Move

Biometric Training: How to Turn Your Data Into a Daily Plan

Cameron Usher · Founder, PX

11 August 2026 · 16 min read

12 years in competitive sport; trained alongside the Australian Olympic squad; coached 100+ people one-on-one.

You are probably wearing the answer to your training problems right now. A watch or a ring that measures your heart rate, your sleep, your recovery, your strain. It has been measuring them for months. And this morning, like every morning, it showed you a number, and you looked at it, and then you did whatever you were already going to do.

That is the gap biometric training closes. Biometric training means using your body's own signals, heart rate variability, resting heart rate, sleep, training load, to decide what today's session should be. Not to summarise yesterday. Not to give you a score to feel something about. To make the actual decision: what you do today, how hard, for how long, and when.

Most people with a wearable have never trained this way, because the device was never designed to do it. It was designed to measure. Measurement is not direction. This guide covers what your signals actually mean, why reading any one of them alone will mislead you, and how a decision gets made from all of them together, before your day starts rather than after it has already happened.

What biometric training actually is

Strip away the vocabulary and training has always been three questions. What should I do today? How hard? For how long?

For most of history, you answered them one of three ways. You guessed. You followed a static plan written weeks ago by someone who cannot see how you slept last night. Or you paid a coach to watch you closely and adjust as they went, which works, and which is why elite athletes have coaches, and why almost nobody else does.

Biometric training is the fourth way. Your body broadcasts its state constantly: how well you recovered overnight, how much load you are carrying from the last two weeks, how much adaptation you can absorb today. A modern wearable captures most of that broadcast. Biometric training reads it and turns it into today's answer.

The distinction that matters is between measurement and direction. Measurement tells you that your heart rate variability dropped overnight. Direction tells you that because it dropped, alongside what your last three days of load and last night's sleep looked like, today's intervals become a forty-minute steady session, and tomorrow's hard day moves to Thursday. One is a fact. The other is a decision. Almost every device and app on the market today gives you the first and quietly leaves you to make the second yourself, which is exactly the job you were hoping the technology would do.

If you already track everything and still feel none the wiser about what to actually do, that is not a discipline problem or a knowledge problem. It is a synthesis problem, and it is the specific problem this whole approach exists to solve.

The signals, and what each one is really saying

Data-driven fitness starts with knowing what each signal can and cannot tell you. There are six worth understanding.

Heart rate variability (HRV). The variation in time between consecutive heartbeats, measured in milliseconds, usually overnight. It reflects the balance of your autonomic nervous system: broadly, higher variability suggests a system with capacity to take on stress, suppressed variability suggests a system still dealing with the stress it already has. Training load, poor sleep, alcohol, illness and life stress all push it down. It is the most useful single recovery signal we have, and it is also the most misread, for reasons the next section covers. Researchers have tested HRV as a daily programming tool for years, and a systematic review and meta-analysis of HRV-guided training found it performs at least as well as, and often better than, pre-planned training, precisely because it times the hard days to the days you can absorb them.

Resting heart rate (RHR). Slower to move than HRV and easier to read. A resting heart rate a few beats above your normal for no obvious reason is your body telling you it is working harder than usual at baseline, whether from fatigue, heat, dehydration or an oncoming cold. On its own it will not tell you which. Trending alongside other signals, it sharpens the picture considerably.

Sleep, in two parts. Duration is the headline: how many hours you actually got, against what you need. Architecture is the detail: how much of the night was deep sleep, where the physical repair work concentrates, and how much was REM. A seven-hour night built on deep repair work and a seven-hour night of shallow, broken drifting are not the same night, whatever the total says, which is why reading duration alone undersells what your wearable already knows. If you want the full picture of what a rough night should change, there is a dedicated guide to how to train after a bad night's sleep.

Training load. What you have actually done, accumulated over days and weeks, weighted by how hard it was, not just how long. Sports science has modelled this for decades, going back to Banister's training-impulse work in the 1970s, and the principle holds: adaptation follows load, and so does breakdown, depending entirely on whether recovery keeps pace.

Soreness and self-report. The signal that needs no hardware. How your legs actually feel walking downstairs is real data, and any system that ignores it in favour of purely automated inputs is throwing information away.

Adherence. The quiet one, and the most underrated. What you have actually done versus what was planned, across the last week and the last fortnight. Nothing predicts what you can do next week better than what you really did, not what the plan assumed you did.

The thread running through all six: none of them means anything against a population average. An HRV of 45 milliseconds is unremarkable for one person and a red flag for another. The only baseline that matters is your own, built from your own rolling history, so that every reading is a comparison against the person you were over the last two weeks rather than against strangers. Any system doing serious wearable workout data analysis works this way. If a product is comparing you to a global average, it is describing the average, not you.

Why single signals lie

Here is the uncomfortable truth about the recovery score on your wrist: a single low reading, on its own, is close to meaningless.

