Ask almost any serious coach how they track training load, and within two sentences you’ll hear “session RPE.” The protocol is simple enough to be irresistible: multiply a single 1–10 rating by the duration of the workout, log the number, and you have a training load. The metric has spread through team sports, endurance squads, and strength-and-conditioning facilities with remarkable speed, and it has done genuine good—it gave coaches a quick way to quantify effort when heart-rate monitors were impractical or unavailable. But session RPE has a problem that most users never confront: it is not a direct measurement of physiological stress. It is a conscious perceptual integration of multiple afferent signals—ventilatory drive, muscle metabolite accumulation, core temperature, cardiovascular strain, and central nervous system output—filtered through the athlete’s interpretation of the question. When you treat session RPE as a universal, modality-agnostic load currency, you are assuming that a “7” after a heavy squat session means the same physiological thing as a “7” after a two-hour endurance ride or a repeated-sprint interval set. It does not.
This article traces the physiological pathways that produce perceived exertion, identifies the specific conditions under which RPE systematically diverges from actual metabolic and neuromuscular stress, and offers practical calibration anchors that coaches can use to keep RPE honest. The argument is not that RPE is useless—used carefully, it is one of the most practical tools available to working coaches. The argument is that RPE without calibration is like a compass that has never been checked against true north: it points somewhere, but not necessarily where you need to go.
What RPE Actually Measures: The Afferent Integration Model
Perceived exertion is not a single signal. It is a composite output assembled by the central nervous system from a stream of afferent feedback originating in the working muscles, the cardiovascular system, the respiratory system, and the thermoregulatory system. The model that best captures this is the one advanced by Gunnar Borg in his foundational work on perceived exertion, later elaborated through the central governor theory proposed by Timothy Noakes: the brain continuously integrates afferent signals from the periphery and generates a conscious sense of effort that reflects, but does not directly quantify, the underlying metabolic state.
Several distinct signaling pathways contribute. Group III and IV muscle afferents respond to mechanical pressure, stretch, and chemical changes in the interstitial fluid—including accumulation of hydrogen ions, extracellular potassium, and lactate. These chemosensitive afferents relay information about local muscle environment to the dorsal horn of the spinal cord and onward to the sensory cortex. In parallel, baroreceptors and chemoreceptors in the carotid and aortic bodies feed information about blood pressure and blood gas status. Pulmonary stretch receptors and metaboreceptors in the diaphragm and intercostal muscles signal respiratory effort. Thermoreceptors in the skin and core provide input on heat load. The brain integrates all of this and produces what we experience as effort.
The critical point is that this integration is weighted, not additive in a simple way. When core temperature rises, thermal strain can dominate the perceptual signal even if metabolic stress has not changed proportionally. When ventilation approaches its mechanical limits, respiratory effort can become the primary driver of RPE independent of what is happening in the locomotor muscles. When glycogen is depleted in specific fiber populations, the metabolite profile shifts in ways that may or may not correspond to the overall workload completed. A coach who logs a session RPE of 8 is recording the athlete’s integrated perception of all these signals—not the metabolic cost of the session.
Where RPE Diverges From Physiological Reality
Understanding that RPE is an integration signal immediately reveals the conditions under which it will mislead. Four scenarios deserve specific attention because they are common in real training environments and because the direction of the error is predictable.
Repeated Sprint Sets and Neuromuscular Fatigue
Consider a repeated-sprint protocol—say, 6 × 30-second maximal efforts with 4-minute recoveries. After the third or fourth repetition, the athlete’s RPE will climb sharply, often reaching 9 or 10 on subsequent sprints. The natural interpretation is that metabolic stress is accumulating proportionally. But blood lactate typically peaks after the second or third sprint and then plateaus or even declines as the session progresses, because clearance rates adapt and because the athlete’s power output drops. What is actually driving the rising RPE is neuromuscular fatigue—specifically, the decline in motor unit recruitment capacity and the accumulation of extracellular potassium in the muscle interstitium, which impairs membrane excitability. The athlete feels like they are working harder because each sprint requires greater central motor drive to produce less power, but the metabolic cost per sprint is falling, not rising.
If a coach compares this session’s load to a continuous 45-minute tempo run at a session RPE of 8, they are equating a neuromuscular-dominated stress with a metabolic-dominated stress. The training adaptations are different, the recovery timelines are different, and the subsequent programming implications are different—but the session RPE log treats them as equivalent.
