Calibrating True RIR: A Safe Gym Test for Technical vs Absolute Failure

Calibrating RIR means checking, once, how many reps you really have left when you think you have two. You do it with one safe test set on a machine: call your 2 RIR, keep going until a rep clearly breaks down, and count what you had in the tank.

Most lifters find the number is bigger than they guessed. This article gives you the exact drill, how to keep it safe, and how to read the result. For the full picture of load, rep ranges, RIR and rest, see our complete guide to load, rep ranges, RIR and rest intervals.

Why bother calibrating RIR at all?

Reps in reserve only works if your “2” is close to a real 2. If your felt 2 is actually 5, a whole programme built on “stop 2 reps short” is quietly much easier than you think. If you want a refresher on the scale itself, start with our RIR and RPE scale guide.

The research on this is mixed, so let’s be fair to both sides.

A 2022 scoping review with exploratory meta-analysis pooled 12 studies and 414 participants. On average, people underpredicted reps to failure by about 0.95 reps, but between-study differences were huge (I2 of 97.9%) (Halperin et al., 2022). Training status did not clearly change accuracy, and predictions were better when made closer to the end of the set. We dig into the full accuracy research in how accurately lifters gauge RIR; this article is about testing your own number.

On the optimistic side, 24 resistance-trained men and women predicting 1 and 3 RIR on barbell bench press were off by only 0.65 reps on average (absolute error), with a minor tendency to underpredict (Refalo et al., 2024).

On the other side, 14 and then 24 people with at least a year of training predicting failure on knee extensions typically underpredicted by about 2.0 reps (95% CI 0.0 to 4.0) (Armes et al., 2020).

Why might they disagree? Different exercises, different loads, and different ways of asking the question. Heavier, lower-rep sets and predictions made close to failure were more accurate in the pooled analysis. The practical takeaway: expect to be off by 1 rep or so, sometimes 2, and find out your own number.

An important caveat: these are mostly short, lab-based studies in young adults, often men. Your result in a crowded gym on a tired day can differ.

Technical failure vs absolute failure: what is the difference?

People use “failure” for at least two different moments.

• Technical failure: the next rep cannot be done with your normal form and tempo. You might still grind out an ugly rep, but you would not count it.
• Absolute (mechanical or momentary muscular) failure: you try your hardest and the weight stops moving. Research papers usually define it as momentary muscular failure.

Between those two sits the zone where reps get slower and sloppier but still move. If you want to standardise where you draw that line across a training block, see technical vs mechanical failure.

This matters for another reason too. A 2023 meta-analysis of 15 studies found no evidence that training to momentary muscular failure beats stopping short for muscle growth (Refalo et al., 2023). So calibration is not about making you train to failure. It is about knowing where failure is so you can stop confidently short of it. We cover that debate in training to failure vs RIR.

Which exercise should you use for the test?

Pick an exercise where a failed rep is harmless.

Good choices

• Leg press (set the safety stops, never lock your knees hard)
• Chest-supported row or chest-supported T-bar row
• Seated cable row, lat pulldown, pec deck, leg extension or leg curl

Poor choices for this test

• Barbell back squat, bench press, overhead press or deadlift, unless you have safety arms and a spotter you trust

The chest-supported row is great because the pad removes lower back fatigue as an excuse. If the load becomes too much you simply let it down. Our guide to failure on machines vs barbells goes deeper on why.

The step-by-step calibration protocol

1. Choose a machine exercise. Leg press or chest-supported row. Check the safety stops or pins first.
2. Warm up properly. Ramp up over 2-3 lighter sets. Do not do the test cold.
3. Pick a load for roughly 8-12 reps. Aim for a set that would normally be moderately hard. Very high-rep sets (above about 12 reps) were less accurate in the pooled analysis, so staying in this range keeps the test cleaner (Halperin et al., 2022).
4. Call your 2 RIR. As you lift, say “two” out loud (or tap a note) at the rep where you honestly believe you could do exactly 2 more. Do not stop. If you have a training partner, have them mark the rep number.
5. Keep going with the same tempo and form. Each rep must meet your normal standard: full range, no bouncing, no hips lifting, no jerking.
6. Stop at technical failure. The set ends when a rep slows dramatically or form breaks, or when you would not count the next one.
7. Count the gap. Reps completed after your call is your true reserve at that moment. If you called 2 and got 4 more, your “2” was really about a 4.

