RPE, RIR and failure: how hard to lift when you also run

RIR and RPE are the same scale in lifting. RIR is the reps you had left at the end of a set, and the lifting RPE scale is built on it: RPE 10 is no reps left, RPE 9 is one, RPE 8 is two.¹ On how hard to push, the strength research is clear enough: training to failure doesn't build more strength than stopping a couple of reps short.² For someone who also runs, stopping short has a second benefit. Failure sets slow recovery for 24 to 48 hours, and that's exactly when your next run happens.³

RIR and RPE are one scale#

The classic RPE scale asks how hard something felt. For lifting, Zourdos and colleagues tied it to something you can count. You finish the set and ask how many more clean reps you had. That number sets the rating.¹

The top rows are from Zourdos; the bottom three, where the scale stops counting reps, are from Helms' guide to using it.¹⁴

How good are people at guessing? Better with practice, and better close to failure. In the original study, 15 experienced squatters rated a true 1RM closer to 10 than 14 novices did. In both groups, the slower the bar moved, the higher the rating, with correlations of -0.88 and -0.77.¹ Helms' guide adds that ratings get less accurate the further you are from failure, and that beginners should log RIR but not rely on it until they've checked themselves. The check is simple: rate a set, then take a later set at the same weight to failure and see how far off you were.⁴

If you run, you already have the habit. I put it like this to lifters: RPE is the one word you don't need translated. Runners use the same scale for the same reason, when the watch and the legs don't agree.

Picking a weight from RIR#

The scale also lets you swap reps for reserve. Helms' guide gives a rough rule: 6 reps at 0 RIR, 5 at 1 RIR, 4 at 2 RIR and 3 at 3 RIR all use about the same load.⁴ So a set of 4 at RPE 8 is roughly your 6-rep max weight, stopped two reps early.

That's handy when your legs are tired from a run. You don't need to change the plan, just the weight. If the bar feels heavy on the first set, drop it until the target reps land at the target RIR. The guide treats its own conversion table as a starting point, not an exact tool, and lets you adjust the load set to set.⁴

What failure buys you#

For strength, very little. Grgic's meta-analysis pooled 15 studies of failure against non-failure training. Strength gains weren't significantly different (effect size -0.09). Neither was muscle growth (0.22). When the non-failure groups were allowed to do more total volume, they gained more strength than the failure groups (-0.32).² Every study was in young adults. In the trained subgroup, failure did come out slightly ahead for muscle size (0.15), which is where the case for it is strongest.²

The newer dose-response work agrees. Robinson and colleagues estimated RIR for each study and ran it as a sliding scale instead of a yes-or-no. Strength gains were similar across a wide range of reps left in the tank. Muscle growth was the exception: it kept rising as sets ended closer to failure.⁵ The authors call it exploratory, because the RIR values were estimated from how studies described their training.

On size alone, the effect of failure is small. Refalo's meta-analysis found a trivial edge for going to set failure (0.19), with a confidence interval that starts at zero. Strictly defined momentary failure showed no significant edge (0.12).⁶

Failure vs stopping short: pooled effect sizes

	Low (effect size (positive favours failure))	High (effect size (positive favours failure))	Mark (effect size (positive favours failure))
Strength (Grgic)	-0.22	0.05	-0.09
Muscle size (Grgic)	-0.11	0.55	0.22
Size, set failure (Refalo)	0	0.37	0.19
Size, momentary failure (Refalo)	-0.13	0.37	0.12

Point estimates and 95% confidence intervals from Grgic 2022 (15 studies) and Refalo 2023 (15 studies). Only the Refalo set-failure interval reaches zero; the others cross it.

Read the top bar first. Strength sits right on zero. The size bars lean slightly toward failure, but none of them is far from zero. Refalo's own training study makes the point in one muscle group. Eighteen trained people trained one leg to failure and the other to 1 to 2 RIR for eight weeks. Quad thickness grew 0.181 cm on the failure leg and 0.182 cm on the other. The failure leg lost more bar speed and more reps along the way.⁷

The same question, measured with a bar#

Velocity loss is another way to measure closeness to failure. As a set goes on, each rep slows down, and you can stop when it's slowed by a set percentage. Pareja-Blanco compared stopping squat sets at 20% velocity loss with stopping at 40%, over eight weeks in 22 young men.⁸

Squat strength improved about the same in both groups, even though the 20% group did 40% fewer reps. The 20% group improved its countermovement jump more, 9.5% against 3.5%. The 40% group grew more muscle in two of the quad muscles, and its share of the fastest fibre type, IIX, fell. The 20% group's held.⁸ For a hybrid athlete, that's a reasonable trade to notice: closer to failure bought some size and cost some jump.

The hybrid cost lands on your next run#

Advice written for lifters stops here. Yours doesn't, because your next session might be a tempo run.

Morán-Navarro had 10 trained men do the bench and full squat three ways: 3 sets of 10 to failure, 6 sets of 5 with the same weight, and 3 sets of 5. The 6 x 5 protocol matched the failure protocol's total reps. Training to failure caused a bigger immediate drop in jump height and bar speed. Recovery from the two non-failure protocols was significantly faster between 24 and 48 hours after the session.³ Same work, half the reps per set, and you're fresher a day or two later.

Doma's review explains why that matters for running. Fatigue left over from a lifting session can impair an endurance session for hours to days. The proposed reasons are changes in how muscles are recruited, altered running mechanics that raise the energy cost of each stride, soreness and lower muscle glycogen. If that keeps happening, the review argues, your endurance training quality suffers over time.⁹ Refalo's scoping review adds the lifting side of the bill. Going to failure likely brings more neuromuscular fatigue, more muscle damage and more discomfort after the set than stopping short, which can hurt recovery, the next session and how long people stick with training at all. It also notes that "failure" has no agreed definition across studies, so the non-failure groups in them vary a lot.¹⁰ Run before or after lifting? covers the order and spacing side of this.

I explain failure to runners as running yourself to the bonk. It's the same wall, just under a bar. It's a tool for specific days, not a daily habit.

How close to failure to train when you also run#

This is our application of the studies above, not a tested protocol.

1. Leave 2 to 3 reps in reserve on most sets of your main lifts, which is RPE 7 to 8. Strength gains look the same across that range, and you recover faster.²⁵
2. Take the last set of a lift to 1 rep in reserve if you want to. That's where your ratings are most accurate anyway.⁴
3. Save true failure for the last set of small isolation work, where Helms suggests putting it, and keep it off squats and deadlifts the day before a run that matters.⁴
4. In the 24 to 48 hours before a key run, stay further from failure or move the lift.³

If size is your main goal, the case for getting closer to failure is stronger, and the trade-off with your running is real. If you're lifting mostly to get or stay strong, the evidence says you're giving almost nothing up by leaving reps in the bank. Logging the RIR of your top set each session also gives you a cheap fatigue check. If the same weight rates a point higher than last week, the day after a hard run, that's the carry-over described above showing up in the gym, and a reason to ease off rather than push. How many sets to do is a separate question, covered in how much lifting you need when you also run, and session RPE and training load covers rating the whole session rather than each set.