How to Predict 1RM Without Maxing Out: Submaximal Methods, Velocity-Based Estimation, and Practical Protocols

The one-repetition maximum (1RM) — the heaviest load an athlete can lift for a single complete repetition with correct form — is the cornerstone of strength training prescription. Nearly every percentage-based program relies on a current 1RM to calculate working loads. But traditional 1RM testing has serious drawbacks: it demands maximal effort that carries injury risk, requires extensive warm-up and recovery time, fatigues the athlete for subsequent training, and produces a number that begins depreciating the moment it is measured because strength fluctuates daily.

Fortunately, you do not need to actually lift your maximum to know what it is. Submaximal prediction methods allow accurate 1RM estimation from loads the athlete can handle comfortably and safely. The most advanced of these methods — velocity-based 1RM estimation — can even track 1RM changes in real time during regular training sessions, eliminating the need for dedicated testing days entirely.

This guide covers the three main categories of submaximal 1RM prediction, explains the science behind each, provides step-by-step protocols, compares their accuracy, and shows how to implement daily 1RM tracking for training autoregulation.

Before diving into prediction methods, it is worth understanding why submaximal estimation is often preferable to traditional 1RM testing — even when testing is logistically feasible.

Injury risk. Maximal attempts push tissues to their structural limits. While injury rates during well-supervised 1RM testing are low in research settings, the practical context matters. Athletes testing 1RM at the end of a training phase may have accumulated fatigue, minor tissue irritation, or technique degradation — all of which increase risk under maximal load. For exercises with high technical demands (Olympic lifts, deadlift) or significant spinal loading (back squat, overhead press), the risk-reward calculus of maximal testing warrants careful consideration.

Fatigue cost. A traditional 1RM test requires progressive warm-up sets and typically 3-5 maximal attempts, often consuming 30-45 minutes for a single exercise. The fatigue generated compromises training quality for 24-72 hours afterward. In sports where strength testing must coexist with skill work, conditioning, and game preparation, this fatigue cost is difficult to justify — especially when submaximal methods can produce equivalent information without the recovery debt.

The 1RM is not static. Perhaps the most compelling argument for submaximal prediction is that the true 1RM fluctuates daily based on sleep quality, nutrition, stress, accumulated training load, and other factors. A 1RM tested on Monday morning after a weekend of rest may be 5-10% higher than the true 1RM on Thursday afternoon after three training sessions. A number that is already outdated when recorded has limited utility for daily load prescription. Velocity-based estimation addresses this by providing a current-day 1RM estimate from warm-up sets, allowing training loads to reflect the athlete's actual capacity on that specific day.

Population-specific concerns. Certain populations should rarely or never perform true 1RM testing:

• Youth athletes — Technical proficiency may not be sufficient to safely express maximal force. Prediction from 3-5RM sets is preferred.
• In-season athletes — The fatigue cost is unacceptable when game performance takes priority.
• Rehabilitation patients — Maximal loading of healing tissue is contraindicated. Submaximal estimation provides training targets without exceeding tissue tolerance.
• Large groups — Testing 30 athletes' 1RM on multiple exercises is logistically impractical. Submaximal methods allow group testing in a fraction of the time.

The question is not whether submaximal prediction is 'as good as' maximal testing — it is whether the small accuracy trade-off is offset by the substantial practical advantages. In most real-world contexts, it is.

The simplest and most widely used approach to 1RM prediction is performing a set to failure (or near failure) at a submaximal load, then using a mathematical equation to estimate 1RM from the load and repetitions completed. Dozens of equations have been published; the most validated include:

Epley Equation (1985):

1RM = Load × (1 + 0.0333 × Reps)

Simple and effective for rep ranges of 2-10. Example: 100 kg × 5 reps predicts 1RM = 100 × (1 + 0.0333 × 5) = 116.7 kg.

Brzycki Equation (1993):

1RM = Load × (36 / (37 - Reps))

Slightly more conservative at higher rep ranges. Same example: 100 × (36 / (37 - 5)) = 112.5 kg.

