tensr.fitness

Notes Arc B — Not all reps are equal Post 05 of 14

Junk volume has a fingerprint.

Honest at every supported sample rate

A set of 12. Rep 1 peaks at 400 N. Rep 12 peaks at 240 N. The first six reps came in above 350 N; the last three came in below 270. If you had stopped at rep 9, the set would have looked the same on a count log. It would have looked very different on a force–time curve.

The middle reps were the work. The last three were noise.

What set decay actually is

Set force decay is the gap between the first rep’s peak and the last rep’s peak in the same set, expressed as a percentage of the first peak. A set that goes 400 N → 240 N decayed 40%.

The single number is useful. The shape of the decline is more useful, and decomposes into a few related views:

Why it matters

Decay is not failure. It’s information.

Flat decay (5% or less) means the set was too easy. The load wasn’t enough to fatigue you across 10–12 reps, and your body adapted to almost none of it. If you’re chasing hypertrophy, that’s an empty set.

Smooth, gradual decay (15–30% across the set) is the productive shape. The load was challenging; you stayed above the stimulating threshold for most of the reps; the last few cost you something. This is what a working set looks like on the curve.

Cliff decay (peak force collapses on the last 2–3 reps, dropping below 60% of starting peak) means you stopped too late. The reps after the cliff were below the stimulating threshold and probably came at the cost of form, joint stress, or recovery you didn’t need to spend. They didn’t add stimulus — they added cost.

The shape names the load mistake. A flat session asks you to add weight or shorten rest. A cliff session asks you to cut a rep or two. A smooth session is the dose; do it again next week.

What to track together

The decay-and-fatigue cluster is one diagnostic view of a set. Individually any one number is noisy. Together they tell a story.

MetricWhat it tells you
Set force decay (%)The headline: how much weaker were the last reps than the first?
Decay slope (N/rep)Was the fall steady or did it accelerate?
Tempo decay (s/rep)Slower reps usually precede weaker ones — early warning.
Fatigue indexA noise-resilient summary; better for week-over-week comparison.
Set consistencyLow variance + low decay = clean set. High variance = something changed mid-set.
Form-shape varianceDid every rep look like the others, or did the shape break?

Read them in this order: start with decay percent, look at shape via decay slope and form-shape variance, sanity-check with fatigue index. Tempo decay is the early warning; consistency is the sanity check.

What gear it needs

Decay-by-peak and decay-by-impulse are honest at every sample rate. These are aggregates over reps; they smooth out under-sampling. An 8 Hz crane scale gives you set decay percent and decay slope reliably.

Tempo decay and form-shape variance want 80 Hz. Tempo decay needs sub-second resolution per rep to be meaningful; form-shape variance compares whole-curve shapes that get pixelated at low sample rates.

For lifters on a low-rate sensor: trust the percent and the slope, treat tempo decay as approximate, skip form-shape variance until you upgrade.

What to do tomorrow

Look at one of last week’s working sets. Ignore the rep count. Look at the curve.

That’s the whole loop. The metric doesn’t tell you the set was bad — it tells you what shape mistake to fix. There’s a place for grinding sets, and tensr won’t lecture you about them; just don’t run grinding sets when you meant to run productive ones, and don’t call a flat set a hard one because it felt heavy on the first rep.

The set ended where the curve says it ended. The rep count was a coincidence.


What this looks like in tensr. Open the app, pair a sensor, and the metrics in this post are on the screen the moment you start a set.

A note on the data. Every force sample you record stays on your device unless you opt into sync. The file format is open — SQLite, CSV, NDJSON, all readable with any tool. More on that in the FAQ.

volume