How Are You, Really?

Wellbeing is not a mood. It is how a person is holding up under what is being asked of them, and the person turns out to be the best instrument we have.

For most of my career, the work was using exercise as medicine. People came to me after a diagnosis, after a surgery, in the middle of a disease that was not going to leave. I wrote programs. I measured strength and range and heart rate. Over time I began to suspect that one of the things that mattered most was something I was not measuring at all.

It was how they were holding up. Whether they slept. Whether anyone at home noticed them. Whether they believed the work was worth it. Two people with the same scan and the same program would walk out of the same six weeks in different bodies, and the difference lived somewhere I had no number for.

I saw the same thing earlier, as a graduate student in exercise physiology. Test results moved with an athlete's state, and nothing in the training explained it. The body was reporting on the inside of the person, and we were reading it as noise.

Years later I helped build a tool that tries to ask about that missing thing on purpose. It is called SportsFlow, and its first module is called Wellbeing. This Inquiry is about that module, but mostly it is about the question underneath it, the oldest one in coaching, medicine and friendship: how are you, really?

The answer everyone gives

The answer is "fine." Ask anyone who has coached, parented or managed people and they will tell you the same. Fine is the national anthem of people under load. It means I don't want to talk about it, or I don't know, or I'm scared that if I say anything else I'll lose my place.

Most of us also say fine to ourselves. That is the harder problem. Often the body knows first. The mind catches up later, and the mouth later still. By the time someone says "I'm not fine," the cost has often been building for weeks.

So the job of a wellbeing measure is not to catch a liar. It is to give a person a better way of noticing, early, what they already half know. If it does that, the person gets the benefit first. Everyone else is a distant second.

Turning that into an instrument meant answering three design questions. What do you ask? What do you compare the answer with? And who gets to see it? The rest of this piece takes them in that order.

Does a questionnaire beat a blood test?

Here is the part that surprises people. When it comes to tracking how training is landing on a person, the humble questionnaire holds up better than the lab.

In 2016, Anna Saw, Luana Main and Paul Gastin pulled together 56 studies that measured athlete wellbeing two ways at once: with self-report, and with objective markers like blood values and heart rate. Two findings came out of it. First, the two kinds of measure mostly did not agree with each other. Second, the self-report measures tracked training load with better sensitivity and consistency than the objective ones did.

That does not mean the body has nothing to say. It has plenty, and the next Inquiry in this series is about listening to it. It means that a person's own report, asked well and asked regularly, is a serious instrument. It is not a soft add-on to the real data. Often it is the real data.

A blood test can tell us something important about the body. A questionnaire can tell us something important about the person living in it. Neither is complete. The mistake is believing that because one produces a more precise number, it necessarily tells a more complete truth.

It also means the questions matter. "How do you feel, 1 to 10?" asked once a week is close to useless. Specific questions, short enough to answer honestly, asked on a steady rhythm, are a different animal. They let a person watch their own line over time. That line is where the information lives.

Here is exactly how far the research carries us. The review supports asking, and asking regularly. It does not tell us which questions are best, and it did not test SportsFlow's. The six daily sliders and most of the battery are our attempt to turn that principle into wording. They are a design, not a validated scale. The exceptions are two of the ten instruments, the clinical screens, which are established tools from outside. The other eight are ours, and this piece is about the choices inside them.

Your line, not the average

Most wellbeing tools compare you to other people. They tell you where you sit against a squad, a norm, a population. This feels scientific. It is mostly useless for one person, and sometimes it is dangerous.

Look at the figure. The person in blue is still above the group average at week twelve. Any system built on comparison would call them fine. But they have dropped thirteen points from their high, steadily, over eight weeks, and slipped out of the range they usually live in. That is a person in trouble who looks healthy on paper.

The reverse happens too. Someone whose usual range sits low can be having the best month of their year and still read below the average. Compare them to the group and you worry about the wrong person.

That is why the SportsFlow battery holds a hard rule: there are no population norms anywhere in it. Every comparison is to the same person, earlier. In the current design, a single reading is treated as carrying a margin of about twelve points either way, so a move smaller than that is read as noise. That margin is a working estimate until the validation data exists. A move larger than that, held over a few readings, is treated as news.

What wellbeing actually covers

I have come to think wellbeing is a bad word for what we mean. It sounds like a spa. What a coach, a parent or a person actually needs to know is three things: Am I ready to do today's work? How am I holding up under the work of the last few months? And where should I put my attention?

The module answers those with ten instruments in three tiers. Four core instruments: a daily check-in of six sliders, and three that look at steadiness, emotional regulation and how attuned a person is to what is going on inside them. Four depth instruments, including presence, self-kindness and gratitude. And two short clinical screens, for mood and for anxiety, which only switch on at clubs that have a named person responsible for wellbeing.

Some of those instruments sound soft. Self-kindness. Gratitude. I want to defend them, because they are not soft at all.

