RPE: the simple scale that makes autoregulation work

TL;DR: RPE is a 1 to 10 scale that rates how hard a set felt. Pair it with a target (e.g. "3 sets of 8 at RPE 8") and your program stops being a fixed script. It adjusts to your real capacity that day (sleep, stress, soreness, yesterday's legs) without you having to do the maths.

Imagine you wrote a 5×5 program six weeks ago. Today it says: squat 102.5kg for 5 reps, five sets. You slept four hours because the dog wouldn't settle. You're halfway through a brutal work week. Your hamstrings are still angry from Monday's deadlifts.

What does the program say? Squat 102.5kg for 5 reps, five sets.

This is the core problem with fixed-percentage programming. The plan was written by a version of you that hadn't lived the past six days. It has no idea you're running on caffeine and hope. So you either grind through and leave the gym wrecked, or you bail on the session and feel like you "failed" a day that was never honestly programmable.

RPE autoregulation is the fix. It's not a new idea (powerlifters have used it for decades) but it's quietly become the default language of modern strength and hypertrophy coaching. And once you get it, you won't want to program any other way.

The RPE scale (and its sibling, RIR)

RPE stands for Rate of Perceived Exertion. The lifting version was formalised by powerlifting coach Mike Tuchscherer in the late 2000s, adapted from Gunnar Borg's original medical scale. Tuchscherer's insight was simple: instead of chasing percentages of a 1RM you took months ago, rate each set on how close it felt to failure. That rating becomes the load signal for the next set, the next session, the next week.

Here's the scale most coaches use today:

RPE	What it feels like	Reps in reserve
10	Absolute max. Bar slows to a crawl, could not have done another rep.	0
9.5	Maybe one more, but bar speed suggests you'd fail it.	0 to 1
9	One solid rep left in the tank.	1
8.5	One clean rep, maybe a grindy second.	1 to 2
8	Two reps in reserve. Challenging but controlled.	2
7	Three reps left. Heavy warm-up feel.	3
6	Four or more reps left. Bar moves fast throughout.	4+

Below RPE 6 you're in warm-up or recovery territory: the weight's moving so fast that rating effort doesn't really mean anything useful.

RIR (Reps In Reserve) is the same information stated in the opposite direction. RIR 2 means you had two reps left, which is the same as RPE 8 (assuming the set was in the 1 to 10 rep range). Many lifters find RIR more intuitive because you're directly estimating reps, not decoding a feeling into a number. "Could I have done two more? Yeah, two." Done.

The two scales are functionally equivalent, and you'll see both in modern programs. Eric Helms uses RIR extensively in his nutrition and hypertrophy work. Renaissance Periodization popularised RIR in bodybuilding circles. Tuchscherer's original Reactive Training Systems still uses RPE. Pick whichever your brain prefers. The underlying mechanic is identical: you are teaching yourself to rate proximity to failure, and using that rating to drive load decisions.

A small but important point: RPE is about effort at a specific rep range, not absolute weight. A set of 5 at RPE 8 is a much heavier load than a set of 15 at RPE 8, but both left you with 2 reps in the tank. The scale normalises across rep ranges, which is precisely why it's so useful.

Why autoregulation beats fixed percentages

Fixed-percentage programming says: "Week 3, Day 1, squat: 82.5% of your 1RM for 5 reps." That looks precise. It is, in fact, guesswork wearing a lab coat.

The problem is the assumptions buried in that percentage. It assumes your 1RM is what you last tested. It assumes your readiness this Monday is identical to the Monday you tested. It assumes no cumulative fatigue, no sleep debt, no life. Your true 1RM drifts day to day by as much as 18%, which is the finding from a 2016 study by Zourdos and colleagues, who compared prescribed percentages to what lifters could actually hit when asked to rate RPE.

Greg Nuckols has written extensively about this at Stronger By Science. His summary: fixed percentages work fine when you're fresh and progressing predictably. They fall apart the moment life gets messy, which, for most of us, is most of the time. Autoregulation is what keeps the programme's intent intact when the specific numbers can't hold.

Eric Helms and his co-authors put RPE-based training to the test in a 2018 study on resistance-trained lifters. One group used a fixed percentage scheme; the other used RPE-driven loads to hit the same target stimulus. Strength gains were comparable, but the RPE group had markedly lower session-to-session variability in perceived difficulty, meaning they were getting a more consistent training stimulus even though the weights on the bar moved around. That consistency is the whole point.

Here's what fixed percentages actually do to a week in your life:

• Great day, programmed weight feels light. You move the weight but leave gains on the table; the stimulus was too low for the reps.
• Rough day, programmed weight feels like a 1RM. You grind, form breaks down, and you take on fatigue you'll pay for later in the week.
• Average day. The weight is roughly right. Congratulations.

One in three, maybe. RPE flips the ratio. You still have a target (say, 3×8 at RPE 8). On a great day that target might be 105kg. On a rough day it might be 92.5kg. Both sessions delivered the same stimulus (8 reps with 2 in the tank) even though the numbers on the bar tell different stories. The programme is adapting to you, in real time, without requiring you to ever email your coach.

