Hyrox Finish Time
Predictor
Enter a past Hyrox time and this predictor returns your likely next finish time with an 80% range, usually within about 11.9% either side. No past race? Use your 5K time plus optional SkiErg, row and wall ball benchmarks for a split-by-split estimate.
Last updated: 2026-10-25 · Free, runs in your browser, nothing uploaded or stored
Runs in your browser. Nothing you enter is uploaded or stored.
Keep going
Next steps
Pacing Calculator
Turn a target finish time into split times for every station, run and roxzone.
Open toolWeakest Station Finder
Enter your splits and see which stations cost you the most time against athletes with your finish time.
Open toolTraining Plan
Weeks to race, level and goal in. A week-by-week plan with runs, station work and simulations out.
Open toolMethodology in one line: coefficients are fitted at build time to 367,515 Open and Pro finishes from 378 races (56,200 repeat-race pairs). This is a statistical estimate from public results, not an official Hyrox figure, and it cannot know your training, health or race-day conditions. See how we estimate below.
How accurate are Hyrox time predictions?
The past-time method is typically off by about 4.8% (median error) when tested on athletes the model never saw, and its 80% range holds the real next result about 8 times in 10. Here is the evidence by division.
| Division | Repeat pairs | Median time | Typical change at median | Median error (held-out) | 80% range |
|---|---|---|---|---|---|
| Open Men | 32,354 | 1:26:13 | -3.6% | 5.7% | -11% to +13% |
| Open Women | 14,318 | 1:34:56 | -2.8% | 5.0% | -9% to +11% |
| Pro Men | 6,510 | 1:19:29 | -2.8% | 4.7% | -9% to +10% |
| Pro Women | 3,018 | 1:25:27 | -1.9% | 4.0% | -8% to +9% |
Held-out means the model was fitted on half the athletes and scored on the other half. The 80% range is the 10th to 90th percentile of real next-race results relative to the model's central estimate.
How we estimate your Hyrox finish time
We use two methods that both come from the same race data: 378 per-race result files, covering 367,515 Open and Pro singles finishes, each with a total, running and workout time. Everything is calculated when the site is built, and the coefficients are open on this page.
Method A: from a past Hyrox time
For 56,200 athletes who raced the same division in two different years, we fit ln(next time) = a + b × ln(past time) separately for Open Men, Open Women, Pro Men and Pro Women. The slope b is below 1 (between 0.79 and 0.88), which is regression to the mean: unusually slow races bounce back and unusually fast ones rarely repeat. The 80% range is the 10th and 90th percentile of the real residuals, so it can be asymmetric.
Method B: from a 5K and station benchmarks
- Running. Total run time = 8 × your 5K pace per km × a pace-fade factor (1.08, 1.15 or 1.25). This factor is an assumption, not measured, because the results data has no 5K times.
- Stations. From the data,
ln(station time) = a + b × ln(run time)across 89,202 Open Men and similar counts for the other divisions (correlation about 0.73). If you add benchmarks, we compare your race-adjusted time (fresh time × 1.12) with your division's median split and blend that estimate with the run-based one, weighting 35%, 50% or 60% for one, two or three benchmarks. Those weights are judgement, not fitted. - Roxzone. Median transition time as a share of running plus stations in your division (about 8.6% for Open Men).
- Uncertainty. The empirical spread of total time given run time (about 7.4% for Open Men) combined with the 5.5% uncertainty in the 5K-to-Hyrox pace factor.
Splits and percentile
For Method A, run and roxzone are predicted from your total with the same data-fitted relationships, and the remaining work time is spread across stations in proportion to your division's median station splits. The percentile compares your predicted time with the full distribution of finish times in your division.
What it cannot know
Your training since the last race, injuries, nutrition, pacing, venue and weather. Athletes who return for a second race are also more committed than average, so first-timers may improve more than the table suggests. Doubles and Relay are not modelled.
How is a Hyrox race split between running, stations and roxzone?
