RPE in strength training: how to score it, and how to tell you need a deload

Two RPE scales exist and they are easy to confuse. The original Borg scale runs 6–20 and was designed for aerobic work. The one that matters for lifting is the resistance-training-specific scale published by Zourdos and colleagues in 2016, which anchors each RPE value to the number of repetitions you had left in the tank — reps in reserve, or RIR.

That anchoring is what makes it useful. RPE 8 is not a vague feeling of "fairly hard"; it means you stopped with exactly two reps still available.

Load tells you what was on the bar. RPE tells you what it cost you. Those are different quantities, and the gap between them is exactly what you want to watch.

If you squat 100 kg for 5×5 at RPE 7 in week one, and the same 100 kg for 5×5 is RPE 9 in week four, your capacity has dropped. Nothing about the program changed — same weight, same sets, same reps — but the cost went up. That is accumulated fatigue, and it is invisible if you only log tonnage.

This is the single most common blind spot in self-tracking. Volume curves look stable while performance is quietly degrading, because volume measures work done, not work available.

First, score the set immediately after you rack it, not at the end of the session. Memory reconstructs badly, and a session you remember as "hard" may have had three easy sets and one grind.

Second, be consistent about what you are rating. Score the set, not the exercise, and use the same definition of failure every time — technical failure (form breaks) is more useful than absolute failure, because it is what you would actually stop at in training.

Third, only compare RPE at the same load and the same rep range. RPE 8 on a set of 3 and RPE 8 on a set of 12 are not comparable numbers, and averaging them produces nonsense.

One high-RPE session means nothing — sleep, caffeine and stress all move it by a point or more. What matters is the direction over three to four weeks at comparable loads.

A practical rule: if average RPE at a given load rises by about one point and stays there for two consecutive weeks, cut volume by roughly 30–40% for one week, or take a deload. If it keeps climbing after that, the problem is usually outside the gym — sleep, food, or life stress — and adding more training will not fix it.

Conversely, if RPE at the same load is falling, you are adapting and can add load or volume. That is the green light most people miss, because they add weight on a fixed schedule instead of on evidence.

Novices are unreliable at it. The 2016 Zourdos work and the literature since both show accuracy improves with training experience; a beginner's RPE 8 might really be a 6 or a 10. For the first few months, treat it as a rough note rather than a decision input.

It is also easy to anchor on the previous week. If you rated 8 last time, there is a pull to rate 8 again. Score the set in front of you, then compare.

And it does not replace load progression. RPE is a regulator, not a program. You still need a plan that says what to do when RPE is low.

Per set: load, reps, and RPE. Per session: total sets per muscle group and duration. That is enough to compute a usable trend — and it is what Vitalis uses to draw the RPE line next to your volume curve.

If you only record one thing beyond load and reps, record RPE. It is the cheapest signal you can collect, and the only one that tells you how the work felt rather than how much of it there was.