Heat Exposure and Thermal Strain
Exercise in hot environments produces a well-documented upward shift in RPE at any given workload. Core temperature elevations of as little as 0.5°C can increase RPE by a full point or more on the 10-point scale, even when oxygen uptake, heart rate, and blood lactate are identical to a thermoneutral control condition. The mechanism is straightforward: thermosensitive afferents and the central integration of heat load add a perceptual cost that is independent of metabolic strain. An athlete doing a moderate interval session at 32°C and 60% humidity will report a session RPE that implies a far greater training load than the same session at 18°C. If the coach logs both as comparable because the RPE × duration product is similar, they will underprescribe the next session in the heat and overprescribe the next session in cool conditions.
This is not a minor calibration issue. In environments where athletes train through seasonal temperature swings—or travel between climates for competitions—uncalibrated RPE can produce load management errors of 20–30% across the year.
Glycogen-Depleted States and the Perceptual Shift
When an athlete trains in a glycogen-depleted state, RPE rises at any given power output or pace. This is partly because glycogen depletion forces a shift toward greater fat oxidation, which produces less ATP per unit of oxygen consumed and increases the relative oxygen cost of submaximal work. But the perceptual shift is disproportionately large compared to the actual metabolic change. The reason is that glycogen depletion in specific fiber populations—particularly type II fibers—triggers group III and IV afferent signaling that is interpreted by the central nervous system as a warning, not merely a report. The brain increases the sense of effort to reduce the likelihood of continuing into a state that threatens homeostasis.
This means a session completed with low glycogen availability will carry a high session RPE that reflects a protective perceptual response, not the actual training stimulus delivered to the muscle. A coach who uses that RPE to plan recovery time may prescribe excessive rest for what was metabolically a moderate session. Conversely, a session completed with full glycogen stores at the same RPE may have delivered a substantially greater metabolic stimulus that deserves more recovery.
Strength Training and the Modality Problem
Perhaps the most consequential misuse of session RPE is comparing strength training loads to endurance training loads. A heavy lower-body session—say, 5 sets of 3 repetitions at 85% of 1RM with 4-minute rests—will produce a session RPE of 7 or 8 for most athletes. A 90-minute Zone 2 endurance ride might produce the same rating. But the physiological demands are almost entirely non-overlapping. The strength session taxes the neuromuscular system, produces minimal metabolic acidosis, and drives mechanical tension and hypertrophic signaling. The endurance ride taxes oxidative metabolism, depletes glycogen progressively, and stimulates mitochondrial biogenesis and capillary remodeling. Logging them as equivalent training loads because the RPE × duration product is similar is like adding apples and oranges and calling the result fruit. Technically true, but nutritionally misleading.
The issue is compounded by the fact that RPE in strength training is heavily influenced by the athlete’s psychological state, familiarity with heavy loads, and bar speed perception. A lifter who is confident under heavy weight will rate a session lower than a lifter who is anxious, even if the mechanical output is identical. The subjective component is not noise—it is real perception—but it means that the metric is measuring something different in the weight room than it is on the road.
When Session RPE Diverges From Physiological Markers: Evidence From Validation Research
Session RPE did not gain its near-universal adoption because it was validated against gold-standard physiological measurements across every modality and condition. It gained adoption because it is practical, free, and produces a single number that can be logged in a spreadsheet. The validation research that does exist tells a more nuanced story than most coaches realize.
The foundational validation work by Foster et al. (2001), published in the Journal of Strength and Conditioning Research, demonstrated that session RPE correlates reasonably well with heart rate-based training load metrics and with blood lactate accumulation during continuous and intermittent exercise. That study is frequently cited as proof that session RPE works. What is less frequently acknowledged is that the correlations were strongest within a single modality and under controlled conditions. The study did not test whether session RPE values are equivalent across different exercise types—and subsequent research has shown they are not.
The evidence for this point is grounded in Reuters and Pew Research Center, which keeps the article’s claims tied to outside reference material rather than product framing.
The afferent integration model advanced by Borg and elaborated by Noakes in his central governor framework explains why. Because perceived exertion is a composite central nervous system output assembled from multiple peripheral signals, the same RPE value can reflect entirely different physiological states depending on which afferent pathways dominate. A session RPE of 7 during a glycogen-depleted endurance ride, a heat-stressed interval session, and a heavy resistance training day may all produce the same number on the log, but the underlying metabolic, neuromuscular, and thermal stresses are fundamentally different. Noakes’s central governor model specifically predicts this: the brain generates a perceptual output designed to protect homeostasis, not to report metabolic cost transparently.