Lombardi Equation (1989):

1RM = Load × Reps^0.10

Power-law model that handles moderate rep ranges well. Same example: 100 × 5^0.10 = 117.5 kg.

Limitations of rep-max equations:

• Accuracy degrades above 10 reps. At 10+ repetitions, muscular endurance becomes a confounding factor. Two athletes with identical 1RM values can achieve very different rep counts at 70% 1RM based on fiber-type composition, training history, and pain tolerance. Predictions from 15RM or 20RM sets can be off by 10-15%.
• Exercise specificity matters. Most equations were validated on the bench press and back squat. They tend to underestimate 1RM for the deadlift (where the strength curve is favorable for higher reps) and overestimate for the overhead press (where fatigue accumulates faster).
• Requires true failure or near-failure sets. If the athlete stops 2-3 reps short of failure, the prediction will underestimate 1RM by approximately 5-8% per rep left in reserve. Accurate rep-max prediction requires genuine effort, which partially undermines the safety advantage over direct testing.
• No daily fluctuation sensitivity. The prediction reflects the athlete's capacity on the testing day but does not automatically adjust for subsequent daily fluctuations.

Practical protocol for rep-max prediction:

1. Warm up progressively to the test load (typically 80-90% of estimated 1RM).
2. Perform a single set to muscular failure or to the last rep that can be completed with acceptable technique.
3. Record the load and reps completed.
4. Apply 2-3 different equations and average the results. This averaging reduces the bias of any single equation.
5. Use 3-5RM loads for best accuracy (prediction error typically 2-5% of true 1RM).

Velocity-based 1RM estimation exploits a well-established biomechanical principle: as the load on a barbell increases, the maximum velocity at which it can be lifted decreases in a near-linear fashion. By measuring the velocity of submaximal lifts, you can extrapolate the load at which velocity would reach the minimum threshold for completing a repetition — your estimated 1RM.

The minimum velocity threshold (MVT) concept:

For each exercise, there is a characteristic velocity below which a repetition cannot be completed. This is the minimum velocity threshold, also called the 1RM velocity or V1RM. Research has established exercise-specific MVT values with remarkable consistency across athletes:

• Back squat: 0.30-0.35 m/s (mean concentric velocity)
• Bench press: 0.15-0.20 m/s
• Deadlift: 0.15-0.20 m/s
• Overhead press: 0.18-0.22 m/s
• Power clean: 0.75-0.85 m/s

The practical implication is powerful: if you know that an athlete's bench press 1RM occurs at approximately 0.17 m/s, you can estimate 1RM from any submaximal set by plotting the load-velocity relationship and finding the load corresponding to 0.17 m/s.

Single-point estimation:

The simplest velocity-based method uses a single submaximal set. The athlete performs 2-3 reps at a moderate load (60-85% of estimated 1RM) with maximal intent. Mean concentric velocity is measured. Using a generalized load-velocity slope for the exercise, 1RM is estimated. Jovanovic (2020) describes this approach using the formula:

Estimated 1RM = Load / (1 - (Velocity - MVT) / Slope)

Single-point estimation is fast and practical but less accurate than multi-point methods because it assumes a population-average load-velocity slope rather than the individual's actual slope.

Two-point method:

Measuring velocity at two different loads (e.g., 60% and 85% of estimated 1RM) allows calculation of the individual's load-velocity slope. The 1RM is estimated by extrapolating this line to the MVT. Research by Garcia-Ramos et al. (2018) demonstrated that the two-point method produces 1RM estimates within 2-4% of actual 1RM, approaching the accuracy of full load-velocity profiling with substantially less testing burden.

