How to Predict 1RM Safely: Velocity-Based Estimation

Across 12 studies reviewed by Banyard et al. (2017, Int J Sports Physiol Perform), velocity-based 1RM prediction methods achieved mean errors of 2.0–4.5% compared to actual measured 1RM — accurate enough for load prescription in most training contexts and achieved without any maximal attempt. For a 150 kg squatter, that margin is ±3–7 kg: precise enough to program at 80% or 85% 1RM without meaningful error. This guide explains the two main velocity-based methods, how to choose between them, their actual error characteristics, and how to apply the predicted 1RM in daily programming.

Direct 1RM testing — incrementally loading a bar until only one repetition can be completed — is the gold standard for measuring maximal strength, but the standard carries costs that are often underweighted:

• Injury risk from technical failure: Maximal attempts place connective tissue (tendons, intervertebral discs, shoulder labrum) under peak mechanical stress. Without a spotter system and technical proficiency at near-maximal loads, injury rates climb. A 2018 survey of competitive powerlifters found that 59% had sustained a training injury in the preceding 12 months, with the majority occurring during high-intensity attempts (Aasa et al., J Strength Cond Res).
• Residual fatigue cost: A properly conducted 1RM test generates meaningful CNS and mechanical fatigue that requires 48–72 hours of reduced training to fully clear. In competition preparation cycles, this fatigue can disrupt the peak taper. Testing 1RM directly more than once per 4–6 week mesocycle is rarely justified.
• Daily variability: True 1RM fluctuates 3–8% day-to-day based on sleep quality, hydration, circadian rhythm, and psychological state (Jovanovic & Flanagan, 2014). A single maximal test captures one point on a distribution — not a stable baseline for load prescription.

Velocity-based prediction avoids all three problems. Testing can be done at submaximal loads (60–80% 1RM), takes under 15 minutes, generates minimal fatigue, and can be repeated daily to track the fluctuating real-world 1RM.

Before velocity-based methods became accessible, rep-based formulas were the primary estimation tool. The most widely used formulas convert the number of repetitions completed at a given load into a 1RM estimate:

Rep-based formulas have two fundamental limitations. First, error increases substantially above 10 repetitions — the formulas were calibrated on data from lower rep ranges and extrapolate poorly. Second, they do not account for fatigue state or inter-day variability. A fatigued athlete completing 8 reps at 80% on a bad day will generate a lower 1RM estimate than they would on a recovered day at the same load — but there is no way to distinguish genuine strength change from fatigue-driven variability.

Every athlete has a characteristic relationship between the load they lift and the velocity at which they move it. This load-velocity profile (LVP) is highly linear (R² typically 0.95–0.99) and is relatively stable across training states. It has two anchors:

• Maximum unloaded velocity (V0) — the fastest the limb can move without external resistance. In the back squat, this is typically 1.3–1.7 m/s for athletic populations.
• Minimum velocity threshold (MVT) — the slowest velocity at which any repetition can be completed, corresponding to the true 1RM. For the back squat, MVT is consistently 0.30–0.35 m/s across trained populations (González-Badillo & Sánchez-Medina, 2010, Int J Sports Med).

Because the LVP is linear, measuring mean concentric velocity at two or more known submaximal loads allows extrapolation to the load that would produce velocity at the MVT — which is by definition the 1RM. This is the conceptual foundation of all velocity-based 1RM prediction methods.

Typical MVT values by exercise:

• Back squat: 0.30–0.35 m/s
• Bench press: 0.16–0.20 m/s
• Deadlift: 0.12–0.18 m/s
• Hang clean: 0.70–0.90 m/s (power exercises use different thresholds)

The MVT method builds a full load-velocity profile using 4–6 submaximal loads and extrapolates to the 1RM. It is more accurate than the two-point method but requires more sets. Use it for baseline profiling at the start of each training block (every 4–6 weeks).

1. Warm-up: 5–8 minutes general warm-up, then 3 sets at 40%, 55%, and 70% of estimated 1RM for 3–5 reps each. Rest 3–4 minutes between warm-up sets.
2. Testing sets: Perform single repetitions (or 2–3 reps at lighter loads) at 60%, 70%, 80%, and optionally 85% of estimated 1RM. Record mean concentric velocity for each load. Rest 4–5 minutes between testing sets.
3. Plot the profile: Graph load (x-axis) vs. mean concentric velocity (y-axis) and fit a linear regression line through the data points.
4. Extrapolate to MVT: Extend the regression line to the exercise-specific MVT (e.g., 0.32 m/s for back squat). The load at that velocity is the predicted 1RM.
5. Apply load zones: Use the profile to derive daily training loads without additional testing. A session target of 0.55 m/s corresponds to approximately 75–78% 1RM for most athletes.

The two-point method reduces the protocol to just two loads — one light (45–55% estimated 1RM) and one moderately heavy (70–80% estimated 1RM) — and uses the two velocity measurements to construct a minimal load-velocity line, then extrapolates to the MVT. This method is ideal for daily session monitoring because it adds only 10–12 minutes to the warm-up.

Accuracy tradeoff: the two-point method shows standard error of ±3–5% for the back squat (Jovanovic & Flanagan, 2014), compared to ±2–3% for the full MVT method. For in-session load prescription, this margin is acceptable. For research-grade baseline profiling, use the full MVT method.

Daily two-point protocol:

• Set 1: 45% estimated 1RM × 3 reps. Record mean concentric velocity.
• Rest 3 minutes.
• Set 2: 75% estimated 1RM × 2 reps. Record mean concentric velocity.
• Extrapolate to MVT using the two-point line.
• If predicted 1RM is more than 5% below the previous session's value, consider reducing session load or investigating recovery status before proceeding.

No prediction method is perfectly accurate. Coaches who build daily training around predicted 1RM values should understand the actual error characteristics to set appropriate precision expectations:

The key insight from this comparison is that the velocity-based methods capture day-to-day 1RM fluctuation that rep formulas and direct tests cannot. An athlete whose true 1RM fluctuates from 145 kg to 158 kg over the week (an 8.5% swing, consistent with research findings) will have that variation detected by velocity-based testing, allowing load to be adjusted accordingly. Rep formulas and infrequent direct tests miss this variation entirely.

Once a daily predicted 1RM is established, it replaces the static 1RM in your percentage-based programming. This is a critical shift: instead of prescribing 80% of a 6-week-old tested 1RM, you prescribe 80% of today's actual capacity — accounting for fatigue, recovery, and readiness.

Practical velocity zones for common training goals:

• Maximal strength (90–100% 1RM): Mean concentric velocity 0.30–0.45 m/s. Use only 1–3 reps per set; these velocities confirm you are within 5–10% of the true 1RM.
• Strength-speed (80–90% 1RM): 0.45–0.60 m/s. Primary range for strength-power development. 3–5 reps per set.
• Speed-strength (60–80% 1RM): 0.75–1.00 m/s. Primary range for power output and VBT protocols. 3–6 reps with maximal intent.
• Hypertrophy (65–75% 1RM): 0.55–0.80 m/s. Higher rep ranges (6–12 reps), moderate velocity targets.

Two implementation rules ensure the prediction is applied correctly:

1. Re-run the two-point protocol at the start of each session — not from the previous session's predicted 1RM. Carry-over error compounds rapidly if you chain predictions from prior days.
2. When the predicted 1RM drops 5% or more from a stable 14-day average, reduce session intensity by 5–10% rather than pushing through the apparent decline. Persistent velocity drops that do not recover within 3 days signal accumulated fatigue requiring a deload.