Is Training to Failure Actually Necessary for Strength?

This week in strength science — New evidence-based training research from 2026 gives a clear answer: no, training to momentary muscular failure is not required for comparable strength and hypertrophy outcomes. Stopping 1–3 repetitions short of failure produces equivalent muscle thickness and 1RM strength gains in trained individuals, and the skill of estimating those remaining repetitions accurately can be developed in as few as six sessions — regardless of age.

This Week's Common Theme: The Case for Effort-Based Autoregulation

Two studies published in 2026 independently examined the same underlying question from different angles: how useful is the repetitions-in-reserve (RIR) framework as a practical training tool?

One study tested whether RIR-based training produces outcomes comparable to failure training. The other tested whether lifters can actually learn to estimate RIR accurately in the first place. Together, they form a coherent argument that RIR-based autoregulation is both effective and learnable — which has real implications for how you structure a program.

Study 1: RIR vs. Failure — Does the Difference Matter for Adaptations?

Key Finding: In trained men and women, stopping 1–3 reps short of failure produced muscle thickness and strength gains statistically indistinguishable from training to momentary muscular failure over eight weeks.

What Was Studied

Vasconcelos and colleagues recruited 19 trained participants (11 men, 8 women; mean age 22.3 years) for an eight-week unilateral leg extension intervention. Each participant trained both legs — one to momentary muscular failure (FAIL), the other stopping at 1–3 RIR — allowing a direct within-participant comparison that controls for individual differences in recovery, nutrition, and genetics.

Results

The estimated between-protocol difference for hypertrophy was negligible: −0.019 cm (95% credible interval: −0.094 to 0.054 cm) for knee extensor muscle thickness. For 1RM strength, the difference was 0.198 kg (−1.87 to 2.6 kg). Secondary analyses using standardized mean differences confirmed no clear advantage for either protocol.

Limitations

  • The study used a single-joint isolation exercise (leg extension), so results may not transfer directly to multi-joint compound movements where fatigue and technique interact differently.
  • Participants were young trained adults; the findings may not generalize to beginners or older populations.
  • Eight weeks is a relatively short intervention window for detecting meaningful divergence in hypertrophy outcomes.

What This Means for Your Training

If you've been pushing every set to absolute failure in the belief that it's necessary for progress, this study suggests you can back off without sacrificing results. Training at 1–3 RIR likely accumulates less systemic fatigue, which may support better session-to-session recovery and more consistent progression over time.

Citation: Vasconcelos T, Ruivo A, Refalo MC, et al. Comparative Adaptations to Different Leg Extension Set-termination Strategies in Trained Men and Women: A Unilateral Within-Participant Study. Journal of Strength and Conditioning Research. 2026. DOI: 10.1519/JSC.0000000000005494

Study 2: Can Lifters Actually Learn to Estimate RIR Accurately?

Key Finding: Both younger and older adults significantly improved their RIR estimation accuracy over six training sessions, with mean absolute error dropping by up to 2.3 repetitions from session one to session six — and no meaningful age-related differences in either baseline accuracy or rate of improvement.

What Was Studied

Wiedenmann and colleagues ran a six-week parallel-group repeated-measures study with 13 younger adults (mean age 33 years) and 13 older adults (mean age 67 years). Participants completed three sets each of bench press and leg press at randomized loads between 75–85% of their one-rep maximum. During each set, they verbally indicated the rep at which they estimated having 4 RIR and 2 RIR, then continued to concentric failure. This allowed researchers to calculate the gap between perceived and actual remaining reps.

Results

Linear mixed-effects models showed a significant main effect of training session on mean absolute error for both exercises at both 2RIR and 4RIR. Error decreased by up to 2.3 repetitions across six sessions. Critically, there was no significant effect of age group and no significant session-by-group interaction — meaning older adults improved at the same rate as younger adults from a comparable baseline.

Limitations

  • The sample size was small (13 per group), which limits statistical power for detecting subtle age-related differences.
  • The study used bench press and leg press only; RIR calibration for other exercises may follow a different learning curve.
  • Participants trained to failure each session to enable accuracy measurement, which may not reflect typical RIR-based programming where failure is intentionally avoided.

What This Means for Your Training

The common criticism of RIR-based training is that it requires accurate self-perception — and that most people are bad at it. This study shows the skill is trainable, and that it improves meaningfully within a short familiarization period. If you're new to tracking your effort by RIR, expect the first few sessions to feel imprecise. That's normal. Accuracy improves with deliberate practice.

Citation: Wiedenmann T, Kunellis D, Rappelt L. Trainability of repetitions-in-reserve estimation in younger and older adults: a parallel-group repeated-measures study. BMC Sports Science, Medicine & Rehabilitation. 2026. DOI: 10.1186/s13102-026-01997-y

What These Studies Mean Together

Taken individually, each study answers a useful question. Taken together, they build a practical framework.

Study one establishes that RIR-based training is effective — you don't leave meaningful adaptation on the table by stopping short of failure. Study two establishes that RIR estimation is learnable — the main practical barrier to using this approach dissolves with a short familiarization period.

This matters for how you think about program design. If stopping 1–3 reps short of failure produces equivalent results with likely less fatigue accumulation, and if the skill of estimating those remaining reps can be developed within six sessions, then RIR-based autoregulation becomes a genuinely practical tool — not just a theoretical one.