The catch: autoregulation only works if your rating is honest. Which brings us to the hard part.

How to rate a set accurately

New RPE users are terrible at rating sets. That's not an insult; it's a skill, and skills take reps. The good news is the skill builds fast if you know where you tend to lie to yourself.

The two classic errors:

1. Underrating when you're strong. The weight moved well, the reps felt crisp, and you tick the box as "RPE 7." In reality, the bar slowed on the last rep, your lockout was slower than the first, and you had maybe one more in you: that was an RPE 8 or 9. Strong days feel easier than they are, and you'll blow up the next session because your "RPE 7" on paper was actually a near-max.

2. Overrating when you're tired. Everything feels heavy. You grind out a set and call it RPE 10. In truth, you had a rep or two left; the set just felt maximal because your nervous system was cooked. You'll drop weight next session you didn't need to drop, and the programme stalls.

The fix is to stop rating by feel and start rating by reps. After the set, ask: "If someone had a gun to my head, how many more reps could I have completed with the same form?" Count those reps. That's your RIR; flip it to RPE. Feel is noisy. Rep count is signal.

Other tips that help:

• Watch the last rep. If bar speed dropped dramatically, you were near failure regardless of how the earlier reps felt. That's an RPE 9+. Film a couple of sets from the side. The bar speed gap between rep 1 and rep 5 is a brutally honest coach.
• Rate within 30 seconds. Memory warps fast. Rate right after re-racking, not after the rest period.
• Use a range early on. "RPE 8 or 9" is fine. Precision comes with practice.
• Occasionally train to failure. You can't calibrate RIR 0 if you've never hit it. A true max set every 4 to 6 weeks resets your ruler.

Within 8 to 12 weeks of consistent rating, most lifters can nail sets within half an RPE point. That's precise enough to run almost any programme.

The simplified Grov version: 3 buttons

Here's the thing: most gym-goers aren't competitive powerlifters. They don't need half-point RPE precision. They need a programme that responds when things go well or badly, without a spreadsheet.

So in Grov, we collapsed the scale to three buttons.

After each work set, you tap one:

• Missed: you didn't hit the target reps. (Equivalent to RPE 10+ / RIR -1.)
• Exactly: you hit the target, maybe with one in the tank. (Equivalent to RPE 8.5 to 9.5 / RIR 0 to 1.)
• Beat: you hit the target and clearly had two or more reps left. (Equivalent to RPE ≤8 / RIR 2+.)

Three buttons. That's the whole rating system. And it works for a specific reason: most autoregulation decisions are categorical, not continuous. The programme doesn't actually need to know whether you were at RPE 8.5 versus 9.0. It needs to know: was that too heavy, about right, or too light? Three buttons answers that question.

The behavioural upside is huge. Choice paralysis is real: every additional option on a post-set rating screen adds friction, and friction is where logging habits die. A 10-point scale looks precise; a 3-point scale actually gets used. And data you collected is infinitely more valuable than data you didn't.

For lifters who want more granularity, full RPE is always there in the notes. But the default, the thing that drives the programme's decisions, is three buttons. That's enough.

What changes based on your rating

This is the part that closes the loop. Your rating doesn't just sit in a log; it drives the next session's load directly.

Here's the simple version of what Grov does after a set:

• Missed: the programme drops the next session's working weight by roughly 5%. Missed reps mean the weight was out of range for today's capacity. Back it off, rebuild, try again with a cleaner stimulus.
• Exactly: the programme holds weight next session. You hit the prescribed reps with the prescribed effort. That's the sweet spot. Stay, consolidate, accumulate volume.
• Beat: the programme adds 2.5 to 5kg (compound lifts) or 1 to 2.5kg (isolations) next session. You had more in the tank, so the stimulus needs to go up to keep driving adaptation.

Those adjustments are small on purpose. RPE-driven progression is meant to be slow and compounding, not heroic week-to-week jumps. A 2.5kg bump every fortnight on your squat is a 65kg year-over-year shift. Over two years that's transformational, and it happened without a single grinding session that wrecked your back.

Load is half the autoregulation story. Volume is the other half: how many sets you do across a week, how that volume is distributed, and when to push it up versus pull it back. We cover the volume side in detail in our piece on sets per muscle per week, which pairs directly with the RPE system you just read about. Load tells each set how heavy to go; volume tells the week how much work to prescribe. Together they make the programme genuinely responsive.

One last note. Autoregulation is not permission to sandbag. If you Beat every set by dialling in at RPE 6, the programme will chase you upward and the weights will catch you eventually, usually in the form of a very humbling session. Rate honestly. The system is only as good as the input you give it.

---

You don't need a coach, a spreadsheet, or a 1RM test to autoregulate. You need a scale, a target, and the willingness to tell the truth about a set you just finished. RPE gives you the scale. Grov gives you the target, the three buttons, and handles the weight math in the background, so every session adapts to the version of you that actually walked into the gym today, not the version who planned the programme six weeks ago.

That's the whole idea behind our thesis: a programme should listen. And if you want to see how RPE-driven load fits inside a full weekly plan with pair structure and volume management, read the complete workout program next.

Start training →