HYROX mobility has no literature. There is no randomised trial, no prospective cohort and no biomechanical study of range and performance in this sport, because the sport is young and the research follows the money and the decades. Every confident article about it — including this one — is reasoning from the stations rather than citing anything.
That is stated first because it is the most useful sentence on the page. What follows is an honest reading of what the race asks of each joint, what the neighbouring evidence says about the interventions people reach for, and a plain account of what Limber does and does not have for it.
The short answer
- No research exists on mobility and HYROX. Everything below is reasoning from the demands.
- The two stations that most obviously depend on range are wall balls and sandbag lunges.
- The ankle is the joint under most load: it decides squat depth and the sled push position.
- Stretching has not been shown to prevent injury in any sport, including the ones with research.
- There is no HYROX tag anywhere in Limber's catalogue, and the sport tags it does carry are never read.
- The most useful thing you can do before a race is measure the ankle and warm up properly.
There is no HYROX literature
Stated plainly, because the rest of the page depends on it.
There is no trial of a mobility programme in HYROX athletes. There is no cohort study linking joint range to race times or to injury in this sport. There is no measured account of the joint angles the stations actually produce. Compare that with the deep squat, which has been recorded with motion capture, or the ankle, which has a reliability literature going back decades.
Grade: absent.
So the honest structure for a page like this is: reason from the movements, borrow evidence only where the movement is genuinely the same as one that has been studied, and label every step. Where this page says "reasoning", it means nobody has measured it. Where it names a study, that study exists and is about something else.
HYROX mobility, reasoned
The standard race is eight kilometres of running, cut into one-kilometre segments, with a workout station between each. Ski erg, sled push, sled pull, burpee broad jumps, rowing, farmers carry, sandbag lunges, wall balls.
Read that list as a set of joint demands and three things stand out.
Almost none of it needs a large range. Running, rowing, carrying and pushing all happen well inside what a normal joint offers. If range were the limiter in this sport, it would be the limiter in an unusual way.
Two stations are the exception. Wall balls ask for a full squat, repeated many times, with the arms finishing overhead — a range demand at both ends of the body, under fatigue, late in a race. Sandbag lunges ask for hip flexor length on the trailing leg and ankle dorsiflexion on the front one, hundreds of times, with a load on the shoulders.
Fatigue is the variable nobody measures. Everything above is about range at rest. What a joint offers at kilometre seven with a sandbag on your back is a different number, and no study on any of this has measured it.
The demands, station by station
Reasoning, throughout. None of this is measured in HYROX athletes.
Wall balls. A squat to depth with an overhead finish. Hemmerich and colleagues measured a full squat in the Journal of Orthopaedic Research in 2006 at roughly 95 degrees of hip flexion, 155 of knee flexion and about 38 degrees of ankle dorsiflexion — a genuine measurement, of a squat, not of this station. If a station is going to expose a range limit, it is this one.
Sandbag lunges. The trailing hip is asked for extension while the front ankle is asked for dorsiflexion, with a load high on the body that punishes a rounded upper back.
Sled push. A long stride into a forward lean, which is ankle dorsiflexion and hip extension under heavy load. Macrum and colleagues showed in the Journal of Sport Rehabilitation in 2012 that restricting the ankle changes squat mechanics immediately; the same logic applied to a sled is reasoning rather than a finding.
Ski erg and rowing. Repeated hip flexion with a spine that has to keep its shape, and on the ski erg a start with the arms overhead.
The running. Small range, enormous repetition — the same problem mobility for runners sets out, with the difference that here it is interrupted eight times by something heavy.
What the evidence next door says
Where the movement really is the same, the neighbouring evidence is worth having, and most of it is unhelpful to the usual advice.
Stretching does not prevent injury. Lauersen, Bertelsen and Andersen pooled the randomised injury-prevention trials in the British Journal of Sports Medicine in 2014: strength training reduced sports injuries to under a third and overuse injuries to about half, while stretching showed no significant protective effect. Pope and colleagues found no reduction in injury when 1,538 army recruits were randomised to stretch before exercise, in Medicine & Science in Sports & Exercise in 2000. Grade: reasonably solid.
Long holds before an effort cost a little. Simic, Sarabon and Markovic put it at about 5 per cent of strength, power and explosive performance in the Scandinavian Journal of Medicine & Science in Sports in 2013, and Behm and colleagues put the practical threshold at around 60 seconds per muscle group in 2016. Grade: solid. For a race that starts with a kilometre of running, that is worth avoiding.
A proper warm-up removes the cost. Blazevich and colleagues embedded static stretching inside a full dynamic warm-up in Medicine & Science in Sports & Exercise in 2018 and found no impairment of performance. Grade: solid. So the pre-race rule is move, then hold only briefly if at all, then move again.
