The calibration difficulty index
A single 0–100 difficulty score for every ski run, computed the same way everywhere, so a “blue” at one resort is comparable to a “blue” at another.
Why a calibrated score
Resort difficulty gradings are self-assigned and inconsistent: one mountain’s blue is another’s black. Every run, though, has objectively measurable terrain. We measure it from open data and turn it into one number that means the same thing everywhere.
What goes into it
The score is driven by the run’s steepness: how steep the steepest sustained pitch is, and how steep it stays on average across the whole run. A short, sharp pitch followed by a long, easy runout scores differently to a run that holds a steep angle from top to bottom, even if their steepest points are similar. We weight sustained steepness more heavily than the single steepest moment, because that’s closer to how a run actually feels to ski.
The exact formula and weightings are ours and stay unpublished, so we can keep tuning them against real terrain without inviting anyone to game or copy the scoring. What we do publish is how to read the number:
| Score | Skis like |
|---|---|
| 0–40 | Green / easy |
| 40–65 | Blue |
| 65–80 | Black |
| 80–100 | Double black |
These bands are for Australia, New Zealand, the US and Canada (green/blue/black/double-black). European resorts grade more assertively for the same real steepness, so their bands shift down: green 0–30, blue 30–45, red 45–65, black 65–100 (no double-black). A Japanese resort (green/red/black, no blue) merges the middle two 4-tier bands into one wide Red, 40–65, since its single Black already covers what the other schemes split into Black and Double black.
What it is and isn’t
The score is an absolute measure of terrain, not a ranking within the current dataset, so it doesn’t shift as more resorts are added. Terrain steepness comes from OpenStreetMap piste geometry sampled against each country’s own elevation model (Geoscience Australia in Australia, LINZ in New Zealand, USGS 3DEP in the US, Copernicus GLO-30 in Canada and Europe), so it’s an estimate at roughly 8–30 m resolution, not a survey. It also doesn’t capture narrowness, moguls, or exposure on short technical features, which is part of why some official black-run ratings read a little harder than their calibrated score. Conditions (ice, powder, moguls) aren’t in the score either, those are what user ratings add on top.