Reading training data
without fooling yourself.
Most training numbers are easy to compute and hard to interpret. These are notes on the difference: where a metric supports the conclusion people draw from it, where it quietly does not, and how to ask a question that survives the answer.
The same interval, eight weeks apart.
One labelled block, followed across weeks of the same session. Which repeats are fair to compare, the three numbers that hold up, and how much change is the ordinary noise of one day against another.
Read → 14 OCT 2026Which climb was steepest, and where was my VAM best?
VAM is metres climbed per hour, and gradient, weight and wind move it as much as your legs do. How to get every climb out of a ride's elevation trace, what VAM compares fairly, and what it cannot tell you.
Read → 14 OCT 2026Is this number real, or is my power meter lying?
A stuck reading or a dropout makes an average that looks like any other. The three faults that hide inside a summary, why cadence is what tells them from real riding, and the error no check on one file can see.
Read → 14 OCT 2026A trainer ride that arrived as one long interval.
An ERG workout synced from a trainer app as a single interval has nothing for interval analysis to show. How the reps are recovered from the power trace, what each one reports, and where that stops working.
Read → 14 OCT 2026Coach plans and Pro Team: what is shared, and by whose agreement.
A Coach plan lets a coach's assistant read the athletes who choose to share with them. A Pro Team lets two to ten people read each other's training. Who pays, what each person agrees to, the limits, and what each side keeps when it ends.
Read → 29 AUG 2026Your decoupling improved. It might just be block length.
A 44% fall in aerobic decoupling between two sessions looks like adaptation. Most of it came from one block running a minute longer than the other — heart-rate lag does the rest. What the metric needs before it means anything, and two comparisons that survive when it does not.
Read →An assistant will answer whether or not it should.
Give any AI access to your training data and it will compute what you asked for, on whatever you point it at, and describe the result with confidence. Arithmetic has no opinion about whether two efforts were comparable in the first place.
That is not an argument against using one. It is an argument for knowing which questions have a trustworthy answer — and for asking in a way that lets the thing refuse.
Your AI as a training analyst · Questions with no chart · Prompts & charts · Connect your training