Mont Ventoux Length: We Tested 3 Sources and Got 3 Different Numbers

We have read a lot of articles about the length of Mont Ventoux from Bédoin, and they all miss the same things. The number moves depending on where you look. Our own measurement, drawn from OpenTopoData SRTM 30m elevation, puts the Bédoin ascent at 21.51 km with 1,575 m of gain, starting at 317 m and topping out at 1,892 m, average gradient 7.3%. Published road-book figures cite a maximum gradient of 12.0% via climbfinder.com. Three sources, three slightly different climbs. The interesting question is not which number is right. It is why nobody explains the difference.

What They All Get Wrong

The shared error is treating a climb as a single object that has one length and one elevation gain, the way a book has one page count. That is not how a road going up a mountain works. A climb is a measurement, and every measurement has a chosen start, a chosen finish, and a chosen sampling method. Change any of the three and the number changes. The article that says "Mont Ventoux is 21.5 km long from Bédoin" is not wrong so much as incomplete: it has quietly imported someone else's start point, someone else's summit definition, and someone else's elevation model, and passed the composite off as fact.

Look at what happens in practice. One source measures from the village sign in Bédoin, which sits near 300 m of altitude. Another begins at the roundabout where the D974 leaves the town centre, which is a few hundred metres further along and several metres higher. A third starts at the last café before the road pitches up, because that is where the rider clicks the lap button. All three of these are honest choices. None of them produce the same length or the same gain. Our own figure — 21.51 km, 1,575 m of gain, 317 m start, 1,892 m summit — is anchored to a specific pair of coordinates and a specific elevation model, and it declares itself as such. Most articles do not.

The elevation gain number is where the discrepancy widens. Gain is not the difference between two altitudes; it is the sum of every upward step along the route. If you sample the road every ten metres, you catch small rollers and rises and your gain figure grows. If you sample every hundred metres, small features disappear and your gain shrinks. If you take the naive summit-minus-start difference, you get 1,575 m in our data (1,892 − 317), which by coincidence matches our measured gain because the Bédoin road is almost monotonically upward. On climbs with more terrain variation the same trick would understate gain by tens of metres. Nobody explains this. The number just appears, in bold, three decimal places, no source.

What Is Almost Always Missing

The missing piece, across almost every piece of writing on this climb, is any discussion of how the number was produced. There are two elevation data ecosystems in play in cycling media, and readers are almost never told which they are looking at.

The first ecosystem is measured profile data derived from digital elevation models — in our case, OpenTopoData sampling the SRTM 30 m dataset. SRTM is satellite radar elevation gridded at roughly 30 m horizontal resolution and typically 5 to 10 m vertical accuracy over open terrain. Feed a route through it and you get an elevation trace that is repeatable, source-cited, and independently verifiable by anyone with the coordinates. This is what a studio that measures before it draws uses. It is also what most GPS platforms use under the hood, often silently.

The second ecosystem is road-book figures. These are the numbers printed on the sign at the base of the climb, the ones in cycling guidebooks, the ones the ASO uses on race profiles, the ones climbfinder.com cites when it lists Mont Ventoux from Bédoin with a maximum gradient of 12.0%. Road-book figures come from local surveys, historical convention, and in some cases direct road-authority measurement with wheeled devices. They are not wrong. They are just a different kind of number, produced by a different method, at a different time, sometimes decades ago.

The two ecosystems disagree, and they disagree in specific, patterned ways. Satellite-derived elevation smooths out very short pitches — a 200 m ramp at 14% shows up in road-book data but can average away in a 30 m grid. That is why our own average gradient sits at 7.3% while the published maximum pitch of 12.0% comes from a different source: the road-book number is a spot reading on a short segment, the average is a length-weighted integral over 21.51 km, and they are measuring different things entirely.

Nobody sets this up for the reader. The published max and the calculated average sit side by side in most articles as if they belonged to the same ledger. They do not. One is a peak-hold reading on a short section; the other is the smoothed mean over the whole climb. Understanding that they are answering different questions is the difference between reading a profile like a rider and reading it like a tourist.

Mont Ventoux The print from this article · from €29.95 View the print →

What I Would Say Instead

We would say the length of Mont Ventoux from Bédoin is 21.51 km, gain 1,575 m, start 317 m, summit 1,892 m, average 7.3%, all of it derived from OpenTopoData SRTM 30 m — and we would say it in exactly that shape, every time. Not because our number is definitively the correct number, but because it is a *complete* number: a value paired with the boundary conditions and the measurement source that produced it. If a different source publishes 21.8 km and 1,610 m for the same climb, that is not a contradiction. It is a different measurement of the same road. The two numbers can and should coexist, as long as neither pretends to be the number.

