
What changed
Nothing.
Numbers and source references
- 84
- half-hour runsresearch/design/s5_lat_wind.md (Spec) and :45
- 2.7×
- wind, 1 s link vs perfect information (2.66)research/design/s5_lat_wind.md:45
- 0.075 vs 0.028
- lost per craft-hour in wind, 1 s link vs perfect informationresearch/design/s5_lat_wind.md:45
- 0.168
- lost per craft-hour in wind at 2 s delayresearch/design/s5_lat_wind.md:45
- 0.050 / 0.248 / 0.474
- calm lost per craft-hour, perfect information / 1 s / 2 s, third layout setresearch/design/s5_lat_wind.md:45
- about a third
- wind losses and time in close calls relative to calm at the same link (0.30× at 1 s, 0.35× at 2 s; breach time 15–21 vs 48–58 s per hour)research/design/s5_lat_wind.md:45 (Descriptive)
- 4.9× vs 2.7×
- link cost in calm vs wind, pooledresearch/design/s5_lat_wind.md:45
- 4 of 6
- layouts where calm's link cost was the largerresearch/design/s5_lat_wind.md:45
- six
- new layoutsresearch/design/s5_lat_wind.md:45
- 5 m/s from 250°
- wind speed and headingresearch/design/s5_lat_wind.md (Spec, Rows) and :45
- 0.5 s
- shortest delay tested in windresearch/design/s5_lat_wind.md:45