HRV is noisy. It moves with a late meal, a glass of wine, a warm bedroom, a stressful email read at 11pm. One suppressed morning tells you almost nothing. Three suppressed mornings, alongside rising training load and shortened sleep, tell you a great deal. The signal was never the problem. The interpretation, one number at a time, in isolation, is the problem.

This is where the single-metric approach that dominates the market quietly fails, in three specific ways.

It cannot see combinations. The state that actually damages people, the one that ends in illness, injury or burnout, is rarely one signal in trouble. It is a pattern: high strain, plus suppressed HRV, plus short sleep, arriving together. Sports scientists call the early version of this overreaching. Any single-signal system is structurally blind to it, because each individual signal might still look tolerable while the combination is flashing red. Catching the pattern requires reading training, sleep and recovery in one place, which is precisely what a training-only app or a sleep-only app cannot do.

It cannot see across domains. Under-eat on a heavy training day and your recovery will pay for it tonight and tomorrow. A brutal week of work stress will sit in your HRV just as surely as a brutal week of intervals. Bodies do not respect app categories. Training, food, sleep and stress are one system, and a reading from one domain routinely has its cause in another. This is the case for an all-in-one reading of your data rather than four apps with four opinions: not convenience, but correctness. Four separate answers about one body is how important patterns stay invisible.

It overstates its own confidence. A score built on one thin reading is presented with the same bold certainty as a score built on fourteen days of clean data. This should bother you more than it does. Honest analysis carries its confidence with it: when the data behind a number is thin, the number should be pulled toward neutral and presented more cautiously, not displayed at full strength. When PX has only weak signal behind a reading, it deliberately dampens the score toward the middle rather than showing you a falsely confident extreme. A perfect 100 built on one wearable reading is shown closer to a cautious 67, because that is what the evidence actually supports. If a system never tells you it is unsure, it is not being honest with you.

None of this is an argument against the science. The science is unusually solid: decades of load modelling, a strong evidence base behind polarized intensity distribution, where most training stays genuinely easy so the hard days can be genuinely hard, and a growing literature on HRV-guided programming. It is an argument about application. The research was always about patterns over time, read in context. The market turned it into a single number, delivered each morning, with no instruction attached.

From reading to deciding

Everything to this point is reading. What separates biometric training from biometric tracking is what happens next: the reading becomes a decision, and the decision arrives before the day does.

Concretely, four things about today are adjustable. What kind of session it is. How hard it runs. How long it runs. And when it lands. A system doing this properly adjusts all four, and tells you why.

Say the pattern from the last section arrives: three weeks of building load, then a short night, and this morning your HRV sits well under your baseline while your resting heart rate sits above it. The overreach pattern. Here is what direction looks like, as opposed to a score: today's planned interval session is replaced with forty minutes of easy aerobic work. Its intensity is capped. The week's hard session moves to the day after tomorrow, when the pattern is likely to have cleared. And the reason is stated, in plain language, on the plan itself: strain has outrun recovery for three days, so today is deliberately easy to protect the week.

That last part is not decoration. An instruction without a reason teaches you nothing and earns no trust. An instruction with a reason does both, and after enough of them you start seeing the patterns yourself. The aim of a system like this is not dependence. It is that six months in, you understand your own body the way a good coach would.

Two more decisions belong in this chapter, because almost nobody makes them with data.

Timing. The same session at the wrong hour is a different session. There is a window in the afternoon, several hours after waking, where most people's power output, coordination and injury resilience peak. Hard work belongs there when your calendar allows it. Easy and recovery work pairs naturally with the morning. A plan that knows this places sessions rather than just listing them, and adapts placement to when you actually train rather than when a textbook says you should.

Environment. Heat changes the equation before you feel it. The same easy run at 32 degrees costs meaningfully more than it does at 18, and pretending otherwise is how a recovery week becomes a strain week. Knowing tomorrow is going to be hot is grounds to ease the target before you run, not to wonder afterwards why the run felt terrible. There is a full guide to training in hot weather covering how much to scale and when the smarter move is shifting the session earlier in the day.

And the cross-domain point from earlier now cuts the other way, constructively: fuel decisions become training decisions. Come in well under your calorie target on a heavy training day and tomorrow's prescription should ease, because the recovery simply is not there to support it. When your training plan, your food and your sleep are being read by one system, that adjustment happens on its own. This is what direction means in practice, and it is the difference between a plan and a feed. The everyday version of it, the one you will actually use most mornings, is knowing how to decide whether to train today in the first place.

Capacity: the part almost nobody does

Every plan you have ever quit had the same flaw. It was written for the person who started it, and it kept assuming you were that person long after you were not.

Two weeks in, you missed some sessions. Real life did what it does. The plan did not notice. So Thursday's session was built on top of a Tuesday and a Wednesday that never happened, and within a month the plan described a fictional athlete, and following it felt somewhere between pointless and dangerous, so you stopped. The failure gets filed under discipline. It was actually arithmetic: the plan lost track of your true capacity and never corrected.

