RPE vs. Percent of 1RM: Which Actually Builds Strength Faster?
I ran percentage-based programs for eight years before I touched RPE. Then I made more progress in nine months than the prior two years combined. Here is what changed.
RPE vs. Percent of 1RM: Which Actually Builds Strength Faster?
I ran 5/3/1 and its variants for almost eight straight years. Percentage-based, by the book, never autoregulated a damn thing. My logbook was beautiful and my squat went from 315 to 405 over those years, which sounds fine on paper but is genuinely terrible progress for the amount of work I put in. The problem wasn't the program. The problem was that on bad-sleep weeks I was grinding into 10/10 sets at "80%," and on great weeks I was leaving easy reps on the platform because the spreadsheet said today was a 70% day.
Then I tried an RPE-capped block — Bryce Lewis's 28 Free Programs — for one mesocycle. Within nine months I'd hit 450. More importantly, I stopped getting hurt every few months.
This isn't an argument against percentages. It's an argument for using both correctly.
What forced the field to evolve
Two facts that, once you accept them, kill pure-percentage programming:
1. Your 1RM drifts daily. Mine moves ±8-12% based on sleep, food, mood, life stress. A "fixed 82%" on a bad day is closer to 90% effort. On a great day, it's 72%. You're never actually training where the spreadsheet thinks you are.
2. Strength gains track proximity to failure, not absolute load. The Grgic et al. 2022 meta-analysis showed RIR (reps-in-reserve — the inverse of RPE) predicts both hypertrophy and strength better than absolute weight on the bar.
If those two things are true, then a system that adjusts load based on how the work actually feels today will beat a fixed system on the same volume.
The case for percentage-based programming
Before I make this sound like RPE is the second coming, the honest case for percentages:
- Predictable progression curves. You know exactly what you're doing for the next 12 weeks. Some athletes need that for adherence — myself included, early on.
- No skill required. Beginners genuinely cannot rate RPE accurately for the first 6-12 months. I sure couldn't. My "RPE 8" in year one of lifting was somebody else's RPE 6.
- Simpler for coaches with many athletes. Easier to write one program for 40 lifters than 40 autoregulated ones.
If you're under two years of training age, percentage-based with conservative jumps is probably better. The autoregulation tax is real until you've built the calibration.
The case for RPE (what convinced me)
- Self-regulates for sleep, stress, sickness. Bad day at work? Your RPE 8 today is naturally lighter. You still get the training stimulus without the injury risk.
- Stops you from grinding RPE 10s on bad days. This was the single biggest change for me. Three of the four times I've tweaked my back in the last decade were percentage-based "80% triples" on days I had no business loading the bar that heavy.
- Honest measurement of adaptation. When the weight I rated RPE 8 last week is now RPE 6 at the same load, I got stronger. The bar speed and the RPE both tell me. The spreadsheet doesn't.
The hybrid that actually wins
Almost every smart modern program (Renaissance Periodization, Juggernaut AI, SBS, RTS, and the defaults in Snap My Gym) uses the same hybrid:
> Percent-of-1RM as a target, RPE as the governor.
In practice this reads: "Squat 5x3 at 82.5%, RPE 8 cap." Three rules I follow:
- If 82.5% feels like RPE 9 today, drop the next set 5%. You're already past the cap.
- If it feels like RPE 6, you stay at the prescribed load — you don't have to push to failure to make progress. Volume at submaximal RPE compounds.
- If you hit the prescribed load at RPE 7 for three sessions in a row, your true 1RM has moved up and the spreadsheet needs to catch up.
Practical RPE primer
This is the rubric I drilled into myself, taped to my squat rack for the first six weeks:
| RPE | What it feels like |
|-----|-------------------|
| 10 | No reps left, form may break. You looked like you were dying. |
| 9 | One more rep maybe. The bar slowed visibly on the last rep. |
| 8 | Two more reps in the tank. Last rep felt heavy but bar speed was steady. |
| 7 | Three more reps, bar speed still fast. Most of your working sets should live here. |
| 6 | Easy — warm-up territory. Should not be a "working" set above this load. |
How to learn the calibration faster
This is the part nobody tells beginners. Three things that compressed the learning curve for me:
1. Film every working set for 4 weeks. Watch the bar speed on rep 1 vs. the last rep. The visible deceleration between those is your RPE.
2. Log a "next rep prediction" before racking. Before you rack the bar, write down "I had X more reps." Then test it on the last rep of a top set every couple of weeks. You'll be wildly off at first. By week 8 you'll be within one.
3. Use a velocity tracker for one mesocycle if you can. Even a cheap one ($120-200). The Vitruve / Enode / Beast units calibrate your "feel" against objective bar speed, and 6-8 weeks of that data builds an honest internal model.
If you can't tell the difference between RPE 7 and RPE 9 yet, you're normal. Log every set with effort notes for 8 weeks. The calibration arrives.