How Accurate Are 1RM Calculators? What the Research Actually Says

Guides → 1RM Basics

On this page

TL;DR: A good 1RM estimate lands within about 5% of your true max — close enough to program from. But accuracy depends on three things: reps (5 or fewer is the sweet spot), the lift (bench predicts best, deadlift worst — every formula underestimates it by 9–14%), and the formula (Mayhew and Wathen came out on top in the landmark 1997 study). Past 10 reps, estimates get sloppy fast.

How Accurate Are 1RM Calculators?

Within roughly 5% of your true max when built the right way — from a hard set of 5 or fewer reps on a big compound lift, using a validated formula. That’s about 10–15 lb on a 300 lb bench: close enough that no program can tell the difference. Built the wrong way (15 reps, isolation lift, random formula), errors of 15%+ are common.

What Does the Research Say?

The landmark study is LeSuer et al. (1997), published in the Journal of Strength and Conditioning Research. They tested seven prediction equations on 67 untrained college students across bench press, squat, and deadlift. Everything since has refined their conclusions, not overturned them.

StudyWho was testedKey finding
LeSuer et al. 199767 untrained college studentsAll 7 formulas correlated highly (r > 0.95) — but only Mayhew and Wathen accurately predicted bench; only Wathen nailed squat; every formula underestimated deadlift by 9–14%
Reynolds et al. 200670 adults, 18–695RM predictions most accurate (R² = 0.993 bench); accuracy degraded substantially at 10RM and 20RM
Mayhew et al. 2008103 womenFormulas more accurate with fewer than 10 reps to failure
Ware et al. 199545 college football playersBench predictions beat squat; none accurate for squat (subjects averaged >10 reps — flawed test)
Dohoney et al. 2002Young adult males4–6RM predicted better than 7–10RM
LeMense et al. 2024 (meta-analysis)71 subjects, back squatVelocity-based estimates overestimated squat 1RM — bar-speed method not viable
Heo et al. 202535 male college studentsPrediction R² ranged 0.59–0.92 by exercise — accuracy is lift-specific

The pattern: high correlation is not accuracy. Every formula correlates above r = 0.90 — which sounds impressive until a formula tracks the trend perfectly while sitting 30 lb too low on every lifter. That’s exactly what happened with the deadlift.

Why Do Calculators Underestimate the Deadlift?

In LeSuer’s study, all seven equations significantly underestimated deadlift 1RM by 9–14% (22–34 lbs) despite correlating above 0.95. The formulas were developed and validated mostly on the bench press; the deadlift — no eccentric phase, full-body demand, grip as limiter — breaks the rep-to-max math they assume.

Practical takeaway: expect your real deadlift to run 5–15% above the estimate, and test deadlift maxes directly more often. Our deadlift max calculator shows all seven formulas so you can see the spread.

How Many Reps Give the Most Accurate Estimate?

Five or fewer. Reynolds (2006): 5RM hit R² = 0.993 on bench. Dohoney (2002): 4–6RM beat 7–10RM. Mayhew (2008): under 10 reps was the cutoff. At low reps the set measures maximal strength; past ~10 reps it measures endurance — a different quality — so the formula extrapolates from the wrong signal.

Does Training Experience Affect Accuracy?

Yes — beginners get the worst estimates. Beginners fail sets from technique breakdown, not muscular failure, so the formula’s core assumption breaks. Treat beginner estimates as programming starting points and re-estimate every 2–4 weeks. Trained lifters fail for the “right” reasons, which is what the formulas assume — Refalo et al. (2023) even found trained lifters predict their own reps-in-reserve within ±1 rep.

Are Bar-Speed (Velocity-Based) Calculators Accurate?

Less than the marketing suggests. A 2024 meta-analysis (LeMense et al.) found velocity-based estimates systematically overestimated back squat 1RM. Velocity tracking is a great training tool (fatigue management); as a 1RM prediction tool, classic reps-based formulas remain more trustworthy.

How Can You Make Your Estimate as Accurate as Possible?

  • Use 2–5 reps, close to failure. A 3-rep max predicts better than a 10-rep max, every time.
  • Use a big compound lift. The formulas weren’t built for curls.
  • Average multiple formulas. No single equation wins everywhere — LeSuer proved that in 1997. Our calculator shows all seven for exactly this reason.
  • Treat deadlift estimates with suspicion. Expect reality 5–15% higher.
  • Re-estimate regularly. Every 4–8 weeks; a 3-month-old estimate is a rumor.
  • Don’t mix lifts or units. Estimate bench from bench sets, kilos or pounds — pick one.

Frequently Asked Questions

Why do different calculators give me different numbers?
Different equations, different curves. Epley, Brzycki, and Lombardi are different mathematical models — they agree near 10 reps and diverge elsewhere. The spread is normal; the average is the most trustworthy single number.
Can I trust a 1RM estimate from 15+ reps?
Not much. Past 10 reps you’re measuring endurance and errors of 10–15% are common. Treat it as a rough starting point and re-estimate from heavier work soon.
Are 1RM calculators accurate for women?
Yes — Mayhew’s work validated prediction in women specifically, finding no meaningful gender difference in relative accuracy at ≤10 reps. The formulas care about reps, not sex.
Should I test my true 1RM instead of estimating?
For programming, no need. Test when you have a reason: a meet, a peaking block, or curiosity with solid technique. Safe protocol: how to test your 1RM safely.
What’s the single most accurate formula?
There isn’t one — it depends on the lift. LeSuer (1997): Mayhew and Wathen best for bench, Wathen best for squat, nothing accurate for deadlift. That’s why averaging beats picking.

Sources

  • LeSuer, D.A. et al. (1997). “The accuracy of prediction equations for estimating 1-RM performance in the bench press, squat, and deadlift.” Journal of Strength and Conditioning Research, 11(4).
  • Reynolds, J.M. et al. (2006). “Prediction of one repetition maximum strength from multiple repetition maximum testing and anthropometry.” Journal of Strength and Conditioning Research, 20(3).
  • Mayhew, J.L. et al. (2008). Validation of 1RM prediction in women. Journal of Strength and Conditioning Research.
  • LeMense, C. et al. (2024). Meta-analysis of velocity-based 1RM prediction. Journal of Strength and Conditioning Research.
  • Refalo, M.C. et al. (2023). Accuracy of repetitions-in-reserve prediction in trained lifters.

Keep reading — 1RM Basics silo

D
Daved
Strength training writer and SEO specialist. I build calculators and guides for lifters who want numbers they can actually program from — tested against the research, written from the gym floor.

Similar Posts