Comprehensive Climbers Ranking

[Peyresourde]

Indurain is easily a Top 5 climber in history (I think his Tour weight was around 78kg). Even though it has to be said that his style of just setting a hard pace (a bit like Evenepoel) inflates his watts slightly. He also sometimes liked to attack over the top of climbs just like Evenpoel.

His '95 Tour, without having a 100+ performance, may be the best in terms of climbing consistency at a high level ever. He basically pumped out 95+ performances every mountain stage and easily matched/surpassed Riis and the ONCE squad. Riis in that Tour already had the same general shape as in '96, he just had a bad day on La Plagne.

 

[Pozzovivo]

@Pozzovivo

Sormano '24: Altitude -3, Stage Hardness +4, Approach/Clear Start -2, Follow-up +3, Regularity +2 --> +4 Adjustment
Ganda '25: Altitude -2, Stage Hardness +3, Approach -3, Follow-up +2 --> 0 Adjustment

My model "does not like" Ganda because of the easy approach (descent followed by false flat downhill is the best/easiest possible approach for me).
Sormano also gets a +2 for regularity because there is 1km of flat (or even slight downhill) in the middle of the climb.

All these adjustments are of course debatable, but they fit quite well also with the observed performance of the next riders (Evenepoel etc.).

Thank you for taking the time to explain, @Peyresourde.
I understand that some degree of subjectivity is unavoidable, and I am not trying to challenge your judgement(or model). I would howver like to know more about the methodological approach behind it.

Importantly, how did you develop the model and the categories itself? Was there an exploratory and confirmatory process, for example by splitting the data?

Did you test alternative model specifications with different weightings on a subset of climbs, and then evaluate how well they predicted performance in another dataset with the same riders? In other words, did you arrive at the current structure because it provided the best fit across samples?

I ask because I want to be cautious about the risk of models being shaped, even unintentionally, by prior expectations or narratives about particular riders or races, and I am interested in understanding how you worked to guard against that.

 

[Peyresourde]

Thank you for taking the time to explain, @Peyresourde.
I understand that some degree of subjectivity is unavoidable, and I am not trying to challenge your judgement(or model). I would howver like to know more about the methodological approach behind it.

Importantly, how did you develop the model and the categories itself? Was there an exploratory and confirmatory process, for example by splitting the data?

Did you test alternative model specifications with different weightings on a subset of climbs, and then evaluate how well they predicted performance in another dataset with the same riders? In other words, did you arrive at the current structure because it provided the best fit across samples?

I ask because I want to be cautious about the risk of models being shaped, even unintentionally, by prior expectations or narratives about particular riders or races, and I am interested in understanding how you worked to guard against that.

My starting point was that I was unhappy with the existing metrics.
LR uses ASLP, which only incoorporates altitude, but is IMO too extreme and also undervalues short climbs.

W2W adjusts for stage hardness and altitude. But it devalues TTs extremely. Especially in mixed TTs like PDBF 2020, it is even harder to perform on the climb than after a 180k road stage. W2W just takes the overall kilojoules and slams a fat minus adjustment on the effort.
In contrast, it seems to overvalue long stages. E.g. on the Lombardia climbs almost everyone does a PB which is not likely to be true.

My first premise is to rather adjust too little than too much. There are studies for the effect of altitude and heat, but I also take these with a grain of salt because some of the effects seem very high to me.

My next premise was that the approach to a climb is very important (even more important than how hard the overall stage has been). As for how to incorporate this? I did not use any scientific method, I basically just winged it and did what seemed plausible to me.

Certain prior expectations play a role in my process. E.g. Carapaz did a really high level performance in the Giro last year on the stage he won (on Pietra di Bismantova). It was a small climb and not very steep, no one else analyzed it. Then you catch yourself thinking: Carapaz normally never has high watts, this is a complete outlier --> There must have been a tailwind or maybe the segment was wrong etc.

And for a famously good climber, it may be the other way around: he won and gapped everyone by 30 seconds, yet the performance seems low --> Maybe the tarmac was worse than I thought, or there was a headwind?

