Ten years of categorized climbs - have they changed?

[Danskebjerge]

I've now added this year's climbs to my survey of the Tour de France categories. So in the article that I just published, you'll not only find an overview of the defining parameters of the categorized climbs, but also a comparison between the 2016-climbs and those of several past Tours. In total, 543 climbs are now included in the analysis - some of them date back to 2006.

http://danskebjerge.dk/article-categories-2016.htm

Hope you'll find it useful (and not too nerdy :-D ).

/Jacob T. Johansen, Danskebjerge.dk

 

[infeXio]

Brilliant!

 

[Red Rick]

Very cool. The issues they talk about are actually pretty logical when you think about it. It's perfectly understandable that the Tour usually has more cat 1's then 2's, and that's even when a bunch of HC's are overcategorized. Mountains are really concentrated in France, and the medium mountains are criminally underused. The cat 3's being steeper on average is completely logical as well, as they're largely compromised of hills, which are often steeper then cols. Oh, and I think come cat 1's are really overcategorised too. Some stuff like 8km at 7% is suddenly shouldn't really be a cat 1.

 

[Libertine Seguros]

Côte des Chevrères is the most ridiculous cat.1 in recent years. That wouldn't even be cat.1 in the Vuelta, and they give cat.1 out for climbs like Cordal and Peña Cabarga which are definite cat.2s. How the hell that's cat.1 and Croix-Neuve is cat.2 I will never get.

Hell, Planche des Belles Filles should be cat.2. That's the same length and easier gradient than Peña Cabarga and Urkiola.

 

[Netserk]

Surely it fits well with the other climbs in the category that PdBF is cat. 1? (http://i.imgur.com/uSNlSEs.png From Linkinito)

That said, I too oppose the devaluation of mountain categorization over the last decade.

On another note, I understand why we describe climbs primarily with length and grade, but when doing an analysis, I think it'd make more sense to use vertical gain and grade instead (and altitude and irregularity as secondary parameters).

 

[Libertine Seguros]

You can calculate vertical gain from the length+gradient stats anyway, same as you could work back to the length from the gradient+vertical gain stats.

If you use those curves as a line of best fit then yes, PdBF only just makes the criteria. However I think those criteria are too loose nowadays at the Tour and Vuelta (not at the Giro, which has always been more stingy with the GPM). Even using your secondary parameters, PdBF is a) the same length as the Spanish climbs that are often considered over-categorized (Cordal, Peña Cabarga, Urkiola), b) less steep on average than them, c) not at an appreciably high enough altitude for that to be a factor, and d) more consistent than the latter two at least. It does have a steeper maximum than them, admittedly, which is its main saving grace.

And Côte des Chevrères is 1% steeper in gradient (12% steeper in absolute terms) and 2,4km shorter (amounts to about 40%). Than a climb that I think is over-categorized as first category. Even if you accept that PdBF is a cat.1, Chevrères is cat.2 for sure.

I think another factor is placement within a stage - certainly I think this is a factor in the Vuelta's dishing out of mountains points. Quite often climbs that ought to be cat.2 are given cat.3 early in a stage when unlikely to make a difference, but the same type of climb will be bumped up to a cat.2, even sometimes a cat.1 (Puerto de la Quesera!!!) when placed near a stage finish where action is more likely. That may play a part in justifying both Peña Cabarga and PdBF as cat.1 climbs.

 

[Netserk]

Completely agree with regards to Chevrères. I also think PdBF should be category 2, but not because that particular climb is miscategorized, but as you said that the whole category is too lenient.

The reason I prefer vertical gain over length is that it is a better representation of how long it takes to climb the climb. Length alone doesn't tell as much and depends a lot more on grade. If you have two climbs with the same vertical gain, but where one is steeper than the other, it is obvious that the steepest is most difficult, even though it is shorter, so comparison between climbs makes more sense that way, at least in my opinion.

