How the score is built
Two rankings. Best waves, the default, answers "which are the great waves": up to 60 points for a break's standing among surfers (a Championship Tour, Big Wave Tour or Olympic venue scores 60, a break with a world-class reputation 35, a second peak of the same wave half that) and up to 40 for what the data below says the wave does on its good days. Most reliable answers "where is a random day most likely to be good", and uses the data alone, with one correction: where many spots share the same half-degree wave cell, each extra one pays a small penalty, so a single wave-model cell cannot fill the top of the list. Switch between them in the Sort menu.
Historical baseline, not a live observation. Every spot is matched to an ERA5 ocean reanalysis cell and read across five full years of hourly data, 2018 to 2022: wave height, peak period and wave direction. Wind comes from ERA5 for the local morning window, sea temperature from ERA5 sea surface temperature, and coastal exposure from the full-resolution GSHHG shoreline. The live seven-day forecast shown in each spot is a separate Open-Meteo feed.
Wave quality 35%
How good the wave is when it breaks. Long-period swell, how much deep water there is to deliver it, whether a canyon or a reef pass focuses energy onto the spot, and how abruptly the bottom rises, which decides whether the wave throws or crumbles.
Best days 22%
What it does when a real swell lands: the size it reaches on its top tenth of days and the energy behind it. A spot that is 1.2 m most days and 4 m when it fires beats one that is 1.8 m forever.
Clean days 18%
Groundswell days that also had light or offshore wind in the morning.
Consistency 10%
Share of days with a workable wave, corrected for exposure.
Season spread 4%
How many months clear the consistency bar.
Water and access 10%
Sea temperature as a comfort measure, and how far the nearest airport and town are.
What changed, and why
The first version of this score asked only how often a place gives you a rideable, clean, warm wave. That put Sapinus first and Pipeline 1,545th, because nothing in it described the wave itself: the size was a median, so a spot that is small every day beat one that is huge when it fires; the sea floor was invisible, so a world-class reef and the beach beside it scored alike; and sea temperature carried enough weight to quietly turn the whole thing into a warm-water list.
This version samples the sea floor from ETOPO 2022 at fifteen arc-seconds, reads swell at its ninetieth percentile instead of its median, and cuts the weight on water temperature. The weights were fitted against a calibration set of seventy widely recognised breaks and checked by how high the model places them. That is a modelling choice, not an independent measurement of wave quality; the data-only ranking remains available as Most reliable.
Why reputation counts in the default list: nothing in a global wave model can see a reef ledge or a sand point, so on data alone Pipeline sat in the 800s. The surfers who have ridden these waves for fifty years are the best measurement of that. Most reliable keeps the data-only view for anyone who wants it, and covers all 5,591 breaks (5,756 named peaks) so it still says something about the spots nobody has heard of.
Limits worth knowing: the wave grid is half a degree, so spots within about fifty kilometres can share the same offshore swell cell; some land-adjacent coordinates use the nearest ocean cell. Exposure, orientation, wind and sea-floor context separate them, but cannot turn offshore climate into a measurement at the peak. ETOPO is sampled at roughly 460 metres, enough for shelves and canyons but not a reef ledge or changing sandbar. The model has no verified crowd, localism, rip, rock, access-rule or tide data. It is a comparison tool, never safety advice.
