Training Load Calculator

Session RPE is a single rating from 1 to 10 for how hard the whole session felt, given about 30 minutes after it finishes. Multiplied by the duration in minutes it produces a training load in arbitrary units. The delay matters: rating during the session captures the hardest moment rather than the whole, and rating the next day captures soreness instead of effort. Rate the session as a whole, including warm-up and cool-down in the duration.

Below about 1.5 is generally considered healthy variation, and above 2.0 is the threshold Foster associated with problems when it coincided with high load. But treat 2.0 as a soft boundary for two reasons. First, it is a threshold derived from observational data in specific athlete populations, not a physiological constant. Second, the number itself depends on which standard deviation you compute, and the two conventions differ by 7.4% — enough to move a week from one side of 2.0 to the other.

Most likely the standard deviation convention. Dividing by n and dividing by n−1 give results that differ by a factor of 1.0801 on a seven-day week, so the sample-based monotony is always 7.4% lower. Neither is wrong, and the source papers do not specify. Other candidates are whether rest days are included as zeros (they should be, and excluding them substantially deflates monotony) and whether the window is a calendar week or a rolling seven days.

Entered as zero, always. A rest day is a real data point and it is the single most influential one in the whole calculation, because a zero sits further from the mean than any training day can, which raises the standard deviation and lowers monotony. Omitting rest days is the most common way to get a badly wrong monotony figure, and it errs in the dangerous direction: it makes an unvaried week look more varied than it was.

The standard deviation is exactly zero and monotony is undefined rather than merely high, because the formula divides by it. This page detects that and says so instead of printing infinity or silently substituting a number. Treat it as the most severe possible reading: a week with no variation at all is the exact pattern the metric was built to catch. Adding one genuinely easy day or one full rest day makes every figure computable again.

It compares this week's load against your recent average, so it describes how much your training changed. The claim attached to it — that staying between 0.8 and 1.3 lowers injury risk — is much weaker than its ubiquity suggests. It is a description of change. Whether that change predicts injury is a separate question the evidence does not settle in its favour.

The honest answer is that it is not well supported. Impellizzeri and colleagues showed the ratio is mathematically coupled, that the 7-day and 28-day windows have no physiological justification, and that a randomly generated chronic load was as associated with injury as the real one — concluding the framework should be dismissed. A study of elite soccer players and pentathletes found the most days lost to injury occurred inside the supposed sweet spot. What survives is modest: large spikes are plausibly risky and a very low chronic base is its own risk factor.

Uncoupled is the more defensible choice, meaning the current week is excluded from the chronic average it is compared against. When the acute load sits inside its own denominator the two are correlated by construction, which inflates the apparent relationship. This page takes your four-week average as an input rather than computing it, so you control which convention you feed it — but if you are assembling that average yourself, leave the current week out of it.

No, and this is the most common misuse. RPE is a subjective rating, calibrated differently by every person who gives one, so the units really are arbitrary. A 2,400 AU week means something only against your own previous weeks. The metrics that travel slightly better are monotony and ACWR, because they are ratios of an athlete's own numbers, but even those assume the underlying RPE scoring was internally consistent.

Move load rather than remove it. Take load off your easiest training day and add it to your hardest: the weekly total is unchanged, the standard deviation rises, and monotony falls. A genuine rest day is the most effective single change available, because a zero is the furthest any day can sit from the mean. This is why hard-easy scheduling shows up in the arithmetic and not just in coaching folklore.

Yes. Session RPE was validated across a wide range of sports and it needs nothing but a rating and a duration, which is its main practical advantage over heart-rate-based load measures that struggle with intermittent resistance work. The caveat is that a 300 AU resistance session and a 300 AU run impose quite different mechanical stresses, and the arbitrary unit cannot see that difference.

Strain is weekly load multiplied by monotony, so it rises fastest when a big week is also an undifferentiated one, which is the combination Foster associated with illness and overtraining. Early work reported that a large share of illness and injury episodes followed spikes in individual strain in the preceding days. That is an observed association in particular athlete groups rather than a causal law, so use strain as a prompt to look at the shape of your week, not as a diagnosis.

Yes. It is free, requires no account, and has no usage limits. ApexConverter is funded by contextual advertising, never by selling user data.

No. The engine runs as Vanilla JavaScript inside your browser under our Zero-Server Client-Side Execution model. Your figures are computed locally and are never transmitted, logged, or stored.

It applies the standard closed-form formula in IEEE-754 double precision, rounding only at the display layer. The engine is reconciled against an independent reference implementation before release.

Yes. The interface is mobile-first with numeric keypad hints and is tested down to a 320-pixel viewport with no horizontal scrolling.

Largely, yes. Because computation is client-side, the page continues to calculate without a network connection once it has loaded.

Amounts display in US$ accounting format, but the underlying mathematics is currency-agnostic. The result is identical in any currency, so simply read the figures in your own.

An em-dash indicates the calculation is not defined for the inputs given — typically a division by zero or a value outside the valid domain. We show a dash rather than a misleading number.

Email apexconverter.praxiscalc@gmail.com with the tool URL, your exact inputs, the output received and the output you expected. Verified mathematical errors are patched within 72 hours.