 

[RedheadDane]

I'm pretty sure the climbs haven't changed.
How they're categorised, and which climbs are used on the other hand, might have changed. :twisted:

 

[Red Rick]

There's plenty of Cat 2's in the Giro nowadays that are HC in the Tour. I thought ascents had to be like 1000m in altitude gain to be HC, but now you have stuff like Pla d'Adet and even Arcalis getting HC status. HC's are supposed to be the truly hard climbs, not the ones a lot of amateurs can do in 40-50 minutes and pro's do within half an hour.

 

[Brullnux]

These graphs are slightly misleading in the sense that they give no indication to a climb's irregularity, which can often be why a climb is Cat 1 instead of cat 2.

 

[Red Rick]

They should just use a system like climbbybike uses, where the amount of points is like (altitude difference^2/distance)*x). You could add up different segments to account for irregularity. Then use a cutoff to determine category

 

[RattaKuningas]

I don't mind shorter climbs if they are steeper. Steeper climbs make race much more exciting than going endlessly uphill on gradient about 5-7%.

 

[jmdirt]

I don't remember which climb it was, but in the early 2000 a climb was used in the CdD and the TdF. It was a cat. 1 in the CdD, but a HC in the TdF. There was a lot of discussion about it at the time. One commentator's take was that the col had grown in the time between the to races. As to Libertine's point, the col was the about the same distance into the race, but *closer to the end of the shorter CdD stage. Maybe one of our good Googlers can fill in the gaps for me...

*EDIT: Maybe it was the finish of the CdD stage?

 

[Danskebjerge]

Re:

These graphs are slightly misleading in the sense that they give no indication to a climb's irregularity, which can often be why a climb is Cat 1 instead of cat 2.

It does seem reasonable to take, for instance, the maximum grade into account. But as far as I can tell from looking at hundreds of categorized climbs, it doesn't play a significant role for the organizers - it's the length and the average grade that matters to them.

Maybe that also has to do with the nature of the climbs. The really big climbs tend to have steady grades (possibly more in France than in e.g. Italy), while the shorter climbs may have changing percentages, but they are so short that it doesn't affect the categorization much.

/Jacob

 

[Danskebjerge]

Re:

I don't mind shorter climbs if they are steeper. Steeper climbs make race much more exciting than going endlessly uphill on gradient about 5-7%.

I think you're right. And it seems the various race organizers have realized it too.

/Jacob

 

[Libertine Seguros]

Now if only they'd realise that having more than just one steep climb at the end of the stage would make the races exciting for longer, we'd be cooking on gas. Instead, half the time we get a nice exciting last 5 minutes.

 

[Danskebjerge]

Re:

Now if only they'd realise that having more than just one steep climb at the end of the stage would make the races exciting for longer, we'd be cooking on gas. Instead, half the time we get a nice exciting last 5 minutes.

Or maybe they should simply make the stages shorter?

/Jacob

 

[Netserk]

I don't think __/ is much better than _____________/ though it would be somewhat cool with a Mt. Fuji kind of stage in a GT once.

 

[Escarabajo]

LOL. Arcalis is not a cat 2. Come on.

Cat 1 yes.

We have to consider the long grind to the top which is ignored in some cases. Like Arcalis. Remember in this forum the long complain about the MTT of Galibier in 2011. It only had a few steep kilometers towards the end. Everyone was convinced that it was not hard enough and that it would not have made any differences. The problem with it, and I have said it few times in the past, is the long grind to the top. The length kills the riders, not as much as the steepness. Both are killers but the length is always the bigger factor like Danskebjerge said in his article. The location of the climb is also a factor, but all organizations are well aware of that.

Some awfully steep climbs have created less differences than the longer climbs which look more like a cat 1 than a HC. PSM being another example of the length being the killer.

I did some basic energy requirements calculations when I first started looking into the watts of cyclists few years back, and always the length of the climbs ended up being the king. Some monster climbs like Ventoux, Madeleine, Stelvio, Finestre, Sierra Nevada, Tourmalet, Sestriere, Galibier being more energy consumers than the steeper counterparts.

