How steep is that hill, really?

It depends how far you measure — the same hill can honestly be 32% over 2 m and 25% over 100 m at once. So every gradient published here states the distance it was measured over, computed from national laser surveys by one published method anyone can re-check. One survey, country by country, wherever open LiDAR exists.

How to read a gradient ladder

Every climb carries the same ladder of rolling maxima — the steepest 10 m, 25 m, 50 m and 100 m anywhere on the road — so a short sharp ramp and a long grinding pass can be compared without either being flattered. Rankings use the 25 m rung, and the Length and Overall columns beside it say whether a row is a brief pitch or a sustained climb. A value measured under tree canopy or beside a structure is marked ≈, and every climb carries a evidence-quality letter — saying what the figure rests on, not how accurate the number is — how to read these numbers →

This is an open survey, not a vetted route guide: every number can be re-checked from the open data, corrections are welcome from anyone — and access, surfaces and safety are yours to judge on the ground. The survey is a work in progress, updated as and when new data lands and review time allows. The measurements themselves are the product: every climb page ships a machine-readable OCL document, and each country’s whole page catalogue downloads as a CSV.

Which hills get a page

We measure every paved public road in a region that shows any hint of being steep — tens of thousands of roads so far. Most turn out to be ordinary; they stay in the dataset. Around 25,300 have a page of their own — 600 to 6,600 per country — and the pages are the survey’s public face. A hill earns one in any of three ways:

• It is seriously steep. Somewhere on the hill there is a stretch of at least 25 continuous metres steeper than that country’s cutoff — currently 20% — where the laser has a clear view of the road. Where trees, bridges or walls make the measurement less certain, the bar is higher (26%), and the page passes the survey’s battery of checks before it ranks — classifiers built from thousands of human rulings, listed in the methodology. A case the checks cannot settle goes to a person, and is listed but not ranked until it is; as new checks are built, earlier pages are run through them too.
• It is a big climb. A long, unrelenting ascent (on the order of 1.5–2 km at around 10% or steeper), or one of the 30 biggest net rises — bottom to top — in its country.
• People already know it. Championship venues, guidebook climbs and race courses, from the recognition sources each country’s page lists. Mentions on other lists are noted on pages but don’t create a page by themselves.

The cutoffs are editorial, not physics: they are set to keep each country’s pages a manageable, human-reviewable number, and they loosen as review capacity grows. A hill below the cutoff is still measured, still in the dataset, and still checkable — absence of a page is never a verdict on the hill. Each country’s page says what its own open data supports.

The surveys

United Kingdom3484 climbs · 388 km of total ascent (rough)all four nations, where the open British laser record reaches — and where the survey began

UK rankings →Map · CSV

United States5760 climbs across the six New England states and New Jersey, New York, Pennsylvania, Hawaii, Oregon, Washington, California, Utah and North Carolina · 3465 in the rankingsthe Green Mountain gaps, the Northeast's hillclimb race courses, and America's steepest mile, on the public-domain USGS 3DEP 1 m laser surveys

US rankings →Map · CSV

Switzerland3762 climbs across 26 cantons of Switzerland · 2054 in the rankingsthe great alpine passes and the steepest village ramps, on swisstopo’s swissALTI3D 0.5 m laser surveys

Swiss rankings →Map · CSV

Austria4394 climbs across 9 Länder of Austria · 1378 in the rankingsthe great alpine passes and the steepest village ramps, on the national 1 m airborne-laser terrain model (BEV and the Länder)

Austrian rankings →Map · CSV

France6595 climbs across the Alps, the Vosges, the Pyrenees, the Massif Central, the Jura, Corsica, Provence, the north-west, the north-east, the Centre and the south-west (77 departments) · 4325 in the rankingsthe great cols and the steepest village ramps, on IGN’s LiDAR HD 0.5 m laser surveys

French rankings →Map · CSV

New Zealand698 climbs across 16 regions of New Zealand · 359 in the rankingsBaldwin Street, the great mountain passes and the steepest village streets, on the national 1 m LiDAR terrain model (Toitū Te Whenua Land Information New Zealand)

NZ rankings →Map · CSV

Norway613 climbs across 14 fylker of Norway · 385 in the rankingsthe great mountain roads and the steepest village ramps, on the national 1 m terrain model (Kartverket)

Norwegian rankings →Map · CSV

Corrections welcome

Seen a sign we don’t show, a gate, a seasonal closure, or a better name for a road? Submit an update — no account needed, photos welcome via the GitHub option there. On-the-ground observations are recorded as evidence and tested against the measurement, exactly like the signs and map tags already in the dataset.

Last updated 17 October 2026 — what changed. The method travels: anywhere that publishes open elevation data can be next. Reuse is welcome: any individual climb’s page and charts are CC BY 4.0 and the dataset is ODbL 1.0, with attribution to the national surveys the numbers are measured from. How it works, in full: methodology · the OCL format · licensing & reuse · why this survey exists.