Protocol for two-point velocity-based 1RM estimation:

1. Warm up progressively to the first test load (approximately 60% of estimated 1RM).
2. Perform 2-3 reps at the first load with maximal concentric intent. Record mean velocity of the best rep.
3. Rest 3-4 minutes.
4. Load approximately 85% of estimated 1RM.
5. Perform 1-2 reps with maximal intent. Record mean velocity of the best rep.
6. Plot the two points (load vs. velocity), calculate the linear slope, and extrapolate to the exercise-specific MVT.
7. The load at which the line crosses the MVT is your estimated 1RM.

The entire procedure takes 8-12 minutes including rest periods, uses loads well within the athlete's capacity, and produces an estimate that can be updated every training session.

The most comprehensive and accurate submaximal 1RM prediction method is full load-velocity profiling. By measuring velocity at 4-6 different loads spanning a wide range of intensities, you construct an individualized load-velocity curve that captures the athlete's unique neuromuscular characteristics.

Protocol for full load-velocity profiling:

1. Select 4-6 loads spanning approximately 40-90% of estimated 1RM. Example for an athlete with estimated squat 1RM of 150 kg: 60 kg, 80 kg, 100 kg, 115 kg, 130 kg.
2. Ascending order: Always test from lightest to heaviest to serve as a progressive warm-up.
3. Reps per load: 2-3 reps at loads below 70%, 1-2 reps at loads above 70%. Record the fastest rep at each load.
4. Rest periods: 2 minutes between lighter loads, 3-4 minutes between heavier loads.
5. Maximal intent: Every rep must be performed with maximal concentric velocity. Submaximal effort invalidates the entire profile.
6. Plot and extrapolate: Plot load (x-axis) vs. mean concentric velocity (y-axis). Fit a linear regression. The load at which the line intersects the exercise-specific MVT is the estimated 1RM.

Why linear regression works:

The load-velocity relationship in most strength exercises is remarkably linear across the 40-100% 1RM range, with R-squared values typically exceeding 0.95 (Gonzalez-Badillo and Sanchez-Medina, 2010). This linearity means a simple best-fit line accurately describes the relationship, and extrapolation to the MVT produces reliable 1RM estimates. Some researchers have proposed quadratic or exponential models, but the additional complexity provides minimal accuracy improvement for most exercises and athletes.

Individual vs. group load-velocity relationships:

While the general shape of the load-velocity relationship is consistent across athletes, the specific slope and intercept vary based on individual fiber-type composition, neural characteristics, and training status. A velocity-dominant athlete (more fast-twitch, less training experience) will have a steeper slope — their velocity drops more rapidly as load increases. A force-dominant athlete (strength-trained, higher proportion of slow-twitch) will have a flatter slope — they maintain velocity better under heavy loads. Using an individualized profile rather than population averages improves prediction accuracy from approximately 5-7% error to 2-4% error.

Updating the profile:

The load-velocity profile changes as the athlete adapts to training. A strength-focused training block may flatten the slope (improved force at high loads). A speed-strength block may steepen it (improved velocity at light loads). Re-profile every 4-8 weeks to maintain prediction accuracy, or use the two-point method during regular training to check whether the slope has shifted significantly.

Important considerations:

• Exercise technique consistency — The profile is only valid if technique is consistent across loads. Changes in squat depth, pause duration, or bar path with increasing load introduce systematic error.
• Sensor quality — Velocity measurement accuracy directly affects 1RM prediction accuracy. A velocity measurement error of 0.02 m/s can translate to a 3-5% error in predicted 1RM. High-frequency sensors (800 Hz+) minimize this error.
• Fatigue during profiling — If the athlete fatigues across the 4-6 loads, heavier loads will show artificially low velocities, overestimating 1RM. Adequate rest between loads is essential.

Not all prediction methods are equally accurate. The table below summarizes the typical prediction error (expressed as a percentage of actual 1RM) for each method, based on published validation studies.

