Why TrakRun exists, and how it decides.

The problem it was built for

Almost every running app is built to make you run more. The streak, the badge, the weekly total against last week’s — the whole apparatus points one direction, and it works, right up until the point it stops working. Most running injuries are not caused by one heroic session. They are caused by four quiet weeks of climbing volume with the hard days creeping closer together, which is exactly the pattern a more-is-better app rewards.

The information needed to see that coming is already on your phone. Your sleep, your resting heart rate, your heart rate variability and every run you have recorded are sitting in Apple Health right now, in most cases recorded by three or four different apps that never speak to each other. TrakRun reads all of it as one picture and answers a question none of those apps ask: given everything you have done and how you have recovered from it, what should today be?

How the numbers are built

Two principles govern all of it. Everything is measured against your own baseline rather than a population average, and every score shows the factors that produced it rather than handing down a verdict.

• Readiness is built from your overnight recovery signals — heart rate variability, resting heart rate, sleep — read against what is normal for you, then weighed with your recent training and what you have said about how you feel. Every contributing factor is itemised.
• Training load uses the acute:chronic workload ratio, alongside how fast your volume is ramping and whether your hard days have anything easy between them — because the ratio on its own misses both of those, and both matter more than the ratio does.
• Pace zones are anchored on your own predicted 5K rather than a generic table. TrakRun uses a real hard effort where you have one, models it from pace and heart rate together where you don’t, and falls back to your training baseline before it resorts to a population default — and it tells you which of those it used.

That last detail is the pattern for the whole app. Where a number rests on an assumption, the app says so rather than presenting an estimate as a measurement.

Where the confidence ends

The research behind consumer recovery metrics is thinner than the confident version you usually read, and pretending otherwise would be the easiest way to lose your trust the first time a score got it wrong. So, plainly:

• The acute:chronic workload ratio is contested. It is useful as a direction of travel and unreliable as a threshold. Anyone quoting you a hard “danger zone” figure is overselling what the evidence supports.
• Heart rate variability is noisy. A single night tells you very little. Trends against your own baseline tell you something, which is why TrakRun reads it that way and why one bad night does not move the score much.
• Wrist optical heart rate degrades exactly when effort changes fastest. That is why chest strap support exists, and why interval data from the wrist should be read with more caution than steady running.
• The app cannot see most of your life. Work stress, illness coming on, a bad week, altitude, heat. It reads what your devices record, and that is a genuine subset of what determines whether today should be hard.

None of which makes the numbers useless. It makes them one input among several, and the most important of the others is still how you actually feel. TrakRun is written to inform that judgement, not to replace it.

What it will not do

• Compare you to other runners. There is no leaderboard and no benchmark against strangers. A stranger’s 5K time is not information about your training.
• Punish you for resting. Recovery overrides the streak, every time. There is no counter here that turns a sensible rest day into a loss.
• Carry adverts. Not in the free tier, not anywhere.
• Sell your data. Your health data is private to your account — never public, never sold, never shared with data brokers, never used for advertising. Ask us to delete your account and everything attached to it goes with it. The specifics are in the Privacy Policy.

Talking to us

A real person reads both of these. For anything about the app, how a number was calculated, or something that looks wrong: support@trakrun.com. For data, deletion requests or anything covered by the privacy policy: privacy@trakrun.com.

If TrakRun tells you something you think is wrong, we would genuinely rather hear about it than not. The reasoning is shown on every score precisely so it can be argued with.