The same session-RPE score does not establish that two runs are interchangeable. The calculation combines duration and perceived overall effort into one number. Once you keep only that number, you lose information needed to describe the session. A useful training review keeps the inputs alongside the score.

Two runs, both 240 points

Synthetic example: these are invented entries for explaining the calculation, not a training recommendation or measured athlete results.

  • Run A: 30 minutes × session-RPE 8 = 240 arbitrary load units.
  • Run B: 60 minutes × session-RPE 4 = 240 arbitrary load units.

The arithmetic is correct in both cases. B lasts twice as long and has half the reported session-RPE. Nothing in the product of those two inputs tells you the pace, distance, terrain, interval structure or purpose of either run. Neither entry tells you what happened to the runner's fitness afterward.

That is why matching points cannot decide a substitution. Replacing A with B changes the recorded duration and overall effort. The equality establishes a shared score under this formula; it does not establish equal training effects. The example also cannot tell you which run would be appropriate for a particular person.

What the original study compared

Foster et al.'s February 2001 paper compared session-RPE with heart-rate-zone scores in 12 recreational cyclists and 14 male college basketball players. Ratings covered the whole session, collected 30 minutes afterward. For 30-minute steady cycling, mean scores were 130 versus 110 respectively (Table 4, p.113). The study compared monitoring methods, not fitness adaptations from equal-score runs. Small, selected samples and non-running protocols limit direct application to recreational runners. This is one historical study, not a consensus review.

These are two separate comparisons: the paper compared different scoring methods; our synthetic example compares different inputs to the same formula. Its findings should not be presented as an experiment proving that runs A and B produce different adaptations.

Keep the inputs in the weekly review

For the two entries above, a useful note would read: “Both scored 240. A lasted 30 minutes with session-RPE 8; B lasted 60 minutes with session-RPE 4. Their session structure and purpose still need to be described before discussing a replacement.” This makes the unresolved question visible instead of hiding it behind an equal total.

  • Keep duration and the whole-session rating separately. An effort rating for one hard interval is a different observation from a rating for the entire run.
  • Keep the sport and a short account of what was completed. Add known distance, session structure and relevant conditions rather than trying to reconstruct them from points.
  • Check that the same scale and calculation method were used. A number from another load model is not automatically comparable because it is also called training load.
  • Identify incomplete entries before comparing totals. A blank effort rating supplies no basis for calculating that session's RPE load.

A recording gap is not a lighter week

Extend the synthetic example with run C: 40 minutes at session-RPE 6 adds another 240 points. With A, B and four explicitly confirmed rest days, the seven-day history is complete.

  • Complete record: 240 + 240 + 240 = 720 load units for the seven days.
  • Copy with C’s effort rating removed: 480 known load units; C’s load and the complete weekly total remain unknown.

Only one field is missing from the copy. The recorded runs have not become shorter or less frequent. Reporting 480 as the weekly total and calling it a one-third load reduction would count the recording gap as a decline. NuroStride’s calculator returned 720 for the complete history and no calculable weekly load for the copy. Even seven documented days do not guarantee complete load values.

For an actual week-to-week comparison, name the missing rating and the known subtotal. Do not insert zero or a previous session’s rating just to produce a total. A missing day is not a confirmed rest day either. Even two complete totals describe only this load model; they do not establish a particular training effect.

What NuroStride actually calculates

NuroStride's method uses minutes × a reported 0–10 session-RPE value for this load option. Its local Python calculator returned 240 for each synthetic entry above. Running the same example with the effort rating absent returned an unknown session load, not zero. This checks arithmetic and missing-data handling; it is not a test of coaching effectiveness.

The kit is used in your own AI environment. It includes no ready-made watch connection. Its method requires missing days to stay unknown and rest days to be explicitly documented. Those rules describe the record available for review; they do not prove recovery or readiness to train.

For a comparison, retain the original observations: duration, perceived overall effort and what the run contained. Equal points can be a starting observation. They leave the training decision unresolved, and they provide no medical clearance or performance guarantee.

Sources and product behaviour checked on 20 October 2026. All example inputs are synthetic. The source study was published in 2001; this is a measurement explainer, not a report of a new trial.