I do this as a hobby, and I intentionally do a lot of the adjustments manually and by feel. Otherwise I would have no fun doing it. I also think this way may be more accurate than some automated system using a unified method (where would you even get accurate data for all parameters? Strava is an option, but it did not exist back in the day and many riders don't post power).

So certainly, for some performances I input the data once, it seems good and I never look at it again. While for some others, I put a lot of effort in to get it right/plausible.

 

[Pozzovivo]

My starting point was that I was unhappy with the existing metrics.
LR uses ASLP, which only incoorporates altitude, but is IMO too extreme and also undervalues short climbs.

W2W adjusts for stage hardness and altitude. But it devalues TTs extremely. Especially in mixed TTs like PDBF 2020, it is even harder to perform on the climb than after a 180k road stage. W2W just takes the overall kilojoules and slams a fat minus adjustment on the effort.
In contrast, it seems to overvalue long stages. E.g. on the Lombardia climbs almost everyone does a PB which is not likely to be true.

My first premise is to rather adjust too little than too much. There are studies for the effect of altitude and heat, but I also take these with a grain of salt because some oft he effects seem very high to me.

My next premise was that the approach to a climb is very important (even more important than how hard the overall stage has been). As for how to incorporate this? I did not use any scientific method, I basically just winged it and did what seemed plausible to me.

Certain prior expectations play a role in my process. E.g. Carapaz did a really high level performance in the Giro last year on the stage he won (on Pietra di Bismantova). It was a small climb and not very steep, no one else analyzed it. Then you catch yourself thinking: Carapaz normally never has high watts, this is a complete outlier --> There must have been a tailwind or maybe the segment was wrong etc.

And for a famously good climber, it may be the other way around: he won and gapped everone by 30 seconds, yet the performance seems low --> Maybe the tarmac was worse than I thought, or there was a headwind?

I do this as a hobby, and I intentionally do a lot of the adjustments manually and by feel. Otherwise I would have no fun doing it. I also think this way may be more accurate than some automated system using a unified method (where would you even get accurate data for all parameters? Strava is an option, but it did not exist back in the day and many rider don't post power).

So certainly, for some performances I input the data once, it seems good and I never look at it again. While for some others, I put a lot of effort in to get it right/plausible.

Great reply, very reasonable and credit to you. Still I do think that doing a path analysis with a "split data" approach could be of high value to find the best model fit. Maybe racing data is just too noisy idk.

I understand that when considering the performance on a climb, the lead in to the climb is very very improtant. So, for instance, Superbagneres in last years TDF should be massively upgraded in your view compared to other sources? The official segment starts after doing like 1,5km of 9% for some reason. Is this Felix Gall's best perfomance?

Anyway, looking forward to seeing the next lists you plan on posting, and how it will prove what we knew in our hearts - that Pozzovivo was in fact the best climber of all time Any standout performances from him in your dataset btw?

 

[Blobs]

Thanks for that informative presentation, Peyresourde! It's amazing how those climbing times of Pantani still stand. It lasted twenty years until those times were matched, in a completely different era, with completely different material.

Thirty years actually

 

[Peyresourde]

Great reply, very reasonable and credit to you. Still I do think that doing a path analysis with a "split data" approach could be of high value to find the best model fit. Maybe racing data is just too noisy idk.

I understand that when considering the performance on a climb, the lead in to the climb is very very improtant. So, for instance, Superbagneres in last years TDF should be massively upgraded in your view compared to other sources? The official segment starts after doing like 1,5km of 9% for some reason. Is this Felix Gall's best perfomance?

Anyway, looking forward to seeing the next lists you plan on posting, and how it will prove what we knew in our hearts - that Pozzovivo was in fact the best climber of all time Any standout performances from him in your dataset btw?

I plan the start posting the next list later today.

And yes, certain aspects of analysis can still be improved. it is by no means a completely finished product and I adjust things from time to time. The more often a certain climb is used, the more feel and context you also get.

