How Hard Should You Push? An RPE-Based Framework

The most useful question you can ask in a workout is not "should I train harder" or "should I take it easy" — it's "how hard is this, on a scale of one to ten?". The science of perceived exertion gives you the answer, and the data say a lot of well-meaning people are stuck at one of two failure modes: never quite working hard enough to drive adaptation, or working so hard so often that progress stops, sleep breaks down, and injury arrives.

What RPE actually is

Borg published the original 6–20 RPE scale in 1982 Borg 1982. The numbers were chosen to roughly correspond to heart rate × 10 (so RPE 13 ≈ HR 130 bpm in healthy adults). It was the first systematic attempt to give athletes and clinicians a way to talk about effort that didn't require equipment.

The modern variant most widely used in resistance training is the RIR-based 0–10 scale developed by Zourdos and colleagues Zourdos 2016:

For endurance work, a 0–10 conversational scale is more practical: 1–3 = can sing, 4–6 = can hold a conversation, 7–8 = short sentences only, 9–10 = barely able to speak.

The dose that drives adaptation

Schoenfeld's 2017 how the dose changes the result meta-analysis found that each additional weekly set per muscle group adds roughly 0.37% to muscle mass, with the curve plateauing around 10+ weekly sets per muscle group Schoenfeld 2017. Critically, the productive sets in this literature are sets taken close to (but not always at) failure — an RPE of 7–9 in most studies.

The Helms RIR review summarises a converging picture: working sets at 0–4 RIR (RPE 6–10) drive hypertrophy and strength similarly, with two important caveats Helms 2016:

• Sets above RPE 9 (1 RIR or fewer) accumulate fatigue rapidly and impair recovery between sessions when used on every set.
• Sets below RPE 6 are below the threshold to drive adaptation in trained adults.

Translation: most working sets should sit at RPE 7–8. Save RPE 9–10 efforts for occasional max attempts or specific hypertrophy finishers, not the bulk of your training.

"Effort is the variable that drives adaptation. Volume shifts how much. Frequency shifts how often. But effort is the trigger. If sets are too easy, the trigger doesn't fire." — per Helms 2016 RIR-RPE review, Strength & Conditioning Journal

Weekly load: session-RPE × duration

Foster's 2001 paper introduced what is now the most-used training-load metric in sport science: session-RPE × minutes-of-training = "session load" in arbitrary units Foster 2001. A 60-minute workout at RPE 6 = 360 units. The 2022 25-year retrospective showed it has held up across endurance, team sport, and strength populations Foster 2022.

Halson's 2014 review documented why this matters Halson 2014:

• Acute fatigue (within a session): normal, recovers in hours.
• Functional overreaching (1–2 weeks of high load): short-term performance dip, full recovery + supercompensation in 1–2 weeks. This is the productive zone.
• Non-functional overreaching (2–4 weeks of unrelenting load): performance drops and stays down. Mood, sleep, motivation degrade. Recovery takes weeks.
• Overtraining syndrome (months of accumulated overload): performance loss, hormonal disruption, mood disorders. Recovery takes months to years.

The 2013 ECSS / ACSM consensus statement on overtraining is unambiguous: there is no biomarker that reliably distinguishes overreaching from overtraining in real time Meeusen 2013. The only protective tool is load monitoring — not after the fact, but week to week, so that spikes are caught and de-loaded before they cascade.

The acute:chronic workload ratio

Gabbett's 2016 review reframed sports medicine's understanding of injury risk Gabbett 2016. The acute:chronic workload ratio (ACWR) is this week's load divided by the rolling four-week average. The data converge on a sweet spot of 0.8–1.3:

• ACWR < 0.8: under-trained, deconditioning. Injury risk rises if you then jump back to your old load.
• ACWR 0.8–1.3: productive zone. Tissue tolerance keeps pace with stress.
• ACWR > 1.5: "spike". Injury risk over the following 1–2 weeks rises a lot.

The 2016 IOC consensus on load and athlete health endorsed this framework and added a critical nuance: chronic high training load is protective Soligard 2016. Trained tissue tolerates load that untrained tissue cannot. This is one of the most under-appreciated findings in modern sports science.

