Training to Failure: How Close Is Close Enough

The gym has a folklore about the last rep. Either you took every set to the floor and you are serious, or you left reps in the tank and you are leaving gains on the table. Both stories are simpler than the work.

A set has a proximity to failure: how many clean reps were still available when you racked it. That number, not the theatre of the last grind, is what the literature is actually talking about when it talks about intensity of effort. Volume vs intensity already separates load from effort. This article is about the effort end — how close is close enough, and when the last rep costs more than it buys.

Failure Is a Definition, Not a Mood

Concentric failure is the point at which you cannot complete another rep with the same technique. The bar does not go up. That is 0 reps in reserve, RPE 10.

Technical failure is earlier for a serious lifter: the point at which the next rep would no longer be the lift you are training. The squat turns into a good morning. The bench turns into a bounce and a wriggle. The deadlift turns into a hitch you will spend a month unlearning. If you only count concentric failure, you will count some of those reps as work. They are practice of a different pattern.

Reps in reserve (RIR) is the count of clean reps you could still have done. 3 RIR is three clean reps left. 1 RIR is one. 0 RIR is concentric failure on a clean rep.

RPE, in the resistance-training sense Zourdos and colleagues validated, is the same idea on a 1–10 scale tied to those leftover reps. RPE 7 is about 3 RIR. RPE 8 is about 2. RPE 9 is about 1. Helms, Cronin, Storey, and Zourdos then wrote the practical paper: how to use that scale so a session's effort matches the session's job.

If you cannot tell 2 RIR from 4 RIR, the scale is not broken. Your calibration is. The fix is occasional, planned sets that actually reach 0–1 RIR on a lift you can fail safely — a machine, a spotter, or a last back-off set — so the easier sets have a real reference. Guessing "that was an 8" every week without ever visiting 10 is how the scale drifts high.

What the Last Rep Buys

Schoenfeld, Grgic, Ogborn, and Krieger's meta-analysis on low-load versus high-load training is the load paper most people already know: you can build muscle with lighter weights if the sets are taken close enough to failure, and you build more strength when the load is heavier. Lopez and colleagues' later network meta-analysis pointed the same direction. Load is not the only slider. Proximity to failure is how a light set becomes a hard set.

That is the useful reading for this article. A set of 12 at a weight you could have taken to 20 is not a hypertrophy set. It is a warm-up that lasted too long. A set of 5 at a weight you could have taken to 6 is a strength set that did its job. The last rep is not magic. The neighbourhood of failure is.

Schoenfeld, Ogborn, and Krieger's volume meta-analysis still matters here. Muscle growth tracks weekly hard sets in a dose-response window. "Hard" is the proximity question. Adding a failure rep to a set you already took to 1 RIR does not add a new set. It adds fatigue to a set that was already in the window. Ralston's meta-analysis on weekly set volume and strength says the same thing in strength language: more hard sets, up to a point. The last grind is not a substitute for another quality set you can recover from.

Plotkin and colleagues showed you can overload by adding reps at the same load, not only by adding load. That is proximity in another costume. If you stop at 8 every week with 4 left, you never applied the new dose. If you stop at 8 with 1 left, and next month that 8 is cleaner, you did. Failure is one way to make sure the dose arrived. It is not the only way, and it is a poor way to do it on every working set of a squat.

How Close Is Close Enough

For most working sets on the competition lifts, 1 to 3 RIR is the neighbourhood.

3 RIR is enough to count as work when the load is already heavy and the pattern is still crisp. It is the right place to live on a volume day, on a second squat session, on a set you need to repeat twice more.

2 RIR is the default hard set. You felt the bar slow. You could have done two more clean reps and you did not. That is a hypertrophy set and a strength set.

1 RIR is the top of the useful window for compounds. One clean rep left. The next one would have been ugly or would have been failure. You will not do many of these in a week if the lift is a squat or a deadlift.

0 RIR is a tool. Last set of a machine. Last set of an isolation. A planned top set with a spotter when you are testing the scale. It is not a personality, and it is not a weekly quota.

The reason the window is "close" rather than "over the edge" is recoverability. A failure set on a hard compound costs you the rest of the session and, often, the next one. The same proximity on a leg extension costs you a local burn. Spend the expensive last rep where the exercise is cheap.

Helms' application paper is explicit that RPE is a programming language, not a dare. You prescribe 8s because you want 2 RIR, not because 8 is a nicer number than 9. If every prescribed 8 is actually a 9.5, you are not following the programme. You are writing a harder one in the margins.

When Failure Is the Right Buy

Low-load work needs to finish closer. If the weight is light enough that 3 RIR still feels like a conversation, you have not recruited the set. Take those sets to 0–1 RIR, preferably on a machine or a single-joint lift, and stop calling them "easy volume."

Isolation work can afford 0 RIR more often. A curl that ends at failure does not scramble the next day's squat. A set of laterals that ends at failure is a local problem. Use failure here when you need to be sure the set was hard and you do not have a good load to make 2 RIR meaningful.

Calibration sets belong on the calendar. Once every week or two, on a lift you can fail without inventing a new pattern, take a back-off to 0–1 RIR so the rest of the week's 7s and 8s stay honest. If you never visit the end of the set, "2 RIR" becomes "I was tired of the set."

