This is a published test protocol, not a completed hands-on ranking. Paid accounts, mobile apps and offshore trials have not been tested here. No scores or affiliate recommendations are inferred from vendor claims.
Three identical tasks
- Aegean island leg: locate the named official sectors; compare wind/gusts at departure and the exposed channel, with run and valid time.
- Biscay crossing: compare two user-supplied departure times along the same checked track; retain wave height, period, direction and arrival conditions.
- Bahamas/Caribbean: obtain the correct OFFNT5/OFFNT3 zone text, cyclone message and low-bandwidth update; distinguish wave data from reef-navigation data.
Tools and status
- Windy and Windy.app: two different products; interactive browser tasks pending.
- PassageWeather: chart interpretation and update/valid-time task pending.
- PredictWind: browser/free-tier and Offshore app tasks pending; paid routing not accessed.
- SailGrib: Android device test pending.
- OpenCPN: desktop/plugin and actual GRIB import test pending.
Evidence sheet for each test
Record product/app version, platform, account tier, UTC test time, identical coordinates/track and requested valid hour; save a screenshot or output where permitted. Mark run time, units, coverage, missing data, warning access, offline behaviour and bytes transferred. Distinguish directly observed behaviour from documentation. A smooth map is not proof of forecast accuracy.
Acceptance criteria before a recommendation
Complete all three tasks, identify failures and paid-only functions, repeat on a second date and have another person reproduce the result. Verify accuracy only against suitable independent observations, not another app displaying the same model. Until then this lab remains excluded from search indexing; the existing tools guide remains available.