One interesting climb is Oude Kwaremont, because it is raced twice in the same race (actually 3 times total) and Pogacar more or less goes full gas on both. In 2025 he went much faster the first time and still had 4 guys on his wheel. 40 km later he was slower but dropped everyone. So you could argue the correct adjustment would be to make both performances equal or even make the second one better.

But then I would have to adjust the second effort by an insane amount like +30 or even +40 points (basically equivalent to 1 W/kg!!!)

Superbagneres is indeed Gall's best performance with 89 (+7). For historic reasons (to be able tp compare with 1986 and 1989) I take the whole 42 minute climb from Luchon. If I just took some segment starting higher up, the Index would probably be the same, but with much higher adjustment. E.g.: 89 (+15)


For Pozzovivo: In my dataset I tried to include every effort of 80 or higher. For other climbs with lower performances, I often just take the winner/highest performer. That is to say that I have only 3 entries for Pozzovivo im my dataset and his best effort is probably not even in there.
The 3 I have:

72 (-1) on Lago Laceno/Mollela (Giro 2012)
72 (-2) on Zoncolan (Giro 2018)
69 (+12) on Fedaia (Giro 2008)

 

[Netserk]

do you have punta veleno?

 

[Peyresourde]

do you have punta veleno?

No. But there seems to be video on Youtube. So I might do it when I have some time.

 

[Peyresourde]

@Netserk thanks for the tip. i did not know this climb.

I tried to calculate it just now, unfortunately the results seems unrealistically low. There must be a major mistake in my input (e.g. the ascent time I have must be 3+ minutes to slow). So I won't publish this result.

 

[Pozzovivo]

Can you share your thoughts on the Tour of Norway performance by RE and JV that received so much attention in 2022, @Peyresourde?

In my view, there was a strong tailwind that day, although opinions differ. It certainly stood out as a personal best at the time for several riders and still appears to be a personal best for some, if we rely on the w2w and LR records.

 

[hayneplane]

This is a great piece of work that adds a lot of insight to the last 35 years of pro cycling.

Do you have any data to what the best performances ever are on climbs that were not the last classified climb of the day eg Tourmalet 2023 where the race was blitzed to pieces or Mortirolo in the giro where it was followed by a climb of Aprica.

 

[Peyresourde]

Now I will start posting my second list (Top 50 climbers). This should be of higher interest also to anyone who is not a fan of Pogacar or Pantani.

*An important note is that my list does not actually contain the real top 50, there are some qualifying criteria:
1) A rider must have at least 10 legit entries in my database AND
2) Must have won either the Tour OR won several other Grand Tours OR has caught my eye (yes, this is subjective but it is my list)

Notable Exclusions:
- Everyone before the 80s and Lucho Herrera (best climber of the 80s) because I do not have 10 good entries for them
- Roche and Pereiro/Landis for the same reason
- Some modern riders that would easily make the list (Lipowitz, Aysuo, Pidcock etc.). If they win some more I will include them in the future

I also do not claim to have the actual top 10 performance of every rider. Especially before 1986 I only have some famous climbs like Alpe d'Huez, Pla d'adet, Puy de Dome and Mont Ventoux. So I cut riders like Hinault and Fignon some slack in the rankings as I am probably missing some of their best performances.


Methodology: I use the 10 best performance Indices of every rider and calculate a weighted average, where the best performance counts slightly more then the next best and so on. I do this because I think the peak performances are more telling. Some famous climbers like Andy Schleck probably do not even have 10 performances in their whole career where they went all out.

Keep in mind that the specific numbers and the gaps between riders actually have a meaning. If Rider A has an Index of 80 and Rider B has 92, then the gap is 12 points = 0.3 W/kg.