Testing is not training. A 1RM or an AMRAP is a measurement. You may go to 0 RIR. You do not then do four more measurements that day and call it volume.

When the Last Rep Is Too Expensive

Heavy compounds at high percentages. A squat at 90 percent already lives near failure. Taking it to 0 RIR is how you get a miss, a form breakdown, and a week of a worse squat. Leave 1–2 clean reps. The strength work already happened.

Deadlifts, for most lifters, even more so. The last ugly pull is a skill you will rehearse. If the programme wanted a grind, it would have said so and it would have given you fewer sets.

High-frequency weeks. If you squat three times, none of those sessions can be a failure contest. The third session is there because the first two left you able to show up.

When the bar speed dies and the shape dies at the same time. A slow clean rep at 1 RIR is still the lift. A slow ugly rep at 0 is a different lift. Rack it.

When you cannot say what RIR you are at. If the set is a fog, you are not training proximity. You are surviving the set. Cut the reps or the load until you can name the leftover reps. Then the next week can be a real progression, which is the actual iron law.

A Week That Uses the Neighbourhood

A useful template, not a religion:

Main strength sets on squat, bench, and deadlift: 1–3 RIR. Most of them at 2. The top set of the week may be 1. None of them are 0 unless the session is a test.

Back-off volume on those lifts: 2–3 RIR. You should be able to do every prescribed set. If set four is a 10, the load was a 9 on set one and you lied to the log.

Accessories: 0–2 RIR. Machines and single-joint work can take the last rep. Rows and presses that still look like the competition lift should stay in the 1–2 window.

One calibration set per week, on something you can fail cleanly.

Write the RIR down. Not later. On the set. The log is how you find out that your "8s" have been 9.5s for a month, which is how fatigue shows up before the weights go backwards. Volume vs intensity already told you that creeping RPE at the same load is a fatigue signal. This article is the other half: if the RPE never creeps because you never get close enough, the signal is missing for a different reason.

A useful check at the end of the week: count how many compound sets finished inside 1–3 RIR, and how many finished at 0 or at 4+. The first number should be most of the week. The second number should be a short list you can explain — a test, a calibration, or a set you should have racked sooner. If you cannot explain it, it was not programmed. It was a mood.

Key Takeaways

• Failure is 0 RIR: no more clean reps. Technical failure is earlier, and it is the one that matters on the competition lifts.
• RPE and RIR are the same proximity scale. They only work if you occasionally visit the end of a set and calibrate.
• Muscle can be built with lighter loads when sets finish close to failure; strength still wants heavier loads. Proximity is how a light set becomes a hard set.
• Weekly hard sets drive the adaptation. The last grind on a set already at 1 RIR is mostly extra fatigue.
• Most compound working sets belong at 1–3 RIR. 0 RIR is a tool for machines, isolation, calibration, and tests.
• If you cannot name the leftover reps, you are not managing proximity. Cut the set until you can.

Put It Into Practice

This week: On every working set of squat or bench, write a number: leftover clean reps. Do not change the programme yet. Just see whether your "hard" sets are 1 RIR or 4.

This training block: Put compounds at 1–3 RIR and move true failure to one accessory or machine set per session. Add one calibration set per week. If the compounds get cleaner and the accessories still finish near the end, the neighbourhood is right.

Long term: Progress the load or the reps at the same RIR. That is overload. Chasing 0 RIR every session is how the RIR number stops meaning anything, because every week is a test.

Analysed by FINAL FORM AI. Your form, perfected.

References

1. 1Zourdos MC, Klemp A, Dolan C, et al.. 'Novel resistance training-specific rating of perceived exertion scale measuring repetitions in reserve.' J Strength Cond Res. 2016. PMID: 26049792
2. 2Helms ER, Cronin J, Storey A, et al.. 'Application of the repetitions in reserve-based rating of perceived exertion scale for resistance training.' Strength Cond J. 2016. PMID: 27531969
3. 3Schoenfeld BJ, Grgic J, Ogborn D, et al.. 'Strength and hypertrophy adaptations between low- vs. high-load resistance training: a systematic review and meta-analysis.' J Strength Cond Res. 2017. PMID: 28834797
4. 4Lopez P, Radaelli R, Taaffe DR, et al.. 'Resistance training load effects on muscle hypertrophy and strength gain: systematic review and network meta-analysis.' Med Sci Sports Exerc. 2021. PMID: 33433148
5. 5Schoenfeld BJ, Ogborn D, Krieger JW. 'Dose-response relationship between weekly resistance training volume and increases in muscle mass: a systematic review and meta-analysis.' J Sports Sci. 2017. PMID: 27433992
6. 6Ralston GW, Kilgore L, Wyatt FB, et al.. 'The effect of weekly set volume on strength gain: a meta-analysis.' Sports Med. 2017. PMID: 28755103
7. 7Plotkin D, Coleman M, Van Every D, et al.. 'Progressive overload without progressing load? The effects of load or repetition progression on muscular adaptations.' PeerJ. 2022. PMID: 36199287