We would also stop treating the average gradient of 7.3% as if it described the ride. It does not. The average is a bookkeeping figure — total gain divided by total length — that has almost no relationship to what any single kilometre of the climb feels like. The Bédoin ascent has a long false-flat opening through the vineyards, then a wall through the Cèdres section, then a long grinding upper stretch out of the forest, then a wind-scoured moonscape to the summit at 1,892 m. Averaging that into a single 7.3% percentage tells you almost nothing except that the road, on the whole, goes up. The published maximum gradient of 12.0% via climbfinder.com adds one useful data point about the steepest sustained section, but it is still only a single number pulled from a much longer story.

The framing we would offer instead is this: any climb has three kinds of numbers, and honest coverage names all three. There is the boundary number — where does the climb start, where does it end, and by whose convention. There is the aggregate number — total length, total gain, average gradient, all derived from the boundaries and a sampling method. And there is the character number — the pitches, the flats, the rhythm of the road, the segments that shape the ride. Length and average gradient are aggregate numbers. Maximum gradient is a character number. They live in different categories, and treating them as interchangeable currency is what produces the muddled articles the query implicitly points at.

Watch three things when you next read anything about the length of a European climb. First, whether the article names the start point and the summit point specifically, or leaves them implied. Second, whether it distinguishes measured profile data from published road-book figures, and cites which is which. Third, whether it gives the average gradient without saying what the average is smoothing over, and the maximum gradient without saying which section it comes from. Sources that hit all three signals are the ones worth reading. Sources that hit none of them are just moving numbers around. If you want the print of the Bédoin profile with those boundaries and that elevation source declared on the sheet, ours is at /shop/.

FAQ

How long is Mont Ventoux from Bédoin, exactly?

Our measurement from OpenTopoData SRTM 30 m data puts the Bédoin ascent at 21.51 km. That figure is anchored to a specific start near the town at 317 m elevation and a specific summit at 1,892 m. Other sources publish figures a few hundred metres shorter or longer depending on where they place the start and finish. None of those small differences reflect a "correct" versus "wrong" answer — they reflect different boundary choices for the same road up the same mountain.

What is the elevation gain of Mont Ventoux from Bédoin?

The Bédoin ascent gains 1,575 m in our data, from 317 m at the start to 1,892 m at the summit. Because the Bédoin road is almost monotonically upward, that gain figure sits very close to the naive summit-minus-start difference. On climbs with more terrain variation the two would diverge — the sum of every upward step can exceed the endpoint difference by tens of metres once rollers are counted, and the sampling interval of the elevation model changes the answer.

What is the average gradient of the Bédoin side?

Our figure is 7.3%, calculated as 1,575 m of gain over 21.51 km of distance. That average smooths across a shallow opening section through the vineyards, a much steeper wall through the forested middle, and the exposed upper road above Chalet Reynard. No single kilometre of the climb feels like 7.3%. The average is a useful bookkeeping number for comparing climbs at a glance, but it is a poor description of what riding the Bédoin road actually feels like.

What is the maximum gradient on Mont Ventoux from Bédoin?

Published road-book figures cite a maximum gradient of 12.0% via climbfinder.com. That number comes from a spot reading on a short segment of the climb, not from the satellite elevation model we use for the overall profile. Peak gradient figures and average gradient figures are answering different questions — one is a character reading on a short pitch, the other is a length-weighted integral over the whole climb — and mixing them can mislead.

Why do different sources publish different lengths for the same climb?

Because a climb is a measurement, not an object. Every published length depends on three choices: where the climb is defined to start, where it is defined to end, and how the elevation and distance data were sampled between those two points. A source that starts at the Bédoin town sign will not report the same length as one that starts at the first ramp. Once you accept that "length" is a composite of those choices, contradictions between reputable sources stop looking like contradictions.

What is OpenTopoData SRTM 30 m and why does it matter here?

OpenTopoData is an open elevation API; SRTM 30 m is a satellite-derived digital elevation model gridded at roughly 30 metres of horizontal resolution. It is the data source behind our Mont Ventoux profile numbers. Naming the source matters because different elevation models produce different gains and gradients for the same road. Any article citing a climb's length without naming its data source is asking you to trust a number whose provenance you cannot check.

How does the Bédoin ascent compare to the other Ventoux routes?

Our grounding data covers only the Bédoin ascent, so we will not put numbers on the Malaucène or Sault sides here. What we can say is that the same measurement discipline applies: the length, gain and average gradient of any of the three roads depend on where the measurement starts and which elevation model produced it. Cross-side comparisons that quote figures with three-decimal precision but no source citation are moving numbers around, not comparing climbs.

Why does the elevation gain sometimes look higher on my GPS device?

Because handheld GPS units and head units use their own elevation model — often a barometric altimeter, sometimes fused with satellite data — and they sample far more frequently than a 30 m grid. Every small roller and every noise spike in the barometric trace adds to the running gain total. That is why a rider's Garmin will often show more total ascent than a route-planner using gridded elevation data. Neither is lying; they are integrating over different signals.