Training max vs one-rep max: use the right number

"Do 5 reps at 85%." Eighty-five percent of what? If your program means your training max and you use your one-rep max, the same instruction can turn a solid working set into a near-maximal grinder. Most loading mistakes in percentage-based programs come from using the wrong number in the bottom of the fraction.

This guide defines the three numbers that get confused (tested 1RM, estimated 1RM and training max), shows with worked kg and lb examples how much the bar weight changes depending on which one you use, explains rounding to real plates, and tells you when an estimate is too uncertain to build a program on.

TL;DR

• Tested 1RM: the heaviest weight you actually lifted for one rep with your normal technique.
• Estimated 1RM: a calculated number from a heavier rep set (Epley, Brzycki or similar). Useful, but it carries error.
• Training max (TM): a deliberately reduced number that a program bases its percentages on. Jim Wendler's beginner 5/3/1 article uses 80–85% of your max; his 5/3/1 Beach Body challenge uses 90% of your best gym lift. There is no universal TM percentage.
• The same "85% × 5" can mean 120 kg, 107.5 kg or 100 kg for a lifter with a 140 kg squat, depending on the denominator.
• Keep the exact TM in your log, round each set to plates you own, and when unsure between two numbers, use the lower one.
• An estimate from more than about 10 reps, a guessed RIR, a changed technique or an old max is too shaky to set a TM from. Retest with a heavier set of 3–6 reps.

What to do this week

• Find your program's instructions and write down exactly what its percentages are "of": 1RM, TM, a 5RM, or nothing (some programs add weight session to session without percentages).
• If you need an estimate, use a recent heavy set of 3–6 reps with good form, stopped with a few reps left if you prefer, and run it through the 1RM calculator with your reps in reserve. You do not need to grind to failure to get a usable number.
• Calculate your TM with the template's percentage, and round down if unsure.
• Before your first session, work out every set's bar weight and check it with the plate calculator.

Three numbers, three jobs

Tested 1RM

A tested 1RM is the most precise number you can have, but only for that day, that lift and that technique. It depends on skill with heavy singles, fatigue, sleep and how the attempt was judged. A gym single with a high squat is not the same number as a competition-depth single.

You do not need to test one. If you do, build up with warm-up singles, use a spotter or safety arms, and stop when bar speed or technique drops. A coach who can see you lift is the best judge of when a heavy attempt makes sense.

Estimated 1RM

An estimate converts a rep set into a single. Our 1RM calculator shows two common formulas and their midpoint:

• Epley: 1RM = weight × (1 + reps ÷ 30)
• Brzycki: 1RM = weight × 36 ÷ (37 − reps)

Example: 120 kg × 5 reps to failure gives 140 kg (Epley) and 135 kg (Brzycki), a midpoint of 137.5 kg (about 303 lb). A separate pound example: 265 lb × 5 gives about 309 lb and 298 lb, a midpoint of about 304 lb.

The two formulas disagreeing by 5 kg is not a confidence interval. It is a reminder that the formulas make different assumptions, and the real error for you could be larger.

Training max

A training max is a number chosen to be lighter than your true max so that the program's percentages produce work you can complete with good technique, cycle after cycle. In both of Jim Wendler's 5/3/1 articles cited here, all percentages are based on the TM, not on your 1RM.

The TM percentage depends on the template:

• The 5/3/1 Philosophy for Beginners (Wendler, 2011): take 80–85% of your max, and if in doubt, err on the side of too light. See our 5/3/1 for beginners section.
• 5/3/1 Beach Body challenge (Wendler): the TM is 90% of your best gym lift, not a competition lift.

Other programs use other denominators. In GZCLP, Cody Lefever's novice linear progression, the T1 lift restarts at 85% of a newly tested 5RM, so that percentage is of a five-rep max, not a one-rep max. Starting Strength (Mark Rippetoe) and StrongLifts 5×5 (Mehdi Hadim) do not use percentages of a max at all; they add weight session to session.

Because a training-max program takes its set percentages from the TM, every set is a percentage of a percentage of your max. With an estimated max of 100 kg and a TM at 85%, the TM is 85 kg, and a set at 75% of that TM is 63.75 kg before plate rounding: 63.75% of the estimated max, not 75%. These are illustrative numbers, not a recommendation for your program.

Worked example: what is the percentage "of"?

Take a lifter with a tested squat of 140 kg and one written set: 85% × 5. The pound table is a separate example for a lifter with a tested 315 lb squat, rounded to conventional 5 lb jumps.