 

[Libertine Seguros]

PSM is your example? Really? That's one of the steeper climbs in the Tour-ready Pyrenees. After the false flat start, the next 10km average over 8%. You may as well have said Hautacam. Arcalis is a cat.1 for sure, no HC there. Doesn't matter how long it is. Sure, a less steep climb over a long distance can merit the HC (consider, say, Grand Saint-Bernard) but even then, unless they're particularly long, particularly high altitude or include some steeper sections, they can be considered just cat.1. Petit-Saint-Bernard and Port d'Envalira are both 2000m+ passes that are long, but are gradual and consistent and have never been considered HC as a result. Arcalis is not sufficiently long that its length makes it a candidate for HC imo. Even then, a lot of the longest climbs with HC status are very inconsistent ones like Val Thorens or Croix-de-Fer where there's an eternity of steep ramps and flatter sections that mean the raw stats of the climb belie how difficult it really is.

 

[Escarabajo]

You get the idea Libertine. After that it leveled off and Froome did most of the damage there once everyone was cooked. That's just a case among others. On an average like the OP said it is the length that ruled the category.

 

[Danskebjerge]

One of the category parameters is not derived from the properties of the climb itself, but from its placing on the route.

We saw a good example of that yesterday. Stage 10 had a category 3 climb a few kilometers from the finish. It was 1.8 kilometer long and ascended at 6.6%. That's actually close to being an average category 4 (four!) climb, but often these supposedly decisive climbs are rated higher than the data would normally indicate.

Another example: Stage 17. Col de la Forclaz, starting at kilometer 154, is a 13 kilometre-long climb at 7.9% - rated as a category 1 climb. Shortly after, at kilometer 174, the riders will hit the Finhaut-Emosson which is 10.4 kilometres long and has a grade of 8.4%. This one is rated as an HC!

Surely, the objective difficulty of Finhaut-Emosson is not one level higher than Col de la Forclaz. It even has less vertical meters (around 150 less). It does, however, have a 12% steep section right up there to the finish line, and that may be decisive for its categorization. Besides, the organizers rarely have two HC climbs in a row - that in itself might play a role here.

 

[Danskebjerge]

Another example: Stage 17. Col de la Forclaz, starting at kilometer 154, is a 13 kilometre-long climb at 7.9% - rated as a category 1 climb.

Compare this to Arcalis, finishing climb of stage 9: 10.1 kilometre-long climb at 7.2%. So you have a category 1 climb that is both longer and steeper than a category HC climb...

Again, the position of the mountain plays a significant role here, but even so it almost seems silly.

 

[Red Rick]

One of the category parameters is not derived from the properties of the climb itself, but from its placing on the route.

We saw a good example of that yesterday. Stage 10 had a category 3 climb a few kilometers from the finish. It was 1.8 kilometer long and ascended at 6.6%. That's actually close to being an average category 4 (four!) climb, but often these supposedly decisive climbs are rated higher than the data would normally indicate.

Another example: Stage 17. Col de la Forclaz, starting at kilometer 154, is a 13 kilometre-long climb at 7.9% - rated as a category 1 climb. Shortly after, at kilometer 174, the riders will hit the Finhaut-Emosson which is 10.4 kilometres long and has a grade of 8.4%. This one is rated as an HC!

Surely, the objective difficulty of Finhaut-Emosson is not one level higher than Col de la Forclaz. It even has less vertical meters (around 150 less). It does, however, have a 12% steep section right up there to the finish line, and that may be decisive for its categorization. Besides, the organizers rarely have two HC climbs in a row - that in itself might play a role here.

I really dislike this, they already give double points for finishing climbs, so you have Arcalis getting 5x the amount of points that other, harder climbs get. You want people to fight for the KoM, not the GC winner winning it by chance because he happens to win the 2 HC finishes. That way it adds nothing to the race

 

[vedrafjord]

Re:

Côte des Chevrères is the most ridiculous cat.1 in recent years. That wouldn't even be cat.1 in the Vuelta, and they give cat.1 out for climbs like Cordal and Peña Cabarga which are definite cat.2s. How the hell that's cat.1 and Croix-Neuve is cat.2 I will never get.

Hell, Planche des Belles Filles should be cat.2. That's the same length and easier gradient than Peña Cabarga and Urkiola.

Incidentally Côte des Chevrères, PdBF and Peña Cabarga are all cat 2 on Strava, so their algorithm agrees with you.