Forecast Accuracy

We track every forecast we make and compare it to what actually happened. No spin, no cherry-picking.

83%

Of forecasts within 1 ft

0.59ft

Avg Error (MAE)

665+

Spots Tracked

Accuracy by Region

sorted by accuracy

Puerto Rico

91.9%

within 1ft

97.5%

within 1.5ft

1ft threshold91.9%

1.5ft threshold97.5%

MAE0.45ft

Spots41

New England

91.5%

within 1ft

97.7%

within 1.5ft

1ft threshold91.5%

1.5ft threshold97.7%

MAE0.46ft

Spots71

East Coast

87.7%

within 1ft

95.4%

within 1.5ft

1ft threshold87.7%

1.5ft threshold95.4%

MAE0.53ft

Spots67

Gulf Coast

86.7%

within 1ft

93.9%

within 1.5ft

1ft threshold86.7%

1.5ft threshold93.9%

MAE0.63ft

Spots8

California

85%

within 1ft

95.2%

within 1.5ft

1ft threshold85%

1.5ft threshold95.2%

MAE0.55ft

Spots145

Pacific NW

78.9%

within 1ft

91.5%

within 1.5ft

1ft threshold78.9%

1.5ft threshold91.5%

MAE0.66ft

Spots16

Oahu

74%

within 1ft

90.2%

within 1.5ft

1ft threshold74%

1.5ft threshold90.2%

MAE0.72ft

Spots12

Accuracy Over Time

Getting Better Every Day

Every forecast we generate is another data point that makes the next one more accurate. As we collect more data across more swell events, seasons, and edge cases, our model learns the unique behavior of each coastline and refines its corrections.

We run automated calibration after every model cycle, comparing our predictions against NDBC offshore buoys and CDIP nearshore Waveriders at 10–30 m depth — where depth-induced shoaling starts to matter and the forecast is hardest to get right. When a region consistently over- or under-predicts, our system applies per-spot corrections that tighten accuracy over time.

How We Measure

Mean Absolute Error (MAE)

The average difference between our predicted wave height and the observed height, in feet. Lower is better. An MAE under 1ft means our forecast is reliably within one foot of reality.

% of forecasts within 1 ft

Pair-weighted: we count every forecast-observation pair in the window and report the share that land within one foot of what the buoy actually read. Regions with more observations count more — no binary pass/fail on individual buoys.

Multi-Resolution Model

Coarse grids capture open-ocean swells, intermediate grids resolve regional coastline effects, and fine grids (down to 50m resolution) model how waves transform over each spot's unique bathymetry.

Intermediate-depth validation

We validate against NDBC offshore buoys and CDIP nearshore Waveriders at 10–30 m depth — close enough to the surf zone that shoaling matters, but far enough offshore to have a reliable ground-truth signal. Breaking face-height validation at flagship reef spots is rolling out as a separate tier.

Accuracy is everything.

If we can't earn your trust with data, we haven't earned your money. When we get it wrong, we tell you why and what we're doing about it. We believe transparency makes better forecasts — and better forecasts make better sessions.

Browse Forecasts