RPE vs. Percentages: Which Is Better for Strength Training?

80 % is objective. RPE 8 is subjective.

From that, a simple conclusion suggests itself: percentage training is precise, RPE is feel.

It is not that simple.

Because 80 % of what? Of a 1RM that was perhaps tested six weeks ago. And RPE 8 how exactly? A judgement that can be remarkably usable depending on the load and proximity to failure — but is not error-free.

So the more useful question is not: which system is scientifically better?

It is: which information does each system give me — and when do I need it?

Percentage-based training: clear loads from a reference performance

In percentage-based training you derive your working weights from a 1RM or another reference figure.

For example:

Your bench press 1RM is 120 kg.

80 % of that is:

120 kg × 0.80 = 96 kg

The plan might turn that into 3×5 at around 95–97.5 kg.

The big advantage is obvious: the load is defined before you train.

You do not have to decide on every set whether 92.5 or 100 kg is “right” today. That makes percentage plans easy to schedule, reproducible and comparable across weeks.

Not every percentage plan calculates from a true 1RM

5/3/1 is a good example.

There, the percentages are not calculated directly from your actual 1RM but from a deliberately more conservative Training Max. That builds additional headroom into a percentage-based system from the start.

If the difference is not yet clear, it is covered in more detail here:

→ Training Max explained

This matters because “percentage training” is not a single method. 80 % of a true 1RM is something different from 80 % of an 85 or 90 percent Training Max.

Where percentages run into limits

A percentage figure does not know your performance on the day.

Sleep, previous sessions, stress, nutrition and ordinary biological variation can all affect how heavy the same load actually is today.

That does not mean your “daily 1RM” swings dramatically every morning. But a weight that allowed five clean reps last week can feel noticeably different today.

This is exactly where a rigid percentage lacks context.

When the plan says 5×5 at 80 %, the number does not automatically answer:

• whether set 1 is RPE 6 or RPE 8 today
• whether set 5 stays technically clean
• whether you have recovered from the previous session
• whether the reference 1RM you used is still realistic

So percentages are precisely calculated, but not automatically precisely dosed.

RPE: subjective, but systematic

RPE stands for rate of perceived exertion.

In strength training a repetition-based 1–10 scale is often used, closely tied to RIR — reps in reserve.

The word roughly is decisive.

RPE is not a measuring probe. You estimate how many reps you could still have completed under comparable technical conditions.

The advantage: it describes something a percentage does not capture directly — your actually perceived proximity to your limit.

How accurate is RPE, or RIR?

The answer is more nuanced than “beginners are always off by 1–2 points”.

Research on RIR estimation shows several relatively stable patterns:

• The judgement gets more accurate closer to failure.
• At heavier loads the estimate can be more accurate than at very light loads with many possible reps.
• Experience can help, but is not a clear independent predictor in every study.
• Even trained people are not always exactly right.

In one study across various compound lifts, 2-RIR estimates were more accurate at 85 % 1RM than at 65 or 75 %. Other work shows mean absolute errors of roughly one rep or less in trained people on the bench press when 1–3 RIR is estimated.

So the practical takeaway is not:

RPE is inaccurate.

It is:

RPE is a usable steering instrument, as long as you do not treat a subjective estimate like a laboratory value.

Is RPE training better than percentage training?

For maximal strength there are now several systematic reviews of autoregulated training.

A 2021 meta-analysis found an advantage for autoregulated methods over fixed load prescriptions. A more recent network meta-analysis from 2025 also rates RPE-based systems as competitive, while showing that results differ depending on the exercise and the comparison method.

That argues against two simple claims:

1. “RPE and percentages are demonstrably exactly equally good.” — The data is too heterogeneous for that.
2. “RPE is clearly superior.” — That cannot be asserted across the board either.

For hypertrophy the direct comparison is thinner still. Muscle growth depends heavily on whether enough productive volume, sensible loads and appropriate proximity to failure come together. Whether the weight was selected via percentages or RPE probably matters less than the training that actually gets done.

RPE and percentages answer different questions

That is the core of the comparison.

Percentages ask:

Which load is planned relative to my reference performance?

RPE asks:

How close was this specific set to my limit today?

So the systems do not necessarily compete with each other.

A plan can use both at once.

The hybrid approach: plan with percentages, interpret with RPE

Suppose your TM for the squat is 160 kg.

The plan calls for 4×5 at 75 % TM:

160 × 0.75 = 120 kg

Now you can use RPE as a second piece of information.

Scenario A: everything runs normally

• Set 1: 120 × 5 @ RPE 6.5
• Set 2: 120 × 5 @ RPE 7
• Set 3: 120 × 5 @ RPE 7.5
• Set 4: 120 × 5 @ RPE 8

The percentage prescription and the actual effort line up plausibly.

Scenario B: an unusually bad day

• Set 1: 120 × 5 @ RPE 8.5
• Set 2: 120 × 5 @ RPE 9.5

Now RPE provides context: the planned weight is considerably more demanding today than expected.

An editorial training rule might then be to reduce the load slightly for the remaining sets or drop a set. That is not a universal scientific threshold but one way to connect percentage planning with autoregulation.

