Compromised running in HYROX: why runs slow after stations

Compromised running is running on legs that just did something else hard, and in HYROX it's every run after the first. Your runs slow after stations because the station leaves you with tired muscles and a high heart rate and lactate, and the next run starts from there. In 24 Pro men at HYROX Madrid, average running speed fell from 16.1 km/h in run 1 to 13.6 km/h in run 8¹.

That's a drop from about 3:43 per km to about 4:25 per km across one race, in Pro athletes. So it happens to the best racers too. The question is how much, and where.

What happens in your legs#

None of the HYROX studies so far isolate the mechanism. The closest review looks at running after resistance training and lists four reasons endurance performance drops: muscles recruit less well, your running mechanics change so each step costs more energy, soreness, and lower muscle glycogen². That review is about the hours and days after a gym session, not the 30 seconds between a sled and a run, so treat it as a likely explanation rather than a proven one.

The lab data on HYROX itself shows the stations, not the runs, drive the metabolic spikes. In Brandt's simulated race, peak blood lactate was higher after stations (8.5 mmol/L) than after runs (7.7), and the single highest values for heart rate, lactate and effort all came at the last station, wall balls³. Participants spent about 80% of the race above 90% of max heart rate. There isn't much recovering going on anywhere.

Where it falls apart: the run-by-run data#

Two published HYROX studies report pacing run by run, and they don't point at a single bad station.

In Brandt's simulation, the slowest runs were run 5, then runs 8 and 6³. Run 5 comes straight after the burpee broad jumps, run 6 after the row, and run 8 after the sandbag lunges⁴. In the Madrid Pro data, speed declined progressively from the first run to the last¹, which reads more like accumulating fatigue than one station breaking people.

Pro men running speed, HYROX Madrid 2025

	Value (km/h)
Run 1	16.13
Run 8	13.58
Roxzone	10.52

Mean speed, n = 24 Pro men. Source: Gutiérrez-Hellín et al., Int J Sports Physiol Perform 2026 (HYScience study).

My eight runs, lined up with the station before each#

Here are my runs from HYROX Bangkok 2026, from the timing splits⁵. Each run is measured from leaving the Roxzone to coming back into it, except run 1 (from the start) and run 8, which has no separate Roxzone split and runs all the way to the wall ball station.

Ken's eight runs, HYROX Bangkok 2026

	Value (min)
Run 1	4.58
Run 2	4
Run 3	3.92
Run 4	3.92
Run 5	4.08
Run 6	3.93
Run 7	3.92
Run 8	4.9

Run splits from the HYRESULT timing table (Men, Bangkok 2026). Run 8 includes the walk into the wall ball station. Source: hyresult.com, read 2026-10-22.

Runs 2 to 7 sit inside a 10-second window, 3:55 to 4:05. The sleds didn't cost me anything I could see on the clock: the runs straight after the push and the pull were two of my fastest. The slowest of the middle six was run 5, after the burpee broad jumps, which happens to be the same run that was slowest in Brandt's group³. One race and one small lab study isn't a pattern. It's just worth checking whether your own run after the burpees is your slowest middle run too.

Run 8 is the real outlier at 4:54, nearly a minute slower than the runs before it. Part of that is measurement: it includes the walk into the wall ball area, which every other run hands off to a Roxzone split. The rest is the lunges, plain fatigue, or both, and the timing can't separate them.

Measured run 8 against run 1, my fade was 6.9%. But run 1 was slow too, and against the middle runs (3:58 on average) run 8 was about 24% slower. For comparison, the Pro men's average speeds work out to about 3:43 per km on run 1 and 4:25 on run 8, a 19% slowdown¹. So my last run fell off harder than the Pro average, with the caveat that my run 8 includes the walk-in, while my middle runs barely moved.

Your own splits will tell you more than mine. Paste your eight runs, finish time and station total into the HYROX splits analyzer and it gives you your Roxzone time and how much you faded from run 1 to run 8.

The Roxzone is part of the same problem#

The slowest movement in a HYROX race isn't a run. In the Madrid Pro data, the Roxzone was the slowest part of the running component at 10.5 km/h, against 13.6 to 16.1 km/h on the runs¹. That's the walking and jogging between the run course and each station.

My timing table splits the Roxzone into a walk-in before each station and a walk-out after it⁵:

Those add up to 6:48, within a few seconds of my 6:45 Roxzone total (the timestamps are rounded to the second). A lot of that is venue layout: the distance from each station to the run course isn't the same, so walk-in and walk-out times aren't comparable station to station, or venue to venue. What you can compare is your own numbers race to race, at the same venue if you can.

What to do about it#

You can't avoid the slowdown, so the aim is to make it smaller and to stop it from scaring you into a pace you can't hold. Two things help.

First, don't start the first run like it's a 1 km test. The Pro data shows speed falling run by run even in very fit athletes¹. A run 1 that's much quicker than you can hold by run 6 just moves the cost later. The run pacing guide goes through how to set a target per run.

Second, practise running on tired legs, on purpose. Two session formats that do this:

Station-into-run repeats. Pick one station, do a race-sized chunk of it, then run 1 km at goal race pace straight away. Rest two minutes and repeat three to five times. Rotate the station across weeks so the run after the lunges or the burpees gets practised as much as the easy ones.

Here's a real one from my log, on 14 July: five rounds of 1 km run, 25 wall balls at 6 kg and 33 m of burpee broad jumps, with 1:45 rest between rounds. The first 1 km went in 4:54. A week later, four rounds of 1 km, 30 wall balls and 22 m of burpee broad jumps, first 1 km in 4:48.

Legs-then-run blocks. Do a heavy leg-dominant station (sled push, sandbag lunges at race weight) for one full race distance, then run 2 km steady. This is the lunges-into-run-8 problem in miniature, and the run 8 split above is why it's worth practising.

Doma's review makes a point that cuts both ways here. Because a hard lifting session can blunt the quality of an endurance session that follows it, the authors suggest adjusting order, recovery time between the two and intensity to limit that carry-over². That's the right default for most of your week: keep your quality runs away from your heaviest leg days. The brick sessions above are the deliberate exception, where running on tired legs is the point. Keep them as the exception.

To check it's working, run the same brick or race simulation every few weeks and watch two numbers: the gap between your first run and the rest, and the spread of the middle runs. If both shrink at the same effort, you're getting better at running compromised.

How often you need these isn't known, and nobody has compared them in a trial. One a week, inside the running volume you already do, is a reasonable start. Don't stack them on top of it. For how much running to do around them, see how much running HYROX needs, and for the sleds themselves, the sled push and pull guide. The rest of the series is on the HYROX hub. hybridAF is not affiliated with HYROX.