RPE and reps in reserve: how to judge a set honestly

Your programme says “top set of five at RPE 8”. You load the bar, get the five, rack it — and honestly you have no idea whether two reps were left or five. The hard part of RPE has never been the numbers. It is learning to read the last two reps of a set.

What the number actually means

In the gym, RPE (rating of perceived exertion) is used in a stripped-down form that measures one thing: how many more reps you could have done at the moment you stopped. That is RIR — reps in reserve. The maths is simple: RIR = 10 − RPE.

• RPE 10 — nothing left; the next rep would have stalled halfway.
• RPE 9 — one rep in reserve.
• RPE 8 — two in reserve.
• RPE 7 — three in reserve, and the bar is still moving quickly.
• RPE 6 and below — four or more; warm-ups and technique work.

Half points are worth using. RPE 9.5 means “maybe one more, but I wouldn't bet on it”. Most training sits between 6 and 10, because below that the difference between five and seven reps in reserve is too vague to call.

What a set with one rep left looks like

The most reliable signal is speed, not burn. On compound lifts above roughly 80% of your best, reps move at a similar pace until the last two or three, when the lifting phase slows visibly. If your final rep took about twice as long as the first and stuck at the hardest point, you were at RPE 9 or 10.

The second signal is technique coming apart. Hips shooting up ahead of the shoulders in a squat, the bar tipping on a bench press, the lower back rounding on a deadlift — that is not “two more left”, that is the end of the set. A rep rescued by bad form does not count as reserve.

Burning, breathlessness and general discomfort are poor guides. A set of 20 squats hurts far more than a triple at 90%, and yet failure is often another five reps away.

Most people stop sooner than they think

Research generally points the same way: when lifters say one rep was left, they tend to have had two to four. The error is largest in beginners, in higher-rep sets, and on lifts they rarely train. Experienced lifters judging low-rep sets on a familiar movement are reasonably accurate. Reading the scale is a skill, and it sharpens with practice.

Calibrating is straightforward. Pick a lift where failure is safe — leg press, a machine, dumbbells, a cable. Use a load good for eight to twelve reps. Run the set and make a mental note of the rep where you thought “two left”. Then keep going to genuine failure and compare the two numbers.

One or two of these sets a week is plenty, done at the end of a session so they do not spoil your main work, and repeated every month or so. The same honesty pays off elsewhere: estimating a 1RM from working sets is only as accurate as your reps-in-reserve call.

Putting RPE into a programme

The most useful format is one heavier set followed by lighter ones. For a squat day: a single set of three at RPE 8, then three sets of five with about 10% less on the bar. The top set sets the tone; the back-offs carry the volume.

Across a four-week block the target climbs — RPE 7 in week one, 8 in week two, 9 in week three, then a lighter week back at 6 to 7. You do not have to write the weight in advance. You write the RPE and let the kilos follow.

For muscle growth, keep most sets between RPE 7 and 9, which is one to three reps in reserve. For technique and speed work, sit at 6 to 7. Save true failure for isolation exercises and use it sparingly; it costs recovery and adds little over a hard nine. How you spread that effort also depends on whether you are chasing more reps or heavier loads.

The real advantage is autoregulation. If 100 kg for five felt like RPE 7 last week and RPE 9 today, the weight stays put or comes down. Chasing a number written a month ago regardless of how the bar moves is one of the quickest routes to a stalled squat.

Where RPE stops working

The scale is at its best between three and eight reps. Past fifteen, perceived effort climbs much faster than actual proximity to failure and the estimate drifts; there you are better off training to a clear slowdown or a fixed rep target.

Deadlifts are their own problem. A set depends on grip, breathing and position as much as on hips and back, and bar speed drops from the first rep onwards. Plenty of lifters cap deadlift sets at RPE 8 for exactly that reason, which fits with how deadlift plateaus usually get broken.

Two more limits. RPE describes a single set, not the fatigue of a whole session — five sets at RPE 8 will drain you more than one at RPE 9. And mood, caffeine and a noisy gym change how hard something feels more than they change what you can do, so on strange days trust bar speed over the feeling.

When to see a doctor

RPE is a tool for dosing effort, not for tolerating pain. Sharp pain during a rep, tingling or numbness in an arm or leg, a sudden loss of strength, or soreness that lasts beyond a few days or turns up away from training are reasons for a medical opinion, not for a lower number on the scale.

Chest pain, heavy pressure, dizziness or fainting under load need urgent care. If you have high blood pressure or heart disease, or you are coming back from surgery or a significant injury, agree with your doctor how close to failure you should train before you start hunting nines.

The short version

• RIR = 10 − RPE: a 9 means one rep left, an 8 means two, a 7 means three.
• Judge by the speed of the last rep and by form breaking down, not by how much it burns.
• Most lifters overestimate effort by two to four reps; calibrate with one weekly set to true failure on a safe exercise.
• Keep most training between RPE 7 and 9 and reserve genuine failure for isolation work.
• The scale is unreliable above fifteen reps and on deadlifts — use fixed rep targets there instead.