Scenario C: an unusually good day

If 120 kg is suddenly RPE 5 instead of RPE 7–8, that does not automatically mean you should add 15 kg on the spot.

A good program can prescribe submaximal work deliberately.

RPE is feedback — not an order to turn every good day into a max attempt.

When percentages are particularly useful

Percentage prescriptions fit well when you:

• are planning a structured training cycle
• want to distribute loads reproducibly across several weeks
• work with a Training Max or a realistically tested 1RM
• do not want to make many decisions in every workout
• want to train specific load ranges for a sport

That is one reason systems like Jim Wendler’s 5/3/1 can work well without daily RPE steering: the conservative reference figure is already part of how load is managed.

When RPE is particularly useful

RPE or RIR help especially when:

• your daily readiness fluctuates noticeably
• a plan deliberately prescribes target RPEs
• you want to autoregulate loads within a rep range
• your 1RM is not current or not sensibly testable
• you want to judge whether the same load feels easier over the weeks

An example:

3×6 @ RPE 8

Here the weight is not prescribed. You choose a load at which six reps leave roughly two reps in reserve.

That is more flexible — but it also demands a reasonable self-assessment.

Can beginners use RPE?

Yes.

They just should not treat RPE with false precision.

A beginner does not have to decide after every set whether that was RPE 7.25 or 7.5. A coarser breakdown can be more useful:

• clearly easy
• moderate
• hard but clean
• very close to the limit

With experience, that perception can then increasingly be linked to RPE or RIR.

Interestingly, newer studies do not consistently show that more years of training alone automatically produce better RIR estimates. What also seems decisive is how close to failure you are, which load you are working with, and how familiar the specific exercise is.

The most common RPE mistake: confusing autoregulation with arbitrariness

RPE does not mean:

I don’t feel perfect today, so I’ll do less.

And it does not mean:

I feel strong, so I’ll just turn RPE 8 into RPE 10.

A good autoregulated program has rules.

For example:

• target: 3×5 @ RPE 7–8
• if set 1 is already clearly above RPE 8 → reduce the load
• if all sets are clearly below RPE 7 → increase moderately next session

The exact boundaries are program design. The important point is: RPE does not replace structure. It adds to it.

Combining percentages and RPE in practice

The plans in hitPR can run on percentages via a Training Max. On top of that you can log RPE or RIR for every set.

That does not mean a plan’s progression has to become RPE-driven. A fixed progression algorithm can still run on reps, AMRAP results or scheme changes.

RPE then provides a second layer:

Did the same objective performance get subjectively easier?

If you compare 100 kg × 5 across several months, for instance, 100 × 5 @ RPE 8 is different information from 100 × 5 @ RPE 10.

That makes RPE particularly valuable as context in your training log.

Conclusion

RPE and percentage training do not measure the same thing — so neither has to replace the other.

Percentages give you structure from a reference performance. RPE gives you context on how demanding a set actually was.

For maximal strength, autoregulated methods do very well across several reviews, but the evidence justifies neither “RPE is always better” nor “both are exactly equivalent”. In practice the combination is often particularly robust: an objective load prescription provides the frame, and RPE or RIR show you how that frame is actually working today.

Further reading:

• RIR vs. RPE: what’s the difference?
• Training Max explained
• Jim Wendler’s 5/3/1 explained
• nSuns 5/3/1 LP: AMRAP autoregulation in practice

Sources

• Zourdos MC et al. (2016). Novel Resistance Training-Specific Rating of Perceived Exertion Scale Measuring Repetitions in Reserve. Journal of Strength and Conditioning Research, 30(1):267–275.
• Helms ER et al. (2016). Application of the Repetitions in Reserve-Based Rating of Perceived Exertion Scale for Resistance Training. Strength and Conditioning Journal, 38(4):42–49.
• Helms ER et al. (2018). RPE and Velocity Relationships for the Back Squat, Bench Press, and Deadlift in Powerlifters. Journal of Strength and Conditioning Research, 32(2):292–297.
• Hughes LJ et al. (2020). Estimating Repetitions in Reserve in Four Commonly Used Resistance Exercises. Journal of Strength and Conditioning Research.
• Mansfield SK et al. (2020). Estimating Repetitions in Reserve for Resistance Exercise: An Analysis of Factors Which Impact on Prediction Accuracy. Journal of Strength and Conditioning Research.
• Halperin I et al. (2022). Accuracy in Predicting Repetitions to Task Failure in Resistance Exercise: A Scoping Review and Exploratory Meta-analysis. Sports Medicine, 52:377–390.
• Emanuel A et al. (2023). Accuracy of Intraset Repetitions-in-Reserve Predictions During the Bench Press Exercise in Resistance-Trained Male and Female Subjects. Journal of Strength and Conditioning Research.
• Zhang X et al. (2021). Auto-Regulation Method vs. Fixed-Loading Method in Maximum Strength Training for Athletes: A Systematic Review and Meta-Analysis. Frontiers in Physiology, 12:651112.
• Li Y et al. (2025). Autoregulated resistance training for maximal strength enhancement: A systematic review and network meta-analysis. Journal of Exercise Science & Fitness, 23(4):360–369.
• Wendler J. (2009). 5/3/1: The Simplest and Most Effective Training System to Increase Raw Strength.