Capacity is the honest, current answer to a simple question: how much training can you absorb right now? Not at your peak. Not when you started. Now. It should behave like a dial, not a badge. Train consistently for two weeks and it earns its way up, and the plan gets more ambitious. Fall below half your planned sessions for a week, whatever the reason, and it steps down, and the plan gets deliberately smaller until you are back underneath it. Recomputed regularly, weekly at minimum, from what you actually did.

A plan that shrinks when your life shrinks is not a lesser plan. It is the only kind that survives contact with a real life, and the research on adherence says the plan you keep is worth more than the perfect plan you abandon.

Capacity is also the honest frame for the two hardest moments in training. Coming back after time away, where the near-universal error is restarting where you stopped and where week one is when most comebacks break, covered properly in getting back into training after a break. And the harder version of the same problem, returning to training after injury, where the cost of overreaching is not soreness but a setback, and where a deliberately stepped-down capacity that rebuilds gradually is the difference between a comeback and a relapse.

It also settles the most common daily argument you have with yourself. Feeling flat is not the same as needing rest, and guilt is not a reason to train. When the signals say recover, the honest answer is a specific easy session or a genuine day off, prescribed rather than guiltily improvised. If you find yourself having that argument most weeks, read how to tell if you need a rest day, because it has a better answer than willpower.

Making it survive a real life

Here is where most of the theory above dies: Tuesday, 6pm, when the plan says ninety minutes and you have forty, a late meeting, and half your energy.

The textbook plan assumes a life with training at its centre. Almost nobody has that life. The people this approach serves best, founders, executives, parents, people returning from a long way off, have lives where training fits in the gaps, and the gaps move daily. A plan that cannot see the gaps is not a plan. It is a wish with dates on it.

Which is why the most underrated input to a training system is not a biometric at all. It is your calendar. A system that reads it can do things a static plan cannot: place the hard session in the real ninety-minute window on Wednesday rather than the imaginary one on Tuesday, compress today's session honestly to the forty minutes that exist rather than let you skip it entirely because the full version was impossible, and learn when you actually train, not when you intended to, and place future work there.

Forty real minutes, built around compound work, at the right intensity for today's recovery, beats the ninety-minute ideal you did not do. Twice over it beats the guilt spiral that follows the skip. Consistency at honest volume is how ambitious people actually get fit, and there is a full plan built on that principle in a training plan built for busy professionals.

The same honesty applies to the week as a whole. A good week is not five perfect sessions. It is the right load, distributed across the days you genuinely had, with the hard work landing on the days your body could absorb it. Some weeks that is five sessions. Some weeks it is two and a walk, and if that is what your capacity and calendar truthfully supported, that was a good week, and next week starts from reality instead of from apology.

Running it, day by day

Put the pieces together and biometric training stops being a concept and becomes a loop. It looks like this.

Overnight, your wearable records sleep, HRV and resting heart rate, and they join your rolling baseline. Before your day starts, the signals are read together, against your own history, alongside your load, your recent adherence, your calendar, even tomorrow's weather, and today's plan is set: the session, its intensity, its length, its place in your day, with reasons attached. You train, or you rest, deliberately. You log honestly, including how it felt, because your self-report is a first-class signal. The next day is sized by what actually happened, not what was supposed to happen. And weekly, your capacity recomputes from the truth of the fortnight, and the plan grows or shrinks to match the person you currently are.

Run that loop for a season and something compounds that no single session ever delivers: months of hard days that landed on the right days, easy days that were genuinely easy, and a plan that never lost touch with your real life. The research on HRV-guided programming points to exactly why: it is not that any single adjusted day matters much, it is that timing the stress to the capacity, over and over, is where the adaptation lives.

Starting this week

If this is new territory, the on-ramp is shorter than it looks, and most of it is patience rather than effort.

Days 1 to 14: build the baseline. Wear your device consistently, sleep with it on, and change nothing else. You are not looking for answers yet; you are teaching the system, or yourself, what your normal looks like. Readings in this fortnight are calibration, not verdicts, and acting on them early is the classic beginner error.

From day 15: read before you train, not after. Each morning, three signals against your own normal: HRV, resting heart rate, last night's sleep. One look, one decision about today, using the method in how to decide whether to train today. Keep the easy days honest so the hard days can count.

From week four: let the weeks breathe. Compare what you actually did against what was planned, and size next week from the truth. If that sounds like administration, it is, and it is exactly the layer worth handing to a system built for it.

One more thing belongs in the picture, and it is the reason this site does not stop at training. Move is one third of the system. The same signals that size your training are shaped by how you eat and how your mind is doing, and the same engine that reads them can read all three. A body is not divided into app categories, and the plan that finally works is the one that stops pretending it is.

This is what PX does. It reads whatever you already track, builds one plan across Move, Mind and Fuel before your day starts, and tells you the reason behind every call it makes. Not another score. Direction. If you are done collecting data about the person you were yesterday, start your free trial and let PX build tomorrow's plan tonight.