RPE vs RIR: How to Rate a Set (and Let Your App Pick the Weight)

Most training plans tell you a weight and a rep count and leave out the variable that decides whether the set did anything: how close to failure you actually got. Two lifters can both do 5 × 100 kg. For one it is a warm-up; for the other it is everything they have. Same numbers on paper, completely different training effect.

RPE and RIR are how you write that missing variable down.

The two scales are the same scale

RIR — reps in reserve — is the plain-language version: how many more good reps could you have done? You stopped at eight, and you reckon you had two left. That is 2 RIR.

RPE — rate of perceived exertion — is the same information counted from the other end, on a scale where 10 is a set you could not have added a single rep to.

RPE	RIR	What it felt like
10	0	Nothing left. The last rep barely moved.
9	1	One more, maybe. Certainly not two.
8	2	Two solid reps left. Bar speed still honest.
7	3	Three left. Comfortable, controlled.
6	4	Four or more left. This is warm-up territory.

The conversion is just RIR = 10 − RPE. There is nothing to learn beyond that, so use whichever one you naturally think in. Some people find “two reps left” concrete and “RPE 8” abstract; others are the reverse. It genuinely does not matter.

What is worth knowing: below about RPE 6, the scale stops being useful. A set with five reps left is not a hard set you rated gently, it is a warm-up — and warm-ups do not need rating.

Why bother rating at all

Because your capacity is not constant, and a program written in advance assumes it is.

The plan says 100 kg for 3 × 8. On a good week — you slept, you ate, work was calm — that is RPE 7 and you left growth on the table. On a bad week it is RPE 10 on the second set and you grind out the third with a rounding back. Same prescription, two very different outcomes, and nothing in your logbook records the difference.

Rating a set turns that invisible variable into data. Over a few weeks it answers questions you cannot otherwise answer:

• Are my “hard” sets actually hard, or am I stopping three reps early out of habit?
• Is this program getting easier because I am stronger, or because I quietly stopped pushing?
• Did that weight jump land where I intended, or did it blow past it?

How to rate a set honestly

The single most common mistake is misjudging how close failure is — calling a set RPE 9 when there were genuinely three or four reps left. It is not dishonesty, it is inexperience: if you have never taken a set to failure, you do not have a reference point for where failure is.

Two things fix it.

Watch the bar, not the burn. Discomfort rises long before capability runs out. The reliable signal is speed: while reps are moving at roughly the same pace, you are further from failure than it feels. When a rep noticeably slows despite maximum intent, you are within one or two of the end. Discomfort lies; bar speed does not.

Take a few sets to actual failure — safely. Once or twice, on a leg extension, a lat pulldown, a dumbbell curl, keep going until a rep genuinely will not complete. Not on a back squat, not on a bench without pins. The point is calibration: you need to have met RPE 10 to recognise RPE 8.

Rate immediately, too — within a few seconds of racking. A minute into the rest period, memory has already softened it.

Where your sets should sit

For hypertrophy, the useful range for working sets is RPE 7 to 9 — one to three reps in reserve. Close enough to failure that the set counts, far enough that your technique holds and you can train the muscle again within the week. Going to RPE 10 on everything does not add much stimulus but adds a great deal of fatigue.

For strength work on heavy compounds, most quality sets live at RPE 7 to 8. You want the practice at a heavy load, not the grinding maximal single that costs you the rest of the session.

A sensible default: leave the last set of an exercise as your hardest, and keep the earlier sets a notch easier so you can actually complete them.

This is also where phases matter. A well-built block moves the target over time rather than holding it fixed — an accumulation phase around RPE 6–7, a harder phase at 7–8, a peak at 8–10, and a deload back down to 4–6 so you arrive at the next block recovered rather than merely surviving.

Turning a rating into your next weight

A rating you never act on is a diary entry. The point of the number is the decision that follows it.

The rule of thumb, if you are doing it by hand:

• Two or more RPE above target (planned 8, felt 10): drop the load 5–10% for the remaining sets. The set after a maximal one is rarely productive.
• Roughly on target: change nothing. This is most sets, and it should be.
• Two or more RPE below target (planned 8, felt 6): add 2.5–5%. You have room, and a week of easy sets is a week that does nothing.
• Consistently below target across a session: raise next week’s starting load instead of adding the same fixed jump you always add.

That last point is the one people miss. Standard linear progression adds a set amount every week regardless of evidence. Autoregulation replaces “add 2.5 kg because it is Tuesday” with “add what last week’s sets say you can carry” — which is both faster when you are progressing well and safer when you are not.

Doing it without the arithmetic

This is exactly the loop HealthBuddy now closes for you. After a set it asks how it felt — in plain words (plenty more / a few more / maybe one / nothing left), in reps left, or on the RPE scale, whichever vocabulary you prefer. It stores one number either way.

Then it uses the answer twice. Inside the session, if a set lands far enough from its target, the rest screen offers a lighter or heavier load for the sets still to come, already rounded to plates you can actually put on the bar. Between sessions, routines set to Adapt to how sets feel start next week from what happened rather than a fixed increment.

It proposes; you decide — and if you ignore the prompt a few times in a row, it stops asking for that session. Effort rating is free, it works on Apple Watch as well as iPhone, and by default it only asks on your last set of an exercise, which is where the information is worth most.

The short version

RPE and RIR are one scale with two names, and RIR = 10 − RPE. Keep working sets between one and three reps in reserve. Judge by bar speed rather than discomfort, and calibrate by occasionally meeting failure somewhere safe. Then actually use the rating: adjust the sets after it, and let next week start from what your body did rather than what the spreadsheet assumed.

Related reading: progressive overload, explained, how many sets and reps to build muscle, and how to break through a plateau.