Open your wearable app this morning.
There is probably a card called Prediction. Or Forecast. Or Outlook. It shows a number. Usually your HRV. Usually with an arrow.
Read it carefully.
Is it telling you what your HRV will be? Or is it telling you what your HRV is, right now, with a confident-looking label on top?
For almost every athlete, the answer is the second one.
What a Prediction Actually Is
A prediction has two things.
It has a horizon. A specific point in the future. Tomorrow morning. 48 hours from now. Friday.
And it has a counterfactual. An if-then. If you train hard tonight, expect X. If you take it easy, expect Y.
Without those two things, you do not have a prediction. You have a measurement with marketing on it.
Your HRV today is 48. That is a measurement.
Your HRV is trending down over the last three days. That is a trend.
If you do the threshold session you have planned for tomorrow, your HRV on Thursday will probably sit in the low 40s, and your perceived effort on Friday’s long run will be high enough that the session will not deliver the aerobic stimulus it was meant to. If you replace tomorrow with an easy aerobic ride, HRV on Thursday will recover into the high 40s and Friday’s long run will land. That is a prediction.
One of those tells you something you cannot already see by looking at the number.
The other does not.
Why This Matters
You have been told the prediction card is the smart card. The card that thinks ahead. The card that earns the subscription.
It is usually not.
It is usually a relabel of current state, dressed up to feel forward-looking, because forward-looking is what the marketing brief said the feature needed to be.
This matters because if you trust the card, you build training decisions around it. You see the green prediction arrow and you do the hard session. You see the red one and you back off. But the green or red was about your state right now, not about what tomorrow looks like, so the decision you just made was reactive, not proactive.
Reactive decisions about training are fine. Most coaching is reactive. But you should not call it predictive coaching when it is not, because it changes what the athlete thinks the system can do.
What Forward-Looking Coaching Actually Says
Real forward-looking coaching has a verb tense problem if you write it down. It cannot just say “your HRV is”.
It has to say “if you do this, your HRV will likely be” and “if you do that instead, your HRV will likely be”.
It has to name a day. Not soon. Not the next few days. A day.
And it has to explain why the two paths diverge. Because of the load you have already absorbed this week. Because of the sleep debt you are sitting on. Because of the threshold session you have stacked against an early start tomorrow.
That is harder to build than a card that shows today’s number with a futuristic font.
But it is the only version of the feature that pays its rent.
What We Changed
We had a Prediction card on the P247 brief that was showing today’s HRV with the word Prediction next to it.
Once you see it, you cannot unsee it. The card was doing nothing.
It now forecasts a future state. It names a specific day, roughly 48 hours out. It says what HRV is likely to look like on that day if the athlete stays on the path they are on. And it says what the alternative path produces.
It is the only version of the card that is honest about what coaching software can usefully tell you. If it does not have a horizon and a counterfactual, it is not predicting. It is reporting with a costume on.
The Test You Can Run on Any Wearable
Open the prediction or forecast card on whatever device you use.
Ask two questions.
What day is this telling me about? If the answer is unclear or “the next little while,” it is not a prediction.
What would the card say if I made a different choice today? If the answer is the same number, it is not a prediction.
You can run this test in thirty seconds. Most wearable forecasts fail it.
The ones that do not fail it are the only ones worth checking in the morning.
P247 builds the weekly decision layer that sits on top of your wearable data. Predictions that name a day. Recommendations that have an if-then. Honest about what the data can actually say.
X Thread
1/ The “Prediction” card on your wearable is probably not predicting anything. It is showing you a number you already have, with a future-tense label on it.
2/ A real prediction has two things. A horizon - a specific day. And a counterfactual - if you do X, expect Y; if you do Z, expect W.
3/ If the card does not name a day and does not change with your choices, it is a measurement in a costume. Not a forecast.
4/ Test any wearable forecast in 30 seconds. What day is it about? Would the answer change if you trained differently today? Most fail both.
5/ P247 ships predictions with a horizon and an if-then. Honest about what coaching software can actually tell you.