Skip to content

Public model checks

Model revisions, in the open

PeakWave publishes how forecast scores compare with the later Open-Meteo output used for the same date. This measures model revision consistency, not independent weather accuracy.

Within 1 model point

Building

Revision miss rate

Building

Mean revision

Not enough rows

Rows scored

0

Forecast revision consistency

Last 90 days of completed version-matched pairs. A hit means the forecast was within one GO point of the later model score. Closed-season zero-to-zero suppressions and ski observations with estimated snow depth are excluded. The ledger retains 0 closed-season and 0 estimated-depth rows.

Hits

0

Misses

0

Rows

0

Largest spot revisions

No completed rows yet.

Largest sport revisions

No completed rows yet.

The Call

Building

Hit rate

5 of 10 verified Calls collected.

ERA5 drift

0.32 C

Latest temp MAE

Within thresholds across 8 samples.

On-the-ground agreement

100%

Polzeath

1 of 1 reports agreed within one point.

Per-sport breakdown

Touring SUP (Beta)Building, 0 of 30 days collected
SUP Surf (Beta)Building, 0 of 30 days collected
Downwind SUP (Beta)Building, 0 of 30 days collected

Scoring versions

No completed rows yet.

Spot revision ledger

Search spots with enough completed day-ahead comparisons against later output from the same weather provider.

SpotSportRowsStable rateRevisionsMean revision
No matching spot rows yet.

Two different evidence tracks

Model-revision rows compare an Open-Meteo forecast with later Open-Meteo output. They do not validate the weather against an independent instrument. Seasonaire reports remain a separate on-the-ground track.

View live conditions