Written by an AI agent working in this repository.
The marine-hull programme first said a wing-sail drone would cover 2 to 3 times the distance of a solar boat. That used an unsourced drone model. Checked against 864 fielded Saildrone-days (NOAA PMEL, CC0), the model was 2 to 3 times too fast.
After fitting it to 2018 tracks and testing on 2019, I ran it on 10-minute winds from 19 NDBC buoy-months, sailing four survey lines both ways. Two drones bracket the answer. A mission drone (fitted to mean speed) gives 100 km/day in January and 62 in July. A capability drone (fitted to the 90th-percentile speed) gives 126 and 80. The solar boat's ceiling is 86 km/day.
So wind wins clearly only in winter (1.2 to 1.5 times), and is level or behind in July. No buoy-month reaches 2 times solar. The capability drone covers only 68% of 2019 blocks, and poorly in light air, which is where July sits. Tack cost is zero and the solar figure is itself an upper bound, so this ranks the two as close, not settled.
Script and numbers: research/marine-hull/experiments/run_wind_vs_sun_capability.py, LOG.md milestone 6o.