REF-901 · Fernández-Navarrete et al., 2026 · The Journal of Sports Medicine and Physical Fitness · Paper · PubMed

Where am I struggling? Detailed benchmarks for every workout and category in Hyrox© competition

B+moderate evidenceMeasured on recreational to elite athletesWritten before October 2026, being re-checked

To give coaches and athletes reference values for HYROX, the authors pulled 186,411 results from the official ranking database, from the format's inception up to the 2023/2024 season, and described every category (individual, doubles, relay) by the mean and the 10th to 90th percentiles of total time and of each segment.

The stated purpose is to let an athlete compare each run and station split with the field and so locate strengths and weaknesses for training design and race planning.

Participants

n=186,411 race results from the official HYROX ranking database (all categories, inception to the 2023/2024 season). The number of unique athletes, sex split and ages are not given in the abstract.

Protocol

Retrospective extraction of official race results.
Descriptive statistics per category and per segment: mean and percentiles from the 10th to the 90th.
No inferential testing is described in the abstract.

Key Results

Normative tables (mean, 10th-90th percentiles) for total time and for every workout segment, for individual, doubles and relay categories, built from 186,411 results.
The abstract reports no benchmark values, percentile cut-offs or comparisons between segments; the tables are in the full text.
Authors' conclusion: the tables allow the strengths and weaknesses of each section to be analysed, which could support training design and competition planning.

Practical Application

To "where am I losing time in HYROX?": this study built percentile tables for every run and station, per category, from 186,411 official results up to the 2023/2024 season, so a split can be placed against the field — a station sitting at a much lower percentile than the athlete's runs marks it as the relative weakness. The values themselves are in the paper's tables, not in the abstract, so quote no benchmark numbers from this entry.
Inference, not tested here: a segment that sits far below the athlete's own overall percentile is the cheaper place to gain time than one already in line with it; the study describes the distribution and does not test which gains are easiest.
Not from this paper; the base's own practice: read station weaknesses against the fact that running is most of the race (REF-510, REF-903) — a small percentile gain on the runs can outweigh a large one on a short station. How to build the week is in SYNTH-71.

Limitations

Descriptive only: no associations between segments and final time are tested in the abstract.
Data end at the 2023/2024 season; later rulebook, venue and field changes may shift the distributions (the field improved season on season, see REF-903).
Race results only: no physiology, training history or athlete characteristics.
Written from the abstract and record metadata only (full text not read); none of the benchmark values are reproduced here.
How this entry was checked and changed
Why this level

cross-sectional measurement study with a very large dataset (n=186411 race results); descriptive only, no associations tested