Why we use RPE instead of percentages of your 1RM

Key points

• Most programmes prescribe weight as a percentage of your 1RM. We use RPE instead.
• The reason is daily variability. Your real strength on any given day moves around by roughly 5 to 18 percent based on sleep, stress, food, and recovery. A fixed percentage ignores all of that.
• RPE, measured by Reps in Reserve, lets you read the bar today and load it correctly today. It works better than a percentage for trained lifters, and it self corrects when life is messy.
• Percentages are not useless. They are good for testing and peaking. They are a bad fit for week to week training when you actually want to grow.

Why this matters

If you are following a programme that says “3 sets of 5 at 80 percent of your 1RM,” you are running into a problem the programme is hiding from you. Your 1RM was tested on a specific day, in a specific gym, with a specific amount of sleep and food behind it. Today might not be that day. Maybe last night was bad. Maybe you missed lunch. Maybe you are six weeks into a programme and that 1RM number is now stale.

The percentage does not know any of that. It will tell you to load 80 percent and grind. If you are fresh, that 80 percent will feel like 75. If you are tired, it will feel like 90. Either way the prescription is off.

RPE, short for Rate of Perceived Exertion and measured in Reps in Reserve (RIR), is the fix. Instead of saying “load 80 percent,” the programme says “find the weight that leaves you with 2 reps in the tank for 3 sets of 5.” Today’s weight comes from today’s body, not last month’s test.

This is autoregulation, and the research has been pointing this way for over a decade.

The RPE scale

RPE is rooted in how many reps you could have done past the set. That number is RIR. RPE 10 means zero reps in reserve, full failure. RPE 8 means you stopped with 2 reps left.

RPE	RIR	What it feels like
10	0	Absolute failure. Could not do another rep.
9.5	0 to 1	Failed on the last attempt, or maybe one more grindy rep.
9	1	Could do one more rep if your life depended on it.
8.5	1 to 2	Definitely one more, maybe two.
8	2	Could do two more reps with good form.
7.5	2 to 3	Definitely two more, maybe three.
7	3	Could do three more reps. Starting to feel effort.
6	4	Moderate. Warmup plus territory.
5	5+	Light. Could keep going a long time.

Most working sets in our programmes target RPE 7 to 9. Below 7 is warmup or junk volume. Above 9 is grinding for diminishing return on most lifts. Above 10 does not exist, despite what some lifters say at the bar.

Why fixed percentages break for trained lifters

Daily strength is not a fixed number

The strength you can express today is not a constant. Tracking studies of trained lifters have found typical fluctuations of 5 to 18 percent across days, driven by sleep, nutrition, prior training, life stress, and recovery state (Helms et al. 2016, and the broader review in the Muscle and Strength Pyramids 3rd edition). That is a wide band. A 100 kg squat with 10 percent fluctuation is anywhere from 90 to 110 kg of “true” effort on a given day.

A percentage based programme treats your 1RM as a constant. It gives you the same load every Tuesday, regardless of how your Tuesday actually went. That is fine if you happen to be average on most Tuesdays. It is wasteful if you are above average (you trained too easy) and dangerous if you are below average (you grind through what should have been impossible and crash).

1RM tests get stale fast

Even a perfect 1RM has a half life. As you train, your 1RM moves. Most programmes test once at the start of an 8 to 12 week block and prescribe percentages off that single number for the whole block. By week 6 the number is stale. The percentages start under or over loading depending on how progress actually went.

The honest fix is to retest every couple of weeks. The practical fix is to stop using a single number as the basis of every set, and read the bar in real time instead.

Beginners can use percentages. Trained lifters cannot

For a brand new lifter, percentages work fine because the variability is low and the absolute weights are small. The error band of “load 80 percent” is small enough not to matter. Once your training age clocks past a year and your weights are heavy enough that 5 percent is the difference between a clean set and a missed set, that error band starts costing you.

This is the population the research is talking about. Helms, Zourdos, and Graham have all done their RPE validation work on trained lifters. RPE accuracy is much higher in this group, which is also the group that needs autoregulation the most.

How accurate is RPE actually

The usual criticism of RPE is that it is “subjective.” It is subjective in the same sense that pace is subjective for a runner. You can learn it. The question is how accurate the learning gets.

Zourdos and colleagues (2021) tested trained lifters across a range of intensities. The findings:

• At RPE 8 to 10, lifters estimated their RIR within 1 to 2 reps. That is more than accurate enough for programming.
• At RPE 6 to 7, lifters tended to underestimate how many reps they had left. When sets feel easy, you usually have even more reps in the tank than you think.
• Accuracy improved with experience. Newer lifters were less accurate, especially at high intensities.

The practical takeaway: at the working set RPE we actually use (7 to 9), trained lifters are accurate enough that the system works. At lower intensities, the bias is in a safe direction. You have more reps than you thought, not fewer.

The practical takeaway: how to use RPE in a session

How to pick the weight

Read the programme. It will say something like “3 sets of 8 to 12 at RPE 8.” Your job is to find the weight today that lets you finish 3 sets of at least 8 reps, with the last rep of each set leaving 2 in the tank.

