How to Calculate 1RM from Velocity Data

Attempting a true 1RM test carries a reported injury rate of 1.3 per 1,000 sessions in competitive powerlifters (Keogh et al., 2006) — a risk that grows prohibitive in-season, during rehabilitation, or with novice athletes. Velocity-based 1RM estimation sidesteps this risk entirely: by measuring the bar speed at two or more sub-maximal loads, a linear regression line can extrapolate the load at which velocity would reach zero, which corresponds closely to maximal strength. The prediction error for trained athletes ranges from ±2.5–5% in published validation studies, well within the day-to-day biological variability of actual 1RM performance.

The load-velocity relationship in resistance exercise is highly linear between approximately 30% and 100% of 1RM. As load increases, mean propulsive velocity decreases in a predictable fashion. González-Badillo and Sánchez-Medina (2010) established this relationship across the full intensity spectrum in the squat, demonstrating r-values above 0.98 for the load-velocity regression in trained athletes.

This linearity means that if you measure velocity at two or more known loads, you can construct a regression line and extend it to estimate the load at which velocity would theoretically reach zero (the load-axis intercept). That intercept load is your predicted 1RM.

The practical advantages over traditional 1RM testing are significant: no maximal strain on the musculoskeletal system, shorter warm-up protocols, repeatable weekly monitoring without accumulated fatigue, and the ability to detect strength changes (the regression line shifts left when stronger, right when weaker) before they manifest in perceived exertion or form breakdown.

The minimum velocity threshold (MVT) is the mean concentric velocity at which an athlete can no longer complete the concentric phase — effectively their 1RM velocity. Across exercises and populations, MVT values are remarkably consistent within an individual, though they vary between exercises.

García-Ramos et al. (2018) showed that MVT is sufficiently stable within an individual that a single MVT measurement can serve as a fixed intercept in subsequent 1RM estimations without retesting — a major time-saver in field settings. This is the 'individualized' two-point method now recommended over population-derived MVT constants.

A full load-velocity profile requires velocity measurements at 4–6 loads spanning 40–85% of estimated 1RM. This gives a high-precision regression line and establishes the MVT for future reference. Protocol:

1. Complete a thorough warm-up: 10 minutes low-intensity aerobic work, followed by specific warm-up sets at 40%, 55%, and 70% of estimated 1RM (3 reps each, 2 minutes rest).
2. Begin profiling at 40% estimated 1RM. Perform 3 reps with maximum intentional concentric velocity. Record MCV per rep; use the best value (or mean, consistently).
3. Increase load by 10% absolute 1RM increments. Rest 3 minutes between loads at lighter intensities; 4 minutes above 75%.
4. Continue until reaching approximately 85–90% of estimated 1RM. Do not attempt true 1RM — the regression will project it.
5. Plot load (kg or lb) on the x-axis, MCV on the y-axis. Fit a least-squares linear regression. The x-intercept (where the line crosses zero velocity) is your predicted 1RM.

A full profile takes approximately 45–60 minutes including warm-up and is best performed at the start of a training block, then updated every 4–6 weeks.

Two regression approaches are used in practice:

Population-derived equation: Uses pre-established coefficients (slope and intercept) derived from large athlete samples. Example for back squat (González-Badillo & Sánchez-Medina, 2010): %1RM = 121.1 − (58.5 × MCV). Rearranged: estimated 1RM = measured load ÷ (%1RM/100). This requires only a single sub-maximal measurement but introduces individual-level error of ±5–8%.

Individualized regression: Built from the athlete's own multi-point profile. Standard error of estimate (SEE) typically falls to ±2–4% once the individual's regression line is established (Jovanovic & Flanagan, 2014). This approach requires more time upfront but is far more accurate for ongoing weekly monitoring.

The trade-off: population equations are faster and acceptable for initial screening; individual regressions are the clinical standard for precision training decisions.

For weekly monitoring where a full profile is impractical, the two-point method (García-Ramos et al., 2018) requires only two load-velocity pairs measured within a standard warm-up. Protocol:

1. Measure MCV at ~55% of estimated 1RM (3 reps, take the best).
2. Measure MCV at ~75% of estimated 1RM (3 reps, take the best).
3. Use the athlete's stored MVT from the full profile as the lower anchor point (zero velocity).
4. Fit a line through the two measured points anchored to MVT. Read off the x-intercept as predicted 1RM.

The two-point method adds only 8–12 minutes to a warm-up and achieves SEE of ±3–5% relative to the full profile. It is the recommended monitoring approach once an athlete's full profile has been established at the start of a block.

Velocity-based 1RM prediction is not perfectly accurate, and understanding error sources prevents misinterpretation. The main sources of error are:

• Sub-maximal intent on test reps: The regression assumes maximum intentional velocity on every rep. If an athlete sandbags or is unfamiliar with maximal intent cues, velocity at any load will be artificially depressed, causing the predicted 1RM to underestimate true strength by 5–10%.
• Daily readiness variation: MCV at any given load fluctuates ±3–7% due to sleep, hydration, and circadian rhythm. A predicted 1RM on a low-readiness day may underestimate true 1RM by up to 8%. Build a 7-day rolling baseline before making load-increase decisions.
• Technical changes: Any modification in technique (grip width, stance, depth) shifts the load-velocity profile. Regressions built under old technique are invalid once technique changes substantially.

Rebuild the full profile when: the athlete has trained for 8+ weeks since the last profile, technique has been modified, or consecutive weekly predicted 1RM values show a step-change rather than gradual progression.

Different exercises have different characteristic velocities at given %1RM. Using a squat equation to estimate a bench press 1RM will produce large errors. The table below provides reference MCV values at 60, 70, and 80% 1RM for commonly profiled exercises, drawn from González-Badillo, Sánchez-Medina, and García-Ramos research groups:

These ranges are for trained athletes performing maximum intentional velocity on each rep. Use them as a sanity check: if your measured MCV at 70% falls outside these ranges, check warm-up standardization, intent cues, and device placement before concluding the regression is off.