Calf stretching buys a few degrees. Radford and colleagues' review in the British Journal of Sports Medicine in 2006 found static calf stretching produced a gain in dorsiflexion of a few degrees, of questionable clinical importance. Grade: thin. Stance and shoe drop change the demand faster than stretching changes the supply.
Screens do not predict. Moran and colleagues found movement-screen composites predict injury poorly in 2017, and Bahr argued in 2016 that this is a general property of screening. A reading tells you where you are, not what will happen to you.
No HYROX tag in the catalogue
Limber's 56-move catalogue tags moves by sport: running 34, lifting 46, cycling 27, climbing 10, golf 8. There is no HYROX tag at all, and if there were, it would not matter — nothing in the build reads any sport tag. They are parsed into the engine's model and never consulted during selection.
There is no HYROX programme. Ten tracks are drawn — Full-Body Foundations, Posture and Desk-Undo, Hips Unlocked, Shoulders and Overhead, Lower Back Care, Hamstrings, Neck and Upper Back, Ankles and Feet, Morning Wake-Up, Evening Downshift — and not one of them is a sport track. There are no sport packs.
Limber cannot see your training. Health data goes one direction: a finished session is written out, and no workout is ever read back. The engine holds a designed rule that a high-intensity session should lead with hamstrings, glutes, hip flexors, calves and ankles, and shoulders inside an 18-hour window. It has never had an input, because nothing feeds it. A gate with nothing behind it is not a feature and is not listed as one.
The readings that fit
Three of Limber's five instruments read what this race uses.
The knee-to-wall test reads ankle dorsiflexion in centimetres, one side at a time: 2 cm reads 15 on the app's scale, 6 cm reads 55, 10 cm reads 85. It is the only reading in the battery in real units. One centimetre moves the score 5 to 10 points, so take it the same way each time and treat a one-centimetre difference as the instrument rather than as progress.
The deep squat test is four bands and reads the wall-ball position as well as anything can — with the caveat that its top two bands are separated by comfort rather than depth. Squat depth mobility works through which joint tends to run out.
The Range Score averages the score groups that have a reading and drops the weight of an unread test rather than counting it as zero. Four of five stations is the ceiling today — the spine group is fed only by a thoracic rotation test that is not built, which for a sport with a sandbag on the shoulders is a real gap and is printed rather than filled. No population has ever been measured against this scale, so there is no percentile and no projection.
Anything that hurts during training belongs with a clinician who can examine you rather than with a mobility routine. Nothing here diagnoses anything. Thoracic spine mobility covers the region the battery cannot read, and the cluster sits under mobility for training.
Where to start
Measure the ankle on both sides and take the squat band. Warm up properly before a race rather than holding anything long. Put your range work on days when nothing follows it — rest day mobility covers that slot — and spend the rest of your effort on the strength work, which is the intervention with the prevention evidence behind it.
Then re-measure in six weeks. Two numbers of your own beat any number this sport has published, because this sport has not published any.
Questions
Is there any research on mobility for HYROX?
None. There is no randomised trial, no prospective cohort and no biomechanical study of joint range and performance in this sport, so every confident article about it is reasoning from the stations rather than citing evidence. Anyone giving you a target range or a required angle for a station has estimated it, and the honest thing is to say so.
Which HYROX station needs the most range?
Wall balls, by a distance, and sandbag lunges after them. A full squat has been measured at roughly 95 degrees of hip flexion and 38 degrees of ankle dorsiflexion, and wall balls repeat that under fatigue with an overhead finish at the top. Lunges ask for hip flexor length behind and ankle range in front, with a load high on the body.
Should I stretch before a race?
Move rather than hold. Long static holds immediately before a maximal effort cost around 5 per cent of strength and power, with the threshold at roughly 60 seconds per muscle group and trivial effects below it. If you want a brief hold on something specific, do your dynamic work after it, which is the arrangement that removed the impairment entirely in a 2018 trial.
Will mobility work make me faster over the stations?
Nobody knows, and the honest position is that it probably is not your limiting factor. Range in this sport is only obviously a limiter at the two stations that ask for depth, and even there position and technique change the demand faster than stretching changes the supply. Conditioning and strength have far better claims on your training hours.
Does Limber have a HYROX programme?
No, and it has no HYROX tag either. The catalogue tags moves for running, lifting, cycling, climbing and golf, and nothing in the build reads any of those tags when it builds a routine. Ten programme tracks are drawn and every one of them is a body-area track rather than a sport one.
Can the app plan around my race training?
No. Health is write-only in Limber: it writes a finished session and never reads a workout back, so it cannot know what you trained or when you are racing. The engine contains a designed rule for high-intensity sessions that has never had an input. You choose the areas and the minutes, and the engine shows its reason for every move.