Your Reps in Reserve Estimate Is Probably Off. Here's How to Calibrate It

TL;DR: Reps in reserve (RIR) is the best tool we have for autoregulating effort, but it only works if the number in your head matches reality. The research is consistent: most lifters, and beginners especially, stop further from failure than they believe. Your "2 reps left" is often 4 or 5. Calibrate it by occasionally taking a set to true failure and counting, by watching bar speed instead of trusting how the set felt, and by reading your own logged reps over time. A well-calibrated RIR turns effort from a guess into a dial you control.

---

If you already know you should stop most sets within a couple of reps of failure, you've solved half the problem. The other half is harder and almost nobody talks about it: can you actually tell when you're there? Because all the careful programming in the world falls apart if your "2 RIR" is really a 5. You think you're training hard. The set says otherwise.

The Estimate Is the Weak Link

RIR is the number of reps you could still have completed when you ended a set. It's the practical version of autoregulation, and it's a genuinely good idea. The catch is that it depends entirely on a subjective judgment made in the worst possible moment, mid-fatigue, with adrenaline up and your ego quietly negotiating.

The accuracy data is humbling. When Hackett et al. (2012) had trained lifters predict how many reps they had left at submaximal loads, they systematically underestimated, stopping with more in the tank than they realized, and they only got accurate once they were genuinely close to failure. A 2024 scoping review by Bastos and colleagues reached the same broad conclusion across the literature: RIR scales are useful and feasible, but estimation error is real, larger at lighter loads and higher rep counts, and larger still in less experienced lifters, who can be off by several reps. The skill takes weeks of deliberate practice to develop, not a single session of good intentions.

That matters because the error almost always runs one direction. People stop too early and call it hard. The result is a training log full of sets labeled "RIR 2" that were really RIR 4 or 5, which is exactly the zone where, as the dose-response research on how close to failure you should train shows, you start leaving measurable growth on the table.

Why You Overshoot

Three things conspire against an honest estimate.

Discomfort gets misread as proximity. The burn, the labored breathing, the urge to rack the bar: those are real sensations, but they show up well before mechanical failure, especially on higher-rep sets. Your nervous system is asking you to stop. That request is not the same as being one rep from physically failing.

You've never actually measured your own failure point. If you've rarely or never taken a set to the true wall, you have no reference. You're estimating distance from a landmark you've never visited. Calibration requires occasionally visiting it.

Load changes the signal. Estimating RIR is easier with heavier loads and lower reps, where the final reps slow down dramatically and failure announces itself. On a set of 15 to 20, the ramp to failure is gentler and the estimate gets noisier. This is part of why people who live in high-rep ranges tend to under-train without realizing it. If you're unsure what weight even lands you in a given rep range, working backward from a one-rep max estimate gives you a sane starting load to calibrate against.

How to Calibrate It

You don't fix this by trying harder to feel the number. You fix it with reference points.

1. Take periodic sets to true failure, on purpose. Pick a safe exercise, a machine or a dumbbell movement you can fail without danger, and once every week or two, take one set all the way to the point where the rep stops moving despite full effort. Count the reps. Then compare that to what you'd have guessed your last rep was. Most people discover their honest failure is two or three reps past where they normally stop. That gap is your calibration error, and naming it is most of the fix.

2. Judge by bar speed, not by feeling. The most reliable external signal of proximity to failure is involuntary slowing. As you approach the last few reps, concentric speed drops sharply even when you're pushing maximally. When a rep is noticeably slower than the one before it despite full intent, you're within roughly 1 to 2 reps of failure. Velocity is the cue the RIR-based RPE scale was built around when Zourdos et al. (2016) developed it: the scale exists precisely because how a set feels is less trustworthy than how the bar moves.

3. Let your log audit you. This is where consistent tracking earns its keep. If you prescribe a weight for a target of 8 reps at 2 RIR and you hit 8 but the previous week you got 11 with the same load, your "2 RIR" was a fiction. Reps achieved at a fixed load, tracked over time, are an objective check on a subjective estimate. Numbers don't have an ego.

4. Recalibrate during a deload. Fatigue distorts perception. The same set feels closer to failure when you're beaten up, which is one more reason a planned recovery week matters. After a deload, retest your failure point fresh, because the calibration you set while exhausted will be wrong once you're recovered.

What Good Calibration Buys You

When your RIR is honest, every other training variable starts working as intended. Your prescribed proximity to failure actually lands in the productive rep range instead of drifting too easy. Your week-to-week progressive overload reflects real adaptation rather than you simply deciding to try harder. And you spend your recovery budget on sets that earned it, rather than grinding to failure on everything to be "safe."

It's worth remembering why the modest stuff matters here. Refalo et al. (2024) found that training to true failure produced essentially the same hypertrophy as stopping 1 to 2 reps short over eight weeks, at a higher fatigue cost. So you don't need to live at failure. You need to know where it is so you can reliably stop just shy of it. That's the whole game: not maximum effort, but accurate effort.

Make the Number Trustworthy

RIR is only as good as the honesty of the estimate behind it. Treat it like any other measurement, something to calibrate against a reference and audit with data, not a vibe you assign and forget. Take the occasional set to failure to find your wall. Watch the bar slow down. And let your logged reps tell you when your perception and reality have drifted apart.

That last part is hard to do in your head across dozens of sets a week, which is exactly what Protokl is built for. It records your reps and prescribed RIR set by set, flags when your performance says a target was too easy or too hard, and adjusts the next session's prescription accordingly, so your effort stays calibrated without you having to play statistician between sets. Estimate honestly, and let the data keep you honest. Try Protokl and turn RIR from a guess into a dial.