A few warmup sets get you in the ballpark. Work upward. The first working set should feel like RPE 8 at the rep target. If it felt easier, the next set goes up. If it felt harder, the next set goes down.

Calibrate occasionally

Once every few weeks, on a single isolation movement, do an AMRAP set (as many reps as possible). Pick a weight where you estimated RPE 8 at rep 8. Then keep going to true failure. If you got 4 more reps, your RPE 8 is actually RPE 6. Recalibrate. If you got 1 more, your RPE 8 is closer to 9. Also recalibrate.

This is the simplest way to keep your RPE estimates honest. Once or twice a mesocycle is enough.

Trust the bar over the spreadsheet

If a weight that the programme says you should hit at RPE 8 is feeling like RPE 9 or 10, the programme is wrong, not you. Drop the weight, hit the rep target at RPE 8, log the difference. The next session you may go back up. Or you may stay. Both are fine. The point is you trained at the right intensity for today.

Common calibration mistakes

Hero RPE in week one

A common pattern: the lifter starts a mesocycle, hits RPE 8 in week one as prescribed, and chases the load up too fast because the weights felt manageable. By week three the same lift is RPE 9.5 and they are grinding. Week four cooks them. The mesocycle was supposed to ramp into intensity over six weeks. It ramped into intensity in three.

The fix: in week one, your RPE estimates should err on the easy side. RPE 7 in week one is fine. The programme builds intensity over time on purpose. Not every week is a peak week.

Chronic underestimation

The opposite problem: lifters who are conservative on every set, never push past RPE 7, and wonder why they are not progressing. The research is clear that proximity to failure matters. Sets at RPE 7 and above drive most of the adaptation. Sets below 7 are mostly warmup.

The fix: trust the prescription. If the programme says RPE 8, get to RPE 8. The AMRAP calibration above will tell you whether your RPE 8 is real.

Treating RPE 10 as a target

True failure is rarely the goal. RPE 10 burns recovery without proportionally driving more growth (Carroll et al. 2019). It also makes RPE estimates less accurate for the rest of the session. We use RPE 10 occasionally on isolation work, almost never on heavy compounds. If you find yourself hitting RPE 10 on every working set, you are leaving recovery on the table for no extra benefit.

When percentages still make sense

Percentages are not useless. They are useful in three specific contexts:

Testing. A 1RM attempt is, by definition, a percentage of itself. If you want to know what you can lift today, you test today. The result is the percentage anchor for the next short block.

Peaking. In the last two to four weeks before a powerlifting meet, attempts are programmed as a percentage of the projected meet 1RM. You want to rehearse the exact weight you will open and second attempt with. RPE works here too, but percentages lock the rehearsal to specific numbers.

Beginners on linear progression. A novice doing 3 by 5 with linear progression is fine on percentages or just fixed weights. The variability is too low to matter. Once you are past that point, autoregulation pays off.

For everything else, RPE wins. That is the common case for the people reading this blog.

What to do this week

If you are running a percentage based programme right now, you do not need to throw it out. The simplest experiment: for one week, log RPE on every working set. Just write a number from 6 to 10 next to the reps. At the end of the week, look at the spread. If your RPEs are all over the place at the same percentages, that is the variability the percentage programme is hiding from you.

The next week, run the same programme but adjust the weight up or down 5 to 10 percent based on what RPE you actually want, instead of what the percentage says. You are now autoregulating.

This is also where RPE connects to recovery decisions. If your RPE has been creeping up for the same weights and reps over three weeks, that is a fatigue signal. We cover that pattern in detail in our piece on why concurrent training interferes, where the same RPE inflation shows up when running volume is bleeding into lifting capacity. The other half of the picture is volume itself, which we cover in volume landmarks for hybrid athletes.

Loonstep prescribes RPE based programming by default. You log the set, log the RPE, and the next session updates the target weight based on whether you progressed, held, or regressed at the prescribed intensity. Sign up and let the coach handle the autoregulation math while you focus on the bar.

References

1. Helms, E. R., Cronin, J., Storey, A., & Zourdos, M. C. (2016). Application of the repetitions in reserve based rating of perceived exertion scale for resistance training. Strength & Conditioning Journal, 38(4), 42 to 49.
2. Zourdos, M. C., Goldsmith, J. A., Helms, E. R., Trepeck, C., Halle, J. L., Mendez, K. M., Cooke, D. M., Haischer, M. H., Sousa, C. A., Klemp, A., & Byrnes, R. K. (2021). Proximity to failure and total repetitions completed in a set influences accuracy of intra set RPE. Journal of Strength and Conditioning Research, 35 (Supplement 1), S158 to S165.
3. Graham, T., & Cleather, D. J. (2021). Autoregulation by “repetitions in reserve” leads to greater improvements in strength over a 12 week training program than fixed loading. Journal of Strength and Conditioning Research, 35(9), 2451 to 2456.