INFNIT™ UTMB World Series Dataset 2003-2024

The first publicly available individual-level dataset for the UTMB World Series capturing both finishers and DNFs across all Mont-Blanc races (UTMB®, CCC®, OCC, TDS®, MCC) and the Azores Triangle Adventure, 2003–2024. 79,587 runner records | 68,925 finishers | 10,662 DNFs | 54 races | 22 editions Unlike existing public UTMB datasets which contain finisher-only records or pre-aggregated race summaries, this dataset captures individual DNF status per runner — enabling survival analysis, heat stress modelling, and DNF prediction previously impossible with public data. Key findings: 2021 DNF rate: 23.4% — extreme heat conditions documented across multiple ultra-trail events that year (vs. 12–13% in normal years) DNF variance across editions: 12.7%–23.4% Women participation grew from ~10% (2003) to ~17% (2024) 40–44 age category is the largest cohort (17,234 athletes) 100+ nationalities represented Columns (16): year, race_uid, race_slug, race_title, race_date, race_category, dist_km, elev_m, overall_rank, finish_time, runner_id, nationality, gender, age_category, status (F/DNF), source GDPR: Name columns removed. runner_id serves as pseudonymous identifier. Methodology: Scraped from UTMB World Index (utmb.world/utmb-index) using Playwright (Chromium) on 2026-05-22. Each of 54 race-year combinations scraped independently with full pagination. Part of the INFNIT™ research programme — AI-driven ultramarathon performance platform. MA AI, University of Southampton. GitHub: https://github.com/alnakib80-cpu/UTMB-World-Dataset License: CC BY 4.0 — attribution required. Citation: Al-Nakib, F. (2026). INFNIT™ UTMB World Series Dataset 2003-2024. Zenodo. https://doi.org/10.5281/zenodo.20348998