Key findings from the accuracy literature:

• Velocity-based methods outperform rep-max equations when applied correctly. The primary reason is that velocity-based methods do not require failure, eliminating the subjectivity of effort perception and rep termination.
• The two-point method offers the best accuracy-to-effort ratio. It requires only two loads and 8-12 minutes of testing but approaches the accuracy of full profiling. Garcia-Ramos et al. (2018) found no statistically significant difference in prediction accuracy between two-point and full profiling for the bench press.
• All methods are exercise-dependent. Predictions are most accurate for the bench press and back squat, where the force-velocity relationship is most linear and the movement pattern is most constrained. Predictions for the deadlift and Olympic lifts are slightly less accurate due to greater technique variability.
• Within-athlete accuracy is better than between-athlete accuracy. Once an athlete's individual load-velocity profile is established, subsequent predictions for that athlete are more accurate than applying population equations to a new athlete.

Which method should you choose?

• If you have a velocity sensor and want daily 1RM estimates: Two-point method integrated into warm-up sets.
• If you need a quick one-time estimate without a velocity sensor: Epley equation with 3-5RM load.
• If you want maximum accuracy for periodization planning: Full load-velocity profile at the start of each training block.
• If testing youth or rehabilitation populations: Rep-max equations with 5-8RM loads (avoiding both maximal loads and high-rep sets where form degrades).

The most advanced application of submaximal 1RM prediction is daily 1RM estimation, where the athlete's working 1RM is updated every training session based on warm-up set velocities. This enables true autoregulation — adjusting training loads to match the athlete's actual capacity on each specific day.

How daily 1RM tracking works:

1. During the initial profiling session, establish the athlete's individual load-velocity slope and their exercise-specific MVT.
2. In subsequent sessions, the athlete performs warm-up sets as normal, with maximal concentric intent on each rep. Velocity is measured at 2+ warm-up loads.
3. The system uses the warm-up velocities and the previously established slope to estimate that day's 1RM.
4. Working loads are calculated as percentages of the estimated daily 1RM, not the historical tested 1RM.

Example implementation:

An athlete has a back squat profiled 1RM of 160 kg with an established load-velocity slope. Today's warm-up:

The velocities are consistently 6-8% below the profile values, suggesting today's 1RM is approximately 150-152 kg rather than 160 kg. If the prescribed session calls for 4 sets of 4 at 82% 1RM, the working load is set at 123-125 kg (82% of 150-152) rather than 131 kg (82% of 160). This 6-8 kg reduction prevents the athlete from grinding through excessively heavy sets on a low day, reducing injury risk and managing fatigue accumulation.

Benefits documented in the research:

• Superior strength gains. Autoregulated programs based on daily 1RM estimation have produced equal or greater strength gains compared to fixed-percentage programs in several studies (Dorrell et al., 2020), likely because they prevent both underloading on good days and overloading on bad days.
• Better fatigue management. Athletes using daily 1RM autoregulation report lower perceived fatigue and show fewer markers of overreaching compared to fixed-load programs matched for average intensity (Shattock and Tee, 2022).
• Reduced missed reps. When loads are calibrated to daily capacity, athletes complete prescribed reps more consistently, accumulating the intended training volume without the incomplete sets that plague fixed-percentage programs.

Practical implementation tips:

• Maximal intent on every warm-up rep is non-negotiable. Submaximal effort during warm-ups produces artificially low velocities, leading to underestimated 1RM and insufficiently challenging working loads. Athletes must understand that warm-up reps are data collection points.
• Re-profile every 4-8 weeks. The load-velocity slope can shift with training adaptations. A short re-profiling session (4-6 loads) keeps the model calibrated.
• Set velocity loss thresholds for within-set fatigue. Beyond load selection, use velocity to determine when to terminate sets. A common threshold is stopping the set when mean velocity drops by 20% from the first rep — this prevents excessive fatigue while ensuring adequate mechanical tension for adaptation.
• Use a sensor with adequate sampling rate. Daily 1RM tracking requires precise velocity measurement at every session. Sensors below 200 Hz introduce enough noise that daily estimates become unreliable. At 800 Hz, session-to-session velocity measurement variability is below 2%, making genuine daily fluctuations detectable.