I have divided the riders into 8 Tiers, we start here with Tier 8 ('Better buy an E-Bike') and Tier 7 ('No Watts')

Tier 7 (70-75)
Cadel Evans | 73.9 | PB: 80 (-2): 7.06 W/kg for 12:25 on Arrate Usartza (Basque 2009)
Bradley Wiggins | 71.5 | PB: 78 (-3): 5.91 W/kg for 59:10 on Ventoux (Tour 2009)
Laurent Fignon | 69.4 | PB: 79 (-5): 6.54 W/kg for 26:51 on Ruchere ITT (Tour 1984)
Andy Hampsten | 69.3 | PB: 81 (0): 6.09 W/kg for 46:16 on Tourmalet East (Tour 1993)

Tier 8 (<70)
Greg Lemond | 63.2 | PB: 68 (+2): 5.76 W/kg for 43:09 on Alpe d'Huez (Tour 1989)
Bernard Hinault | 62.3 | PB: Probably 80+ on Avoriaz ITT (1979 Tour)

The most shocking entry is Greg Lemond. A rider who has won the Tour 3 times and he does not have a single good climbing performance ?! (I may be missing some of his efforts from '84 and '85). Overall he seems to be the worst climber of all Tour winners from the 80s. The only thing he has going for him is good consistency in the Tours he won.

Bernard Hinault would in reality be higher then Lemond, he has some solid climbing performances especially in ITTs, but I am missing more of his performances as he was active earlier.

 

[Peyresourde]

Can you share your thoughts on the Tour of Norway performance by RE and JV that received so much attention in 2022, @Peyresourde?

In my view, there was a strong tailwind that day, although opinions differ. It certainly stood out as a personal best at the time for several riders and still appears to be a personal best for some, if we rely on the w2w and LR records.

Remco Evenepoel | 80 (-5): 6.50 W/kg for 30:23 on Gaustatoppen (Norway 2022 Stage 3)

It was not a world beating performance, but it was the first good climbing performance of Evenpoel (I think he did not have a single performance >70 prior!). In that context it is understandable that people were sceptical. There was a tailwind, but on these kind of climbs without many spectators and vegetation, it is really hard to tell how strong and I also did not study the video that closely.

But both him and Jay Vine improved on that performance later that year in the Vuelta. this is in line with my prior argument that riders can normally repeat a performance if they have shown it once. I think the riders further back (Plapp etc.) did not do that great (1 minute gap is a lot).

This is a great piece of work that adds a lot of insight to the last 35 years of pro cycling.

Do you have any data to what the best performances ever are on climbs that were not the last classified climb of the day eg Tourmalet 2023 where the race was blitzed to pieces or Mortirolo in the giro where it was followed by a climb of Aprica.

1994 Aprica is one of the most legendary stages of all time. Already on the Stelvio, Vona did a 'world-record' performance for the time, then Pantani topped it on the Mortirolo.

I will probably do an extra post on this stage in the future and also on the Tourmalet - Cauterets Cambasque combo which was used both in '95 and '23.

 

[Red Rick]

Remco Evenepoel | 80 (-5): 6.50 W/kg for 30:23 on Gaustatoppen (Norway 2022 Stage 3)

It was not a world beating performance, but it was the first good climbing performance of Evenpoel (I think he did not have a single performance >70 prior!). In that context it is understandable that people were sceptical. There was a tailwind, but on these kind of climbs without many spectators and vegetation, it is really hard to tell how strong and I also did not study the video that closely.

But both him and Jay Vine improved on that performance later that year in the Vuelta. this is in line with my prior argument that riders can normally repeat a performance if they have shown it once. I think the riders further back (Plapp etc.) did not do that great (1 minute gap is a lot).

1994 Aprica is one of the most legendary stages of all time. Already on the Stelvio, Vona did a 'world-record' performance for the time, then Pantani topped it on the Mortirolo.

I will probably do an extra post on this stage in the future and also on the Tourmalet - Cauterets Cambasque combo which was used both in '95 and '23.

The Bormio-Mazzo section is one of the big wattsavers in cycling.

 

[mikks]

How would your model rate Contador vs Rasmussen climbing style in 2007? My memory is a bit hazy, but wasn’t it more of a constant stop-and-go effort — sprinting, pausing,sprinting, pausing and so on and on.


