Squat 1RM math is identical to bench — interpretation isn't
The four prediction formulas are lift-agnostic. Epley, Brzycki, Lombardi, and O'Conner each fit a curve to the load-vs-reps relationship; none of them know which lift you're doing. So a 100 kg × 5 squat and a 100 kg × 5 bench produce the same numerical estimate. The reliability bands from LeSuer 1997 also apply identically (±2% at ≤5 reps, widening past that).
What changes is how to interpret the result against population data and against your own programming. Squat is more variant-sensitive than bench — back vs front vs box, high-bar vs low-bar, belted vs raw, depth conventions, all materially shift what the same external load “means.” A 200 kg low-bar belted squat to a higher box and a 200 kg high-bar raw squat to depth are not the same accomplishment, even though the calculator returns the same 1RM estimate.
A worked example — 140 kg × 3 squat
Run 140 kg × 3 reps on back squat. Same four-formula math as bench:
- Epley: 140 × (1 + 3/30) = 154.0 kg
- Brzycki: 140 × 36 / (37 − 3) ≈ 148.2 kg
- Lombardi: 140 × 3^0.1 ≈ 156.3 kg
- O'Conner: 140 × (1 + 3 × 0.025) = 150.5 kg
Cross-formula average ≈ 152.3 kg squat 1RM with a HIGH reliability band of ±2% (~149.3–155.3 kg). Three reps is well inside the high-confidence zone, and the 8 kg spread across formulas is what to expect at this rep count. If you walked away from a set-of-3 thinking 160 kg might have been there, the math says the honest top end is closer to 155.
For programming weights off this number, the percentage-of-1RM calculator surfaces both the NSCA traditional bands and the ACSM 2026 widening — squat tends to respond well to the wider hypertrophy band (load ≥30% with RIR ≤3) more than bench does, but the math is the same either way.
Frequently asked
Is back squat 1RM the same as front squat or high-bar 1RM?
No. Front squat typically lands at 70–85% of back squat for the same lifter, and the gap between high-bar and low-bar back squat usually runs 5–15% (low-bar carries more weight for most). The math here doesn't know which variant you trained — it estimates 1RM from your submax weight and reps. The number is specific to whatever variant you actually performed. Don't cross-compare front squat estimates to back squat numbers without that adjustment.
Should the squat 1RM estimate include depth?
The estimate doesn't know your depth — it computes from weight and reps. Powerlifting-legal depth (hip crease below knee crease) is the assumption most lifters use when reporting squat numbers. If you're squatting to a higher box or stopping at parallel-but-shy, your reported 1RM tracks heavier than a depth-judged number would. The math doesn't care; the comparison to standards does. Same load through a shorter ROM is a different lift.
What rep range gives the most reliable squat 1RM estimate?
≤5 reps is the HIGH reliability band (~±2%) per LeSuer 1997's validation across bench/squat/deadlift. 6–10 reps is MEDIUM (~±5%). Past 10 reps the estimate is a programming hint, not a true 1RM — error widens to ±10% or more. Squat sets to failure at 12+ reps are particularly noisy because cardiovascular limits start mixing with strength limits, and the prediction formulas were never calibrated for that interaction.
Why is my squat 1RM different across the four formulas?
The four formulas (Epley, Brzycki, Lombardi, O'Conner) fit different curve shapes to the rep-weight relationship and disagree more as reps increase. At ≤5 reps they typically agree within 4–5%. At 8 reps the spread can be 8–10%; at 12 reps it can exceed 15%. The cross-formula average smooths the disagreement, but the underlying truth is that nobody's squat 1RM-vs-rep curve perfectly matches any single formula. The four-formula spread on your number is itself useful information about how confident to be in the headline.
Does the formula treat low-bar and high-bar differently?
No. The math is bar-position-agnostic — it takes weight and reps and returns an estimate. Population-level data suggests most experienced lifters back-squat 5–15% more low-bar than high-bar (lower bar position lets the hips contribute more), but that's an interpretive overlay, not a math difference. Estimate from whichever style you trained, and label your reported number accordingly.
What I'd do next
- Estimate deadlift 1RM separately
Don't infer deadlift from squat. Population ratios are loose (deadlift typically 1.0–1.3× squat) and your individual ratio is the only one that matters.
- Build squat working weights
Same %1RM table, applied to your squat number. Strength / hypertrophy bands surface both NSCA traditional + ACSM 2026.
- Depth + variant decisions
What depth and bar-position choices change about the same external load.
Also in this cluster
By Jimmy L Wu. Engine shared with the general 1RM calculator — same four published formulas (Epley 1985, Brzycki 1993, Lombardi 1989, O'Conner et al. 1989), same LeSuer 1997-grounded reliability bands. Squat-variant ratios (front:back ~70–85%, low-bar:high-bar +5–15%, belted:raw +5–15%) are training-population observations, not conversion factors — track your own ratio over time rather than calculating off published averages. Engine logic in lib/strength/oneRepMax.ts. Not medical advice — for max attempts, work with a qualified strength coach.