The Fatigue Management Argument

One of the underappreciated costs of training to failure on every set is cumulative fatigue. When fatigue accumulates faster than it can be cleared between sessions, training quality degrades — not in a single session, but over weeks. RIR-based programming gives you a mechanism to modulate intensity intentionally, preserving the quality of your sessions across a full training block.

This is particularly relevant for lifters running higher-frequency programs or those managing a demanding schedule outside the gym. Consistency across weeks matters more than intensity on any given day.

Structured Progression and Intelligent Load Management

For RIR-based autoregulation to work in practice, you need a reliable record of what you've done. Knowing that last week's working set felt like 2 RIR at a given load tells you something actionable: you can likely add load or volume this week. Without that record, RIR becomes a loose subjective impression rather than a structured progression tool.

This is where tracking your training with a consistent logging system becomes more than a habit — it becomes the infrastructure that makes autoregulation functional. Kenso's rule-based progression engine is built around exactly this logic: each session's effort data informs the next session's targets, so progression follows from what you've actually done rather than from a fixed schedule that ignores how you responded.

A Note on Older Adults

The Wiedenmann study's finding on older adults deserves specific attention. The assumption that older adults are less capable of accurately perceiving effort — and therefore less suited to autoregulation-based programming — is not supported by this data. Older adults improved their RIR accuracy at the same rate as younger adults. This extends the practical utility of RIR-based training to a population that often benefits most from individualized load management.

Practical Takeaways

Here's what these two studies, read together, suggest for your programming:

  1. Stop defaulting to failure. Training at 1–3 RIR produces comparable strength and hypertrophy outcomes in trained individuals, with likely lower fatigue cost.
  2. Treat RIR calibration as a skill to develop. Your early estimates will be imprecise. Log them anyway. Accuracy improves meaningfully within six sessions of deliberate practice.
  3. Use session records to anchor your effort ratings. A perceived 2 RIR means more when you can compare it to the same exercise two weeks ago at the same load.
  4. Don't assume age limits autoregulation. The evidence does not support that assumption.
  5. Manage fatigue deliberately across a training block, not just within a single session.

If you're already logging sessions in Kenso, you have the data infrastructure to make RIR-based autoregulation work. If you're not yet tracking this level of detail, Kenso's AI Coach can help you interpret session-to-session effort trends and flag when your progression targets need adjusting based on your actual training history.

Conclusion

The evidence from these two 2026 studies converges on a straightforward conclusion: RIR-based training is a legitimate, learnable, and effective approach for managing training intensity. It doesn't require you to sacrifice results, and it doesn't require years of experience to implement — just a short familiarization period and a reliable way to log what you're doing.

The most durable training approaches are the ones you can sustain across months and years without accumulating unmanageable fatigue or relying on motivation to push through every set. Effort-based autoregulation, grounded in accurate self-perception and consistent tracking, is one of the more evidence-supported ways to build that kind of sustainable structure.

Train with intention. Track what matters.

Ready to put RIR-based progression into practice? Kenso is built for lifters who want structured, data-informed programming — not guesswork. Download the Kenso app on iOS and start logging your effort ratings alongside your loads and reps. Your progression engine will take it from there.

Frequently Asked Questions

Is training to failure necessary for muscle growth?

Based on current evidence, no. A 2026 unilateral study found that stopping 1–3 repetitions short of failure produced statistically equivalent muscle thickness and 1RM strength gains compared to training to momentary muscular failure over eight weeks in trained individuals.

What does RIR mean in strength training?

RIR stands for repetitions in reserve — the number of additional reps you estimate you could complete before reaching muscular failure. A set performed at 2 RIR means you stopped with approximately two reps remaining in the tank.

How accurate is RIR estimation, and can it be improved?

RIR estimation accuracy is imperfect at first but improves significantly with practice. A 2026 study found that mean absolute error decreased by up to 2.3 repetitions across just six training sessions in both younger and older adults.

Does age affect your ability to learn RIR estimation?

According to a 2026 parallel-group study, older adults (mean age 67 years) showed no significant difference from younger adults (mean age 33 years) in baseline RIR accuracy or in their rate of improvement across six sessions.

Why does tracking your training matter for RIR-based programming?

RIR ratings are most useful when they can be compared across sessions. Knowing that a given load felt like 2 RIR last week gives you a concrete reference point for this week's targets. Without session records, effort ratings remain subjective impressions rather than actionable data for structured progression.

Full Citations

  1. Wiedenmann T, Kunellis D, Rappelt L. Trainability of repetitions-in-reserve estimation in younger and older adults: a parallel-group repeated-measures study. BMC Sports Science, Medicine & Rehabilitation. 2026. DOI: 10.1186/s13102-026-01997-y. PubMed: https://pubmed.ncbi.nlm.nih.gov/42632893/

  2. Vasconcelos T, Ruivo A, Refalo MC, Chaves Lucas G, Vasconcelos PR, Brito JP, Vilaça-Alves J, Oliveira R. Comparative Adaptations to Different Leg Extension Set-termination Strategies in Trained Men and Women: A Unilateral Within-Participant Study. Journal of Strength and Conditioning Research. 2026. DOI: 10.1519/JSC.0000000000005494. PubMed: https://pubmed.ncbi.nlm.nih.gov/42617172/