Perceived exertion

Why can two sessions that feel equally hard leave the body in completely different states?

Two sessions, both rated an honest eight out of ten. One leaves you tired that evening and fine by morning. The other leaves you walking downstairs backwards for three days. If the rating is the measure of how hard you worked, how can the consequences differ that much?

The comfortable assumption is that effort is a readout of cost — that "how hard this feels" is the body reporting what the session is doing to it. Suppose we take that assumption apart. What exactly is the sensation of effort measuring, and is the thing it measures the same as the thing we want to know?

Ask first where the feeling comes from. The obvious answer is the muscles: they are working, they send signals, you feel the strain. But that answer runs into an old and stubborn observation. If a muscle is pharmacologically weakened, so that the same task now requires a much larger neural drive to accomplish, the task feels harder — even though the muscle is doing no more work than before. Conversely, blocking sensory traffic coming back from the limb does not abolish the sense of effort.

What does that pattern point to? That effort tracks the outgoing command more than the incoming report. Roughly: your sense of exertion is an internal copy of how hard the nervous system is currently having to drive you, not a summary of what is happening in the tissue. That is why effort rises through a long steady run in which the pace, the muscle, and the oxygen cost are all constant — the drive required to hold that pace is climbing, so the feeling climbs with it.

Now ask what else can raise the required drive. Heat. Dehydration. Poor sleep. Yesterday's session. A stressful week. Low motivation. Each of these makes the same external work require more central effort, and each therefore raises the rating — without changing the mechanical or metabolic damage being done.

And the reverse? Consider a hard downhill run, or heavy lowering under a barbell. Lengthening contractions produce high force with comparatively low metabolic and cardiovascular demand, so the drive is modest and the session feels tolerable. Yet mechanically they are the most damaging thing most athletes do; the soreness and force loss show up a day or two later. The reading was low, the cost was high, and the delay is precisely what conceals the mismatch.

So we have two variables that we had been treating as one. Call them effort and cost. Effort is generated now, centrally, as a control signal — its job is to let you pace, to tell you whether this is sustainable for the next twenty minutes. Cost is distributed, delayed, and plural: glycogen depleted, muscle fibres disrupted, tendons loaded, hormonal and neural fatigue accumulated. There is no reason a single scalar produced for real-time pacing should also be an accurate summary of a multi-dimensional bill that mostly arrives tomorrow.

Once that separation is clear, the original puzzle dissolves and a sharper one replaces it. Why does the rating work as well as it does? Because for most steady aerobic training, effort and cost do move together — which is why session-RPE, the practice of multiplying a ten-point rating by the minutes trained, remains a genuinely useful way to track training load. The scale is not broken. It is a good proxy inside the range where the correlation holds, and it fails at exactly the sessions that sit outside it: heavy eccentric work, very long low-intensity efforts, anything in unusual heat, and anything performed under emotional or motivational conditions unlike normal.

One honest caveat: the origins of the effort sense are still argued over, and the balance between central command and peripheral feedback is not fully settled. What is not in dispute is the practical consequence — the rating and the aftermath can come apart, and predictably so.

Two corrections worth making explicit.

The first: this is not a claim that perception is unreliable or that athletes are bad at judging themselves. The rating is a faithful measurement — of the wrong quantity for this particular question. Treating an accurate measurement of X as an approximate measurement of Y is a very different error from measuring badly, and it is fixed by adding a second instrument, not by trying harder to feel correctly.

The second: the fix is therefore not to distrust the rating but to pair it with something that tracks cost independently. Duration and distance, mechanical descriptors like how much downhill or how much lowering was involved, and next-day markers such as soreness, resting heart rate, or a simple jump test all carry information the effort scale structurally cannot. The point of writing down the rating and the session's content is that the two disagree in an informative way.

• Why does eccentric work cause so much more damage per unit of perceived effort?
• What makes session-RPE hold up so well as a load metric despite this gap?
• If effort is centrally generated, what does that imply for how caffeine, music or a race crowd

Every term the collection defines is gathered in the glossary.