Open any fitness platform’s leaderboard. A 5-foot-3 lifter is ranked on the same chart as a 6-foot-4 lifter, on a movement where leg length determines half the work being done. The 5’3” lifter “places #847 of 2,400 in the squat.” The 6’4” lifter “places #213.” Useful information density: roughly zero.
The right comparison is you, last month. You, on the same band, at the same anchor point. You, after eight hours of sleep, on this exact movement.
What personal benchmarks actually are
tensr keeps every set you’ve ever recorded. The metrics that matter for “is this getting better” are comparisons against that history, on that movement, with that gear.
| Metric | What it compares |
|---|---|
| Session vs personal best | Today’s working-set as a % of the all-time best for this movement |
| Session vs same-day-last-week | Today’s session vs last week’s same day, same movement |
| Curve-shape fingerprint | Per-movement: do this week’s reps look like last month’s? |
| Resistance-profile fingerprint | Has the band/cable setup actually changed the stimulus? |
| Movement velocity-loss equivalent | Tempo-based proxy for the velocity-based-training ”% velocity lost” |
Together these say: for this lifter, on this movement, with this hardware, on this kind of day — is the work better, worse, or the same as the last comparable instance.
Why it matters
Population-level benchmarks have their place — coaching pilots, broad cohort studies, pricing services. But for the lifter showing up at their gym tomorrow, the only comparison that’s actionable is themselves.
Three specific reasons:
Leg length, arm length, frame. A 5’3” lifter doing a deadlift produces different forces and curves than a 6’4” lifter at the same external load. Cross-population comparisons on these metrics aren’t wrong; they’re just not informative for either lifter’s training decisions.
Gear changes the signal. A band that’s six months old produces less peak tension at the same hand position than a fresh band. A cable machine after maintenance can have changed friction. The same workout on different gear produces different N·s totals — which is fine, as long as you know it. The resistance-profile fingerprint catches “wait, this movement looks different now” before you mistake it for losing strength.
Time of day, sleep, food matter. A morning fasted session and an evening fed session aren’t directly comparable for most metrics. Same-day-last-week comparisons reduce that variance: same time, same circadian state, same fueling pattern.
What to track together
The personal-context cluster is one view answering “how does today fit into my history.”
| Metric | When to read it |
|---|---|
| Session vs personal best | Periodic — “where am I in my career?” |
| Same-day-last-week | Weekly — “is this week’s volume profile reasonable?” |
| Curve-shape fingerprint | When something feels different — “did the rep actually change?” |
| Resistance-profile fingerprint | When swapping gear or returning from a break |
| Velocity-loss equivalent | Within a session — “should this set end?” |
The one most lifters underuse: resistance-profile fingerprint. If you swap from a worn 60-lb band to a fresh one, the curve looks different at the same hand position; if you’re not aware, you’ll attribute the new stimulus to “feeling stronger” instead of “the gear changed.” The fingerprint catches that.
What gear it needs
Mostly honest everywhere. The personal-history comparisons (PR, vs-last-week) are summable from impulse and TUT — both forgiving at low sample rates.
Curve-shape fingerprints want 80 Hz. Comparing rep shapes requires sub-second resolution; at 8 Hz, the fingerprint is too coarse to detect meaningful drift.
For lifters on a $30 sensor: the headline comparisons (PR, vs-last-week) are fully usable; the fingerprint comparisons are deferred until they upgrade.
What to do tomorrow
Pull up a movement you’ve done at least four times. Look at the four most recent sessions’ curves overlaid.
Find the rep that doesn’t fit the others. That’s an outlier — either today is different, or that rep is.
Two reads from this exercise:
If today’s outlier is cleaner than the historical reps, you’ve improved. The fingerprint says you’re producing the same force with smoother mechanics — that’s technique consolidation, often the most underrated kind of progress.
If today’s outlier is messier — high variance, dipped middle, jagged force trace — and you can’t immediately explain it (new band, new shoes, unusual fatigue), it’s worth flagging. A single anomalous session is normal; a pattern of anomalous sessions on the same movement is a signal.
A note on cohort comparisons: they’re not anti-pattern. tensr supports opt-in cohort benchmarking for coaches running pilot programs and for lifters who specifically want to compare to a defined group (training partner, online community). The point isn’t that population data is bad — it’s that the default comparison should be personal, and the population view should be a deliberate choice.
You vs you, last week. That’s the leaderboard that matters.