Self-kindness is one of three parts of what researchers call self-compassion, alongside a sense of common humanity and mindful awareness (Neff, 2003). It is the habit of treating your own mistakes the way you would treat a friend's. The research suggests people who have it are not less driven. In one set of experiments, self-compassion after a failure or a mistake increased people's motivation to improve. The authors' reading is that a mistake stopped feeling like a verdict on their whole worth. Gratitude has its own research base: in Robert Emmons and Michael McCullough's studies, keeping a gratitude practice raised wellbeing on several, though not all, measures. In my experience it is also one of the better markers of a person who is still connected to why they do what they do. When it goes, other things often follow.

These instruments carry no wearable data at all. There is no defensible way to infer self-kindness from a heart rate, and a product that tried would be guessing about a person's inner life from their pulse and calling it science. The founding principle of the battery is that the machine serves the person, and the person is never the raw material.

One word for the coach

Here is the tension at the center of this module. A coach needs to know who is ready to train hard today. A person needs a private place to be honest. Those two needs pull against each other, and most products pick one and pretend the other does not exist.

The first version of this module picked privacy. Nothing was shared by default. It was ethically clean and practically useless, because the whole point was to help athletes and coaches look after a group together. So it was rebuilt in three layers.

The first layer is one word, each morning: Ready, Hold or Back off. It is derived from the readings underneath but never exposes them. It goes to the coach by default, and that default is stated to every athlete as club policy before they answer a single question. The athlete sees exactly what the coach sees, in exactly the same words. No secret second view.

A default is still a decision made on someone's behalf. That one word is built from answers the athlete gave privately. It reveals something. Not the readings, but a direction. An athlete who knows that may shade their answers to protect their place, and then the word stops being true. So the word is kept deliberately coarse, the rule is stated before anyone answers anything, and the athlete can always see exactly how today's word was reached. Consent that is not informed is not consent. And a club that ever used the word in selection would break it in a season.

The second and third layers are the person's. Scores are off until granted. The interior, which includes the clinical screens and anything written in notes, stays off unless the person explicitly opens it, and granting a score never grants the body data beneath it.

Two small design choices carry a lot of weight here. A missing check-in shows up as "No read," never as fine. And a score the person has chosen not to share shows up as "not shared," never as a blank. Absence of data must never look like absence of a problem. If you have ever watched someone go quiet for a week before something broke, you know why.

Where to put your attention

The most useful screen in the module is not a score at all. It is a short list called Your attention.

Most dashboards show you your lowest number and let you stew. That is backwards. Your lowest number is often just where you always are. The useful question is what has moved.

So the attention engine is designed to rank findings by the largest move against your own line, not by the lowest raw score. It returns one thing to focus on, one thing that is working and should be left alone, and one thing it noticed but is holding in reserve. It changes at most once a week. Attention is a limited resource and a new alarm every morning teaches people to ignore alarms.

And it does not give advice. It points to a practice. A short reading from the wisdom library, or a station in the emotional regulation series. I am wary of software that tells people how to live. I am less wary of software that says, here is something people have been practicing for two thousand years, it might be worth five minutes today.

A reading that tells you what moved is useful. A reading that tells you who you are is a horoscope.

What it refuses to do

Every honest measure needs a list of things it will not do. These are locked in the Wellbeing module at every tier, and they are written on the coach's own settings screen where nobody can claim they did not see them.

Wellbeing data is never used in selection. Not for seats, not for teams, not for scholarships. The moment a wellbeing score helps decide who makes the boat, every athlete in the club learns to answer it strategically, and it stops measuring anything.

There is no button for a coach to request access to a score. Asking is pressure, and pressure turns consent into compliance.

Revoking a grant is retroactive. If you take a score back, it disappears from the coach's history too, not just from next week.

And the AI coach that sits inside the product follows the same gate. A withheld score never enters the request that is sent to the model. It cannot leak what it was never given. When it is asked about something withheld, it says so instead of guessing.

The part that is still unfinished

Here is where this stands. The field is full of people selling certainty they do not have.

The structure of the module is settled: what is measured, how often, how it is weighted, who sees what. The wording of most of the items is still a working draft that has to go through proper validation before anyone should lean on it. Two of the ten instruments are published clinical screens, and they switch on only at clubs that have a named person responsible for wellbeing. And nothing here is a diagnosis. A low reading is a set of answers about a person's last few weeks. It is a reason for a conversation. It is not a label.

I am fine with that. I would rather publish the honest state of an unfinished instrument than a confident one that is wrong.

None of this is new. Good coaches and good nurses and good grandmothers have always done a version of it. They look at you and they know. What a careful instrument adds is memory and patience. It remembers how you were in March. It notices the slow slide that a busy person misses. And if it is built right, it tells you first.

The most useful thing a coach can do is not a better program. It is noticing that someone is struggling before they have said so, sometimes before they know it themselves. That is not talent. It is attention, repeated. This module is an attempt to make that kind of attention available to more people than one coach can see, and to hand it back to the person first.

The purpose of measuring wellbeing is not to make a person more legible to a machine. It is to make the person more visible to themselves. And, when they choose, to someone who cares enough to notice.

Orion