At a median finish, running takes about 48% of an Open Men race and 49% of an Open Women race, stations most of the rest, and roxzone about 8%. Faster divisions spend less of the race running.
| Division | Running | Stations | Roxzone | Avg km pace |
|---|---|---|---|---|
| Open Men | 48% | 44% | 8% | 5:10 |
| Open Women | 49% | 43% | 8% | 5:48 |
| Pro Men | 46% | 47% | 7% | 4:34 |
| Pro Women | 47% | 46% | 7% | 4:58 |
Shares are for a median finish in each division, from the fitted run-versus-total relationship. For the exact station medians see the pacing calculator.
FAQ
Frequently Asked Questions
How do I predict my Hyrox finish time?
Enter a previous Hyrox finish time and your division, and the predictor applies the typical race-to-race change seen in 56,200 repeat results. If you have no race time, use a recent 5K time, optionally with 1km SkiErg, 1km row and 100 wall ball benchmarks, to build the estimate from running pace and station strength.
How accurate is the Hyrox finish time predictor?
On athletes held out of the fit, the past-time method was off by a median of about 4.8% (roughly 4 to 5 minutes at a 1:20 finish), and the 80% range contained the real next result about 8 times in 10. The 5K method is less certain because 5K time does not capture station strength, so its range is wider.
Can I predict my Hyrox time from my 5K time?
Yes, roughly. The tool multiplies your 5K pace by a pace-fade factor (1.08 to 1.25) to estimate the 8 km of Hyrox running, then estimates station time from the run-to-station relationship in 32,354-plus athletes' real splits. The 5K-to-Hyrox pace factor is an assumption, not measured, so treat that result as a wider estimate.
How much faster do people get between Hyrox races?
A typical athlete at the median Open Men time improves by about 3.6% at their next race, and by about 2.8% for Open Women. Slower first-timers improve the most because the model regresses toward the middle of the field, while athletes already near the front improve the least.
What share of a Hyrox race is running?
Running is about 48% of a median Open Men race, the 8 stations about 44%, and roxzone transitions about 8%. For Open Women the split is roughly 49% running, 43% stations and 8% roxzone.
Why does the predictor show a range instead of one time?
Because a single number would be false precision. Weather, venue, pacing, a bad sled lane or a good wave all move real results, and two athletes with the same past time will not repeat it exactly. The range is the 10th to 90th percentile of what similar athletes actually did, so about 8 in 10 results should land inside it.
Which station benchmarks should I enter?
Use a fresh 1km SkiErg time, a fresh 1km row time and the time for 100 wall balls at your race weight. These are the three stations easiest to test in training. We treat them as about 12% faster than your race-day splits, because stations in a race come after hard running, and compare them with the median split in your division.
Does the predictor work for Doubles or Relay?
No. It is fitted on Open and Pro singles results only (Men and Women). Doubles and Relay times depend on who does which part of each station, so they are not comparable. Use the Hyrox Doubles Split Planner for how to divide the stations between two partners.
Does the venue or course change my predicted time?
Yes, but the predictor does not know your venue. Course layout, floor surface and heat all shift times by a few percent, and that venue noise is part of the spread inside the range. If you race a notoriously fast or slow venue, expect to land toward the matching end of the range.
How is the percentile calculated?
We rank your predicted time against every finisher in the same division across 378 races (195,974 Open Men finishes alone), using 1% steps of the real finish-time distribution. A result of 60% means your predicted time beats about 60% of finishers in that division.
Is my data stored or uploaded?
No. The prediction runs entirely in your browser. Nothing you type is uploaded, stored or tied to an account, and there is no signup. The model coefficients are calculated when the site is built, so the page never sends your times anywhere.
What is a realistic goal after predicting my time?
Aim for the fast end of your range, not beyond it. If the prediction puts you at 1:24:00 with a range of 1:16 to 1:32, a 1:20 goal is a solid target with good training; 1:10 is outside what athletes with your history normally do. Build the plan for your target in the pacing calculator.
Keep exploring
More free Hyrox tools
Every tool runs in your browser with no signup, built on the same race data. See all tools.