Over the past few weeks, we’ve started a structured testing process aimed at something that sits at the core of HYROX performance — but is still poorly quantified: the true cost of stations.
This post is not a final answer or a polished framework.
It’s an update on where we’ve started, what the first data looks like, and how we plan to build on it.
Why test stations at all?
In endurance sport, most disciplines come with clear physiological anchors.
For running, cycling, rowing, or skiing, we can reference threshold pace or power, heart rate responses, efficiency, and durability. These anchors allow us to quantify intensity, cost, and sustainability with reasonable precision.
Not all HYROX stations work like that.
Movements such as wall balls, sleds, lunges, farmer carries, and burpees lack an equivalent “threshold” metric. Unlike cyclical endurance disciplines, these stations impose a mixed load that is not dominated by a single physiological limiter. Local muscular fatigue, movement economy, coordination under load, and neuromuscular stress can all influence how an athlete returns to running — often without a clear signal in heart rate or traditional metabolic markers.
That makes stations one of the largest performance variables in HYROX — and one of the least measured.
The goal of this testing is not to remove intuition, but to add structure around it:
Where does station intensity begin to meaningfully degrade running? How much time saved on a station is offset by increased physiological cost? What is a repeatable “sweet spot” of exertion for a given station and athlete?
Phase 1: Testing approach
To reduce noise and cumulative fatigue, this first phase deliberately avoids full race simulation.
Instead, we fixed:
- Running pace (moderate, repeatable, well below race intensity)
- Full recovery between rounds (to minimise central aerobic carryover)
- One station, multiple controlled intensities
Each round consisted of:
A fixed-pace run One station performed at a prescribed intensity A repeat of the same run Full recovery before the next round
This structure allows us to examine the isolated cost of the station, rather than the compounding effects of fatigue across an entire race simulation.
What we observed
Even with full recovery and controlled running pace, the data behaved in a largely logical way.
As station intensity increased:
Station time decreased Heart rate drift during the post-station run increased Recovery dynamics changed measurably despite identical pacing
This reinforces a key assumption: stations carry a physiological cost that is not captured by time alone.
Moving toward a usable metric
To begin connecting performance and cost, we experimented with a combined “block time” concept:
Station time + fixed run time weighted by the heart rate response following the station
This is not a finished metric. The weighting factors are provisional and will almost certainly change as more data is collected. However, even with limited athletes, small number of rounds, the results were informative.
The most aggressive station effort was not optimal.
The easiest effort was time-inefficient.
The signal pointed toward an intensity between those extremes, rather than at either end — which aligns with practical racing experience.
What comes next
Phase 1 was intentionally narrow:
One station – Limited rounds Fixed constants
That’s a feature, not a limitation.
The next steps are to:
- Replicate the protocol across more stations
- Test across athletes with different profiles
- Keep the structure consistent so the data remains comparable
As patterns emerge, weighting factors can be refined — or replaced entirely — without invalidating earlier tests.
Final thoughts
This isn’t about finding a single number that solves HYROX pacing.
It’s about building a framework that helps athletes and coaches:
- Understand station cost more clearly
- Make better pacing decisions under fatigue
- Train stations with intent rather than guesswork
We’re early in the process — but the first signals are logical, repeatable, and worth pursuing.
More to come.
