Guide
Where are you losing time? How to read your HYROX splits after a race
Split your result into runs, stations and Roxzone, work out the share of each part and compare run 1 with run 8. In the top 100 PRO men, running was 48.5% of finish time and the Roxzone 7.3% (Rappelt et al., 2026). The order of athletes shifted most on the sleds and the wall balls. In the example profile (today 1:19:17, goal 1:15:00) the gap is about 5% of finish time. Look for those minutes first in the parts where you lose the most relative to the rest.
In short
- Split your finish time into three parts and work out each share; the Roxzone is usually the smallest part.
- In the top 100 PRO men, running was 48.5% and the Roxzone 7.3% of finish time (Rappelt et al., 2026).
- The ranking shifted most on the sleds and the wall balls, and slower athletes were more sensitive to station time.
- In 11 recreational athletes, station times were not related to finish time (Brandt et al., 2025); the group was small.
- Compare run 1 with run 8. In 24 PRO men, run 8 was about 42 s per km slower (HYScience, 2026).
Where are your four minutes usually hiding?
In the runs and on the stations. The Roxzone is too small to take four minutes out of. In the top 100 PRO men in season 7, running was 48.5% of finish time and the Roxzone 7.3% (Rappelt et al., 2026). What is left, about 44%, is the stations. That last figure is our own calculation.
Take the Benchmrk example profile, not a real athlete. Goal 1:15:00, Open Men, threshold pace 4:25/km. If this profile raced today, the app predicts 1:19:17. So the gap is 4:17, about 5% of finish time (4:17 of 79:17). On race day the app expects 1:16:42, with a 70% chance of the goal time. Your result tells you where that is easiest to find.
How do you split your result into three parts?
Add up the eight runs, add up the eight stations and take the Roxzone separately. The official results list those times per part. Then work out each part as a percentage of your finish time.
A worked example using the Benchmrk engine example (Open Men, threshold 4:45/km). Suppose this was your result.
| Part | Time | Share |
|---|---|---|
| Runs | 42:24 | 53.9% |
| Stations | 30:52 | 39.2% |
| Roxzone | 5:25 | 6.9% |
| Finish time | 1:18:41 | 100% |
The runs average out at 5:18/km (42:24 divided by 8). The gap to 1:15:00 is 3:41.
How does your split compare with the research?
Use the research as a rough range, not a standard. The two studies differ a lot in group and design.
| Study | Group | Running | Stations | Roxzone |
|---|---|---|---|---|
| Rappelt et al. (2026) | Top 100 PRO men, season 7 | 48.5% (27:38) | about 44% (our calculation) | 7.3% (4:10) |
| Rappelt et al. (2026) | Top 100 PRO women, season 7 | 47.9% (30:17) | about 45% (our calculation) | 7.2% (4:34) |
| Brandt et al. (2025) | 11 recreational athletes, simulated Open race | about 60% (median 51.2 of 86.5 min) | 32.8 min median | not reported separately |
Rappelt looked at results from athletes who are much faster than you. The Roxzone layout differed per race. Brandt had 11 people in a gym, with the runs on a treadmill. In the worked example, the running share (53.9%) sits between the two groups and the Roxzone (6.9%) is similar to the top 100. If your Roxzone is clearly higher, your transitions are worth a look. How to win seconds there is in save time in the Roxzone.
Are you losing the time in the second half?
Put run 1 next to run 8, and runs 1 to 4 next to runs 5 to 8. Losing pace is normal, even at the top. In the HYScience study (Gutiérrez-Hellín et al., 2026), 24 PRO men ran a real race in Madrid, with an average finish time of 69:58. Run 1 averaged 3:43/km, run 8 4:25/km. That is about 42 seconds per kilometre, or 16% lower in speed.
Only PRO men in one race, and we only read the abstract. Use it as a reference for the pattern. Benchmrk sets a stricter bar for training: in benchmark 4, the pace decay test, run 8 stays within 12 seconds per kilometre of run 1. If your difference between run 1 and run 8 is much bigger, your gain probably lies in your start and in running under fatigue. Running under fatigue is a choice Benchmrk makes based on research. How to pace your first kilometre is in run pacing. The training for it is on compromised running.
Which stations do you look at first?
Start with the sleds and the wall balls. In the results Rappelt and colleagues (2026) analysed, 39,696 results in PRO and ELITE, the order of athletes changed most at those stations. On the running, the ranking shifted least. Slower athletes were also more sensitive to station time, especially on the strength stations.