Research on session RPE in resistance training has revealed additional complications. Studies comparing session RPE to total volume-load (sets × reps × weight) have found that RPE is sensitive to proximity to failure and psychological state but does not reliably track mechanical work across different exercise orders or rest intervals. A session of heavy singles at 90% 1RM may produce a higher session RPE than a higher-volume hypertrophy session at 65% 1RM, even though the total mechanical load is substantially lower. The perceptual signal is dominated by neuromuscular strain and bar speed perception in the heavy session, while the hypertrophy session’s perceptual signal is driven by metabolic accumulation—two different physiological realities producing numbers that a coach might be tempted to compare directly.
When session RPE has been tested against objective physiological markers across modalities, the correlations break down. An endurance session and a resistance session producing identical session RPE values show divergent heart rate, lactate, and neuromuscular fatigue profiles. This is not a failure of the metric—it is exactly what the afferent integration model predicts. The metric was never designed to be modality-agnostic. It was designed to capture subjective perceptual load within a given exercise context, and it does that reasonably well. The problem is not with the metric but with the cross-modal comparison that was never validated.
Calibrating RPE Against Objective Anchors
The solution is not to abandon RPE. It remains one of the most practical tools available to coaches, particularly in settings where laboratory testing is not feasible. The solution is to calibrate it individually and modality-specifically, so that each athlete’s RPE is interpreted within the context of what their physiology is actually doing. Several objective anchors can serve this purpose.
Blood Lactate as a Metabolic Anchor
Periodic blood lactate sampling during key sessions provides a direct measure of the metabolic intensity the athlete actually achieved. If an athlete reports a session RPE of 8 but post-session lactate is 4 mmol/L, the perceptual rating is likely inflated by thermal, neuromuscular, or motivational factors. Over time, a coach can build an individual profile of the lactate values that correspond to each RPE level for each modality, creating a calibration table that reveals the athlete’s perceptual tendencies. This does not require lactate testing at every session—even quarterly calibration sessions are enough to detect drift in the RPE-lactate relationship.
Heart Rate Drift as a Cardiovascular Anchor
For endurance sessions, heart rate drift—the gradual increase in heart rate at a fixed power output over the duration of a session—provides a cardiovascular anchor that is independent of the athlete’s perceptual state. If an athlete reports a moderate RPE but heart rate drift exceeds 10% over the session, cardiovascular strain was accumulating faster than perception reflected. This pattern often appears in athletes who are highly motivated or competitive in group training settings—they suppress their perceptual awareness of fatigue until cardiovascular drift becomes severe. Conversely, an athlete who reports a high RPE with minimal heart rate drift may be experiencing perceptual inflation from anxiety, heat, or glycogen depletion without substantial cardiovascular cost.
Velocity-Based Metrics as a Neuromuscular Anchor
In strength training, bar velocity measured with a linear position transducer or accelerometer provides a neuromuscular anchor that RPE cannot. If an athlete reports a session RPE of 7 but bar velocity on the final set is within 5% of the first set, neuromuscular fatigue is minimal despite the perceptual rating. If velocity drops 15–20% across sets while RPE remains at 6, the athlete is either under-rating the session or is one of the rare individuals whose perception lags behind neuromuscular decline. Either way, the velocity data corrects the perceptual report.
Power Profile Consistency as a Performance Anchor
For cyclists and runners, comparing the power or pace achieved in a session to the athlete’s known power profile or critical speed provides a performance-based reality check. If an athlete reports a session RPE of 9 but the average power was 75% of their functional threshold power, the session was not physiologically extreme—something else was driving the perception. If the same athlete reports a 6 but the power was at threshold, the athlete is either under-perceiving or deliberately under-reporting, and the coach needs to investigate which.
Building an Individual Calibration Protocol
The practical implementation is straightforward but requires discipline. Start by selecting one session per modality per month for calibration. In that session, collect RPE at set intervals—not just at the end—and pair it with an objective anchor: blood lactate for metabolic sessions, heart rate drift for endurance sessions, bar velocity for strength sessions, or power/pace relative to known profile for performance sessions. Over three to four months, you will have enough paired data points to identify each athlete’s perceptual tendencies.
The Real Value of RPE: What It Is Good For
A final note on documentation: the calibration data you collect is only useful if you can retrieve it and act on it. Coaches who maintain structured logs of each athlete’s perceptual tendencies—what they over- or under-perceive, under what conditions, relative to which anchors—make better programming decisions than any single metric allows. Whether that documentation lives in a notebook, a spreadsheet, or a dedicated drafting tool like the Unsloppy AI Writing App matters less than the principle behind it: if your calibration data is not documented in a format you will actually revisit, it will not influence future decisions.