Revised UTMB scores

I recently heard on the Distances+ podcast that the UTMB had revised its UTMB Index performance scoring system, which is something that I had heard about in other podcast episodes earlier this year.

The podcast mentioned that the difference between the old and new scores was around 2–3%, which is true, for instance, of French elite trail runner Ugo Ferrari, who got bumped from 795 to 812. (I expected the revision to go upwards due to previous podcast discussions in which elite runners complained of underrated races.)

Having just checked my own scores, the revision was much more massive — my overall score just got bumped from 618 last November to 667, which is almost a +8% change!

The change is primarily driven by my 20K score, which got bumped from 614 to 668 (+8.7%). Part of that change is itself driven by my performance at the Trail de la Colline 15K, which is my highest score ever at 689.

In contrast, my 50K score got bumped from 603 to 623 (+3%), with both my Cortina (Lavaredo 50K) scores remaining squarely under 600. The reason why my 50K score is still higher than 600 is because of my 2024 performance on the NTMF 42K, which is my second-best score ever at 653.

The revised score produces a progression ‘curve’ that makes more sense than the older one, which I unfortunately did not save:

The updated UTMB profile page does not indicate which position I now occupy in the global distribution of the UTMB Index, which was top 3% on 20K and top 5% on 50K prior to the change. My running stats currently show these percentiles with a note.

Resulting updates

I have updated my running stats page with the new scores. The revisions go back five years, which means that all of my race results are affected. All my results got bumped upwards, except the MaxiCross 27K 2022 one, which was apparently downgraded from 555 to 541, and the Ut4M 20 Vercors 2023 one, which went from 615 to 604.

The new scores now indicate that my very first performance above 600 UTMB points was on the very flat Ecotrail Paris 30K in 2022. My first equivalent performance on a mountain trail race is still the Ut4M 20 Vercors 2023. That makes sense to me.

The change means that my older blog posts mentioning UTMB scores are now outdated. Two of my upcoming posts, on the Trail des Pyramides Noires 23K and on the Oisans Trail Tour 39K, will also feature references to outdated scores.

Something else that will have to change is how I use my UTMB index to predict my finish time and placement. I used to look at the first 600-point runner in past races to do that, but with the revised score, I guess that I should instead look at runners closer to 620-630 on 20K, and even higher than that on 10-15K. The 600-point reference point remains an ambitious goal for me on 50K.

Methodological details

Below is an extract of the official UTMB statement on the revised index, in French. The changes cover gradients, altitude, and (perhaps most significantly in my case) field competitiveness:

le modèle a (…) été enrichi avec de nouveaux paramètres permettant de mieux refléter la diversité des profils de course. L’intégration des pentes, de l’altitude maximale ou encore de la configuration de la compétition en tête de course permet de valoriser plus justement les courses dites roulantes, les formats de haute montagne ou les performances réalisées dans des environnements très compétitifs. 

The same information was posted in English on LinkedIn, with additional details, including the mention of the global 2% adjustment, and a link to a methodological note, as a booklet (in English, on LinkedIn) and as a more detailed Medium blog post (in French or in English).

The UTMB Group Medium blog also has a very detailed case study about scoring the CCC 2024 (in French or in English), with screenshots of its internal interface where a lot of the building blocks of the UTMB index appear:

For some reason, the case study is not clearly linked to in the other posts, which is a shame, since it clearly is the most interesting of all publications mentioned above. I ended up posting a comment on that blog post to share a few (and probably wrong) thoughts.

* * *

I had a short email exchange with Colin Fay a few days after writing this post, after accidentally discovering that he recently scraped the entire list of UTMB-indexed trail runners. Colin might publish something on his blog about that at some point.

The Dans la tête d'un coureur podcast has an episode on the new index, but the podcast host did not do a good job at guiding the conversation, and the episode ended up being rather uninteresting.