Brandt and colleagues (2025) found something different. In their 11 recreational athletes, station times were not significantly related to finish time. VO2max was related to it (ρ = −0.71), as was endurance training volume (ρ = −0.68). Eleven people is a small group, and a correlation is not a cause. The two studies do not rule each other out. Possibly a station difference disappears in a small group but shows up in tens of thousands of results.
So compare your stations with your own previous race and with each other. If your sled push or wall balls time is clearly slower than the rest, that is where the gain is. See faster sled push and wall balls without stopping.
How do you split the gap across runs, stations and Roxzone?
Start with a fair, proportional split and then shift towards your weak point. If you take the same percentage off every part in the worked example, 1:15:00 looks like this.
| Part | Now | Proportional at 1:15:00 | Difference |
|---|---|---|---|
| Runs | 42:24 | 40:25 | −1:59 |
| Stations | 30:52 | 29:25 | −1:27 |
| Roxzone | 5:25 | 5:10 | −0:15 |
| Total | 1:18:41 | 1:15:00 | −3:41 |
The calculation: each part times 75:00 divided by 78:41, about 0.953. Running pace then goes from 5:18/km to about 5:03/km. If you already run the second half evenly, shift more of the gap to the stations. If you lose a lot of pace after run 4, more of the gain is in the running.
What does the step-by-step plan look like?
Do this within a week of your race, with your official result next to you.
| Step | What you do | Worked example |
|---|---|---|
| 1 | Add up runs, stations and Roxzone separately | 42:24, 30:52 and 5:25 |
| 2 | Work out the share of each part | 53.9%, 39.2% and 6.9% |
| 3 | Work out the gap to your goal time | 1:18:41 to 1:15:00 is 3:41 |
| 4 | Compare run 1 with run 8 | Fill in your own splits; HYScience reference is 42 s/km in PRO |
| 5 | Rank your stations, slowest first | Look at the sleds and the wall balls first |
| 6 | Split the gap proportionally and shift towards your weak point | Runs −1:59, stations −1:27, Roxzone −0:15 |
| 7 | Pick one or two parts for your next block | For example the second half of the runs and the sled push |
What this means for your finish time is in predict your HYROX time. A sim to test your new split is built with build a HYROX simulation.
How Benchmrk does it
The app predicts your time per run, per station and for the Roxzone, and gives your chance of hitting your goal time. So you see your gap in advance. Five benchmarks recalibrate that prediction. Strength for your weakest stations sits next to your running, and the plan adapts to what you log.
Frequently asked questions
How do you analyse your HYROX splits?
Add up your runs, stations and Roxzone separately and work out what share of your finish time each part is. Then look at the difference between run 1 and run 8 and at your slower stations, especially the sleds and the wall balls.
How much of a HYROX race is running?
In the top 100 PRO men in season 7 it was 48.5%, and in the women 47.9% (Rappelt et al., 2026). In a simulated race with 11 recreational athletes it was about 60% (Brandt et al., 2025).
How much time does the Roxzone cost?
In the top 100 PRO men an average of 4:10, about 7.3% of finish time. The Roxzone layout differed per race, so compare mainly with your own previous race.
At which stations do you lose the most places?
In the results Rappelt and colleagues (2026) analysed, the order of athletes shifted most on the sleds and the wall balls. It shifted least on the running.
How much slower can run 8 be than run 1?
There is no standard from research for that. Benchmrk uses as a training goal that run 8 stays within 12 seconds per kilometre of run 1 (benchmark 4). In 24 PRO men in a real race, run 8 was about 16% slower in speed, about 42 seconds per kilometre (HYScience, 2026).
Sources
- Rappelt et al. (2026), Longitudinal performance development in PRO and ELITE HYROX competitions across the first seven competitive seasons. Frontiers in Physiology 17:1847569 (accessed on 13 November 2026)
- Brandt et al. (2025), Acute physiological responses and performance determinants in Hyrox©. Frontiers in Physiology (accessed on 13 November 2026)
- Gutiérrez-Hellín et al. (2026), Relative Running Intensity and Pacing During an Official HYROX PRO MEN Competition (HYScience). International Journal of Sports Physiology and Performance – Abstract only; the full text is behind a paywall. (accessed on 13 November 2026)