How to read your result

Reps you got after calling “2”	What it suggests	What to do next
0-1	You were closer to failure than you thought (you overpredicted)	Make your call a rep or so earlier, then retest
2	Your call was accurate on this lift	Keep going, retest in 8-12 weeks
3-4	You tend to underpredict (common in research)	To hit a true 2 RIR, go 1-2 reps past the point that feels like 2
5+	Your “hard” sets may be far from hard	Choose a heavier load and retest

Treat these as starting points, not laws. One test set is a single data point, and your result can shift with sleep, fatigue and food that day.

How to stop safely near failure

Safety is the point of doing this on a machine, but a few habits apply to everyone.

• Set the safeties first. On the leg press, set the stoppers so the sled cannot crush you if your legs give out.
• Do not lock out knees hard at the top of a leg press set, since fatigued joints and heavy sleds are a poor mix.
• Never test to technical failure on a free barbell lift alone. If you do try it, use safety arms or a spotter.
• Keep breathing. Do not hold your breath through a grinding last rep.
• Stop at pain, not at burn. The muscle should burn and feel heavy. Sharp joint or tendon pain, dizziness or chest pain means stop.

If you have a heart or blood pressure condition, or are returning from injury, ask your doctor before pushing sets near failure.

Does your RIR accuracy improve with practice?

Less than most people hope. In nine trained men who bench pressed three times a week for six weeks and went to failure on the last set, absolute prediction error did not significantly improve over time (Remmert et al., 2023). Participants tended to underestimate their reserve more in later sessions.

That is a small, short study, so do not treat it as the final word. A broader trained-and-untrained study of 27 men and 31 women on machine curls, pushdowns and seated rows found that sex, training experience and prior experience rating RIR did not significantly affect accuracy, but predictions were better closer to failure and in later sets (Remmert et al., 2023).

What this suggests: just training for years does not automatically calibrate you. No study has tested whether periodic self-tests sharpen your feel either, but they do tell you your current gap, which is the number you actually need.

How often should you recalibrate?

A reasonable habit is every 8-12 weeks, or any time you:

• start a new exercise or machine
• change your rep range substantially (for example 6-8 reps to 12-15)
• return after a long break
• feel that your sets have become too easy

To be clear, the 8-12 week interval comes from coaching practice, not a trial. The research simply shows that accuracy is imperfect and varies, so occasional rechecks make sense. A scoping review of 31 studies concluded RIR scales appear feasible and useful for setting intensity, and that predictions made closer to failure are more accurate (Bastos et al., 2024).

Using what you learn

Once you know your gap, apply it simply.

• If you underpredict by about 2, a “2 RIR” call really means about 4 left. To train at a true 2, go 2 reps further.
• Judge effort late in the set. Predictions closer to failure were more accurate (Halperin et al., 2022).
• Keep a short note in your log: exercise, load, rep of call, extra reps. Over a few tests you will see a pattern per exercise.
• Use our RIR and RPE calculator to convert between RIR and an estimated percentage of your max, and think about where your compound lifts and isolations should sit in proximity to failure rules for compounds vs isolations.

For many lifters, a simple rule works: compound lifts a few reps short, isolations and machines closer to failure. How close is a decision for your program and recovery, not for this one test.

If you want exact numbers to go with this, our free push, pull and leg plans give you structured sessions where you can run this test in week one.

Common mistakes with the RIR test

• Calling 2 RIR too early, then stopping. The test only works if you keep going.
• Letting form drift to squeeze out more reps. Cheat reps inflate your number. Stop at technical failure.
• Testing once and treating it as permanent. It is a snapshot.
• Using an exercise you never train. Calibrate on your real lifts.
• Testing when exhausted or underfed. Compare on typical days.

The bottom line

Calibrating RIR is simple: one safe machine set, a 2 RIR call, and a count of the reps you really had left. Studies show lifters usually miss by about 1 rep on average, sometimes 2, and that experience alone does not close the gap. Test on machines, stop at technical failure, repeat every 8-12 weeks or when something changes, and treat the result as a rough guide, not a rule.