Two lifters both log 3×8 @ 100 kg. The first could have done twelve reps on every set. The second failed on the last rep of the last set and needed a spotter. Their logs are identical. Their training weeks are not remotely comparable.
That gap — between what you lifted and what it cost you — is what RPE and RIR exist to close. It is the single most useful number that most people do not record.
The two scales
RIR — Reps In Reserve. How many more reps could you have done before technical failure? Finish a set with two more in you, that is 2 RIR.
RPE — Rate of Perceived Exertion. A 1–10 scale of how hard the set felt, where 10 means you could not have done another rep.
In strength training as it is practiced today, these are the same number wearing different clothes:
| RPE | RIR | What it feels like |
|---|---|---|
| 10 | 0 | Could not have done another rep. Bar slowed to a crawl. |
| 9.5 | 0–1 | Maybe one more, would not bet on it. |
| 9 | 1 | One more rep, definitely. Two, no. |
| 8 | 2 | Two solid reps left. Bar speed clearly slowing on the last rep. |
| 7 | 3 | Three left. Fast, controlled, comfortably hard. |
| 6 or below | 4+ | Warm-up or technique work territory. |
So: RPE = 10 − RIR. Pick whichever your brain likes. Beginners generally find RIR more concrete — counting reps is easier than rating a feeling. Experienced lifters tend to drift to RPE because it accommodates the half-points and the days when a set is hard for reasons that have nothing to do with the load.
The accuracy problem nobody warns you about
Here is the uncomfortable finding from the research: novice lifters are bad at this, and they are bad in a predictable direction. They systematically underrate. A set a beginner calls "RPE 9, one rep left" is very often RPE 6 with four reps left.
This is not a character flaw. If you have never taken a set to actual failure, you have no reference point for what one rep from failure feels like. You are calibrating against nothing.
Two fixes:
- Take a few sets to true failure, deliberately, on a safe exercise (machine, cable, dumbbell press with a safe bail-out). Not as a training method — as a calibration exercise. You need to have felt a 10 to be able to recognize an 8.
- Anchor on bar speed, not on suffering. Discomfort lies; the bar does not. When the concentric slows noticeably, you are within a couple of reps of failure regardless of how heroic you feel.
What to actually train at
The useful range is narrower than people think.
- Hypertrophy: RPE 7–9 (1–3 RIR) for most working sets. Far from failure means you leave stimulus on the table. Constant failure means fatigue accumulates faster than the stimulus that justifies it — and fatigue is what forces the deload.
- Strength: RPE 6–8 on the main lifts for most of a block, with heavier top singles/doubles rising to RPE 8–9 near the end of a cycle. Technique degrades near failure, and grinding heavy singles is a large fatigue bill for a small adaptation.
- Failure (RPE 10): occasionally, on isolation work, near the end of a block. Not on squats. Not every session. It is a tool, not a lifestyle.
The real payoff: autoregulation
RPE's value is not in describing the past. It is in changing the plan on the day.
Your program says 3×5 at 120 kg. You slept five hours, you are in week nine of a cut, and the first set moves like 140. Under a fixed program you grind through it, dig a deeper hole, and probably miss reps in session two. Under an RPE cap — "3×5, stop if it exceeds RPE 8" — you drop to 112.5 kg, get the quality work, and keep the week intact.
That is the whole idea. The program sets the intent; RPE sets the dose. Fixed loads assume every day is average. No day is average.
Where the signal gets ignored
Most workout apps let you log RPE and then do absolutely nothing with it. It lands in a text field, gets a nice chart, and never influences a single recommendation. Which is a shame, because a rising RPE trend at fixed load is one of the earliest, clearest fatigue signals you can collect. It typically shows up before your lifts start regressing — it is the warning, not the crash.
And it becomes far more informative when it is read next to something most workout apps cannot see: how much you have been eating. RPE climbing at fixed load during a maintenance phase means fatigue — pull volume, consider a deload. The same RPE climb in week eight of a deficit means something different: it is the predictable cost of the cut, and the right move is usually to trim volume and hold the loads, not to panic about the program.
Same number. Opposite interpretations. The only thing that distinguishes them is nutrition data — which is exactly why Gymrock reads your logged sets, your RPE, and your food in one model rather than three disconnected ones. The score is only worth collecting if something is going to read it.
The short version
- RPE = 10 − RIR. Use either. Just use one.
- Expect to underrate at first. Calibrate by taking a few safe sets to real failure.
- Live at RPE 7–9 for hypertrophy, 6–8 for most strength work.
- Watch bar speed, not suffering.
- Use it to adjust the day's load — not just to decorate your log.