I would guess that this kind of pacing wouldn’t produce the fastest climbing times.. Still, from what I remember, visually it was incredibly impressive and looked like a brutally hard effort.

 

[Red Rick]

How would your model rate Contador vs Rasmussen climbing style in 2007? My memory is a bit hazy, but wasn’t it more of a constant stop-and-go effort — sprinting, pausing,sprinting, pausing and so on and on.


I would guess that this kind of pacing wouldn’t produce the fastest climbing times.. Still, from what I remember, visually it was incredibly impressive and looked like a brutally hard effort.

You cannot model that. It was a slow pace until the final 2km or so, then it was a few crazy attacks with intervals.

 

[Peyresourde]

@mikks The model assumes (generally) assumes linear speed, so I cannot model this. Overall, Rasmussen was the slightly better climber that Tour. But the level was not that high (a bit better than the 2010s), even if I consider the accelerations. I have thought about giving some bonus points manually for these kind of climbs, but have not really done it (much).

The next tier includes every Tour winner from 2013-2019: Tier 6 ('Sky Era/2010s')

Tier 6 (75-80)
Andy Schleck | 79.7 | PB: 91 (+3): 6.20 W/kg for 50:05 on Tourmalet West (Tour 2010)
Egan Bernal | 79.2 | PB: 84 (-9): 6.86 W/kg for 24:26 on Villars sur Ollon ITT (Romandie 2018)
Claudio Chiappucci | 79.0 | PB: 85 (+8): 5.98 W/kg for 47:33 on Tourmalet East (Tour 1995)
Chris Froome | 78.6 | PB: 81 (+7): 6.58 W/kg for 17:58 on Farrapona (Vuelta 2014)
Thibaut Pinot | 78.6 | PB: 81 (-2): 6.52 W/kg for 27:07 on Prat d'Albis (Tour 2019)
Geraint Thomas | 77.4 | PB: 81 (-1): 5.99 W/kg for 54:54 on Bondone (Giro 2023)
Vincenzo Nibali | 77.0 | PB: 79 (-3): 6.16 W/kg for 42:44 on Angliru (Vuelta 2013)
Tom Dumoulin | 74.9 | PB: 81 (-1): 6.80 W/kg for 17:36 on Oropa (Giro 2017)
Pedro Delgado | 74.7 | PB: 86 (-5): 6.15 W/kg for 59:28 on Ventoux ITT (Tour 1987)

Schleck and Bernal could both be 1-2 Tiers higher, their problem was not trying hard often enough and the accident respectively, though Bernal still has some hope.

After analyzing all the climbing performances of the 2010s, I really respect Froome and Nibali more than before. I always thought they were not that good (performance-wise).
Now I just think it is impressive, to have such long and succesful careers without any outlier-performances.

In contrast to Pogacar or even Vingegaard today, they were barely better or even worse than their rivals in terms of peak-climbing performances. Yet they still managed to win year after year.
I would say Froome's peak was from 2013-2016, so it is even more fascinating that he won the Tour-Vuelta double in 2017 without a single good climbing performance.

 

[houtdffan]

Froome won the 2017 Vuelta due to a Contador illness in the Pyrenees and him riding Sierra Nevada like a total idiot, trying to attack from the very bottom of the climb, more than 30km from the finish

 

[Netserk]

im most surprised by angliru being nibalis best performance. he seemed a notch better in the tour the following year

 

[Peyresourde]

Nibali has a PB of 79 and an average of 77. This might be the closest of all riders. So he has a lot of performances on a similar level (Verbier 2009, Ventoux 2009, Hautacam 2014). But I would say in general that his personal peak was quite short only from 2013-2014. He also had surprisingly bad form in several GTs over his career.

If you additionally consider that he attacked Horner several times (irregular pacing), I think Angliru is his best.

 

[Krzysztof_O]

im most surprised by angliru being nibalis best performance. he seemed a notch better in the tour the following year

Grandpa Horner was out of this world then. There's still a good chance Rogla will win the Tour in 5-7 years when he finally peaks (with Pogi growing his belly on a yacht in Monaco).