A practical weekly framework

Combine the pieces and you get a simple decision protocol most adults can apply in 2 minutes per workout:

1. Rate every session 1–10 RPE at the end. Multiply by minutes. Log it.
2. Sum the week. Track the rolling 4-week average.
3. Compute the ratio. This week's total ÷ 4-week average.
4. If < 0.8: you've under-trained for context (injury, illness, life). Re-anchor at the 4-week average rather than the previous high.
5. If 0.8–1.3: proceed. Add 5–10% next week if you're feeling good.
6. If > 1.5: deload this week. Drop volume by 30–40%, keep intensity moderate.
7. Every 4–6 weeks: a planned deload regardless of ACWR. Volume down 40–50%, intensity moderate, full week.

Periodisation in two paragraphs

Periodisation is the planned variation of training load over time. The classical "block periodisation" model alternates 3–4 weeks of progressive overload with 1 week of deload Issurin 2010. Hard, harder, hardest, easy. The deload week is when adaptation crystalises — it's not "lost time," it's where the supercompensation happens.

For recreational adults, you don't need a complicated mesocycle plan. The simplest framework that works: 3 weeks of progressive load (each week 5–10% harder than the last), one week of deload (40–50% volume drop, easy intensity). Repeat. After 3–4 cycles, take 7–10 days fully off training. Most people who follow this never hit overtraining.

When you've gone too far

Kreher and Schwartz's 2012 practical guide lists the syndromal pattern of overtraining Kreher 2012:

• Performance: drop in lifts you previously could hit, slower paces at the same RPE, can't finish workouts you used to handle.
• Mood: persistent low mood, irritability, loss of motivation specifically for training, "everything feels heavy."
• Sleep: trouble falling asleep despite physical fatigue, early waking, non-restorative sleep.
• Resting HR: persistent elevation 5–15 bpm above your normal baseline for 5+ days.
• Appetite: often suppressed.
• Frequent minor illness: upper-respiratory infections, slow wound healing.

One bad week is fatigue. Two bad weeks is overreaching. Three+ bad weeks with the cluster above is non-functional overreaching, and it warrants a planned 7–14 day reset — not "powering through."

Recovery is not optional

The 2018 IOC consensus on recovery and performance frames recovery not as the absence of training but as an active process with specific physiological windows Kellmann 2018. The hierarchy of evidence-supported recovery interventions:

• Sleep: the single most-supported recovery tool. 7–9 hours, dark room, regular schedule. See our sleep article.
• Adequate energy & protein intake: chronic under-fueling is the most common modifiable cause of stalled progress.
• Active recovery: light walking, easy cycling, mobility work between hard sessions.
• Planned deload weeks: 1 in every 4–6 weeks.
• Stress management: work and life stress count as training load — the body doesn't distinguish.

Beachside note

Beachside's class programming bakes in deload weeks every 4–6 weeks — coaches scale loads down deliberately so the supercompensation window can do its work. If you train solo, you'll need to apply that discipline yourself. (My family runs the gym; I host this site there. Disclosure.)

Common myths

"No pain, no gain." Pain is information, not a virtue. RPE 7–8 working sets do the bulk of the work; RPE 9–10 every session breaks people.

"If I'm not sore, I didn't train hard enough." Soreness is not a reliable indicator of training quality, as covered in our soreness vs injury article.

"Rest days are for the weak." Adaptation happens between sessions, not during them. The 2018 IOC consensus is unambiguous: planned recovery days are required for sustained progress, not optional.

"More volume always beats less." The dose vs response curve has a ceiling. Beyond ~10 weekly sets per muscle group, additional volume rarely adds and often subtracts (via accumulated fatigue).

Safety notes

• Stop a working set if form breaks down. "One more rep" is not worth a tweaked back or a torn rotator cuff.
• Never train through sharp, localised joint pain. See our medical red flags article.
• If your resting HR has been elevated for 5+ days, you're cooked. Take 3–5 days fully off and rebuild from a base.
• Don't try to "make up" missed training by spiking the next week. ACWR > 1.5 is where injuries happen.

The bottom line

• Use RPE every session. Most working sets at 7–8. Save 9–10 for finishers and occasional max attempts.
• Track session-RPE × minutes as a weekly load measure.
• Keep weekly load increases under 10–15%. ACWR > 1.5 raises injury risk a lot.
• Deload every 4–6 weeks. 40–50% volume drop, moderate intensity, one full week.
• Watch for the overreaching cluster: performance drop + mood + sleep + elevated resting HR + appetite changes. 3+ bad weeks = back off, not push through.
• Recovery is a training input, not a luxury. Sleep, fuel, deload, life stress all count.

References