Kilograms (rounded to 2.5 kg)

Pounds, separate example (tested 315 lb, rounded to 5 lb)

Why the top row matters: the Epley formula estimates that someone with a 140 kg max can do about 120 kg for 5 reps to failure. So "85% × 5" of a 1RM is roughly a five-rep max by that estimate, while the same instruction applied to an 80–90% TM is a hard but submaximal set. Use the wrong denominator in a TM-based program and every set gets about 11–25% heavier than intended (1 ÷ 0.9 to 1 ÷ 0.8).

A template example: 5/3/1 Beach Body

The Beach Body challenge sets the TM at 90% of your best gym lift and, for many of its lifts, uses five reps on every work set across a three-week cycle. With a 140 kg best squat, the TM is 126 kg:

For a separate pound example, a 315 lb best squat gives a 283.5 lb TM and week-3 sets of 215 / 240 / 270 lb.

Notice the last set of week 3: 120 kg × 5, about 86% of the lifter's best, which is close to the Epley-estimated five-rep max. That only works if the best lift was real. If the 140 kg had been an inflated estimate and the true max were lower, that set becomes a grind or a miss. This is exactly why Wendler's beginner article says to err on the side of too light.

After the first three-week cycle, the challenge raises the TM by a fixed amount (10 lb for the squat and deadlift, 5 lb for the press) and repeats the percentages. We summarize only enough to show the arithmetic; read the original for the full challenge.

Rounding to plates

Percentages almost never land on a loadable number. Use these rules:

• Store the exact TM (for example, 119 kg or 267.75 lb) in your log. Round only the bar weight for each set. That stops rounding errors from compounding across cycles.
• Round each set to your smallest jump. With 1.25 kg plates the jump is 2.5 kg; with 2.5 lb plates it is 5 lb. Fractional plates allow smaller steps.
• If you must round the TM itself, round down. That follows Wendler's advice to err on the light side.
• Know your bar. Most barbells are 20 kg or 45 lb, but not all. The plate calculator lets you set the bar weight and shows the plates per side.

How much does rounding matter? On a 107.1 kg target, 107.5 kg is 0.4% off. If your gym has nothing smaller than 2.5 kg plates (5 kg jumps), your choices are 105 kg or 110 kg, up to 2.7% off. In pounds, a 241 lb target becomes 240 lb with 5 lb jumps, or 235 or 245 lb with 10 lb jumps. Small differences, but they add up on the heaviest set of a cycle.

How to set a training max: decision rules

1. Recent tested single, good technique? Use it with your template's TM percentage.
2. No single, but a recent heavy set of 3–6 reps with an honest reps-in-reserve rating? Estimate the 1RM, then apply the TM percentage.
3. Template says "best gym lift"? Use your best recent lift under normal gym conditions, as the Beach Body page specifies, not a competition result.
4. Torn between two numbers? Use the lower one.
5. First week feels too heavy? If the prescribed reps on your top sets come with slow or grinding reps, your TM is too high. Lower it by a few percent now rather than missing reps later. Wendler's beginner article does not allow a TM increase after a cycle with grinding reps.

To rate reps in reserve for rule 2, see our guide to RPE and reps in reserve. It also explains why newer lifters, on average, rated true maximal efforts lower than experienced lifters in the research.

When an estimate is too uncertain

Treat an estimated 1RM as too uncertain to set a TM from when:

• The set was more than about 10 reps. Prediction studies found accuracy got worse above roughly 10 reps to failure, and one recommended using no more than 10 reps in these equations, with 5RM data predicting best.
• The reps in reserve were a guess. Research on RIR-based ratings found less experienced lifters rated true maximal efforts lower than experienced lifters did.
• The lift was a deadlift and you rely on a formula alone. In one study of untrained students, every prediction equation tested significantly underestimated the deadlift.
• Your technique or range of motion changed since the set (deeper squats, a paused bench, a new grip).
• The max is old, or you have since had a layoff, an illness or a large bodyweight change.
• It is a snatch or clean & jerk. These lifts are limited by speed and position; rep-based formulas do not fit them.

A large 2024 meta-regression also found that the number of reps people can do at a given percentage of 1RM differs between individuals and between exercises. A formula is a population average applied to you.

What to do instead: test a heavier set of 3–6 reps with good form, run the first week of your program and adjust the TM from how it went, or start deliberately light and let the template's increases catch you up.

Reviewing progress

• A rising TM is a planning change, not proof of strength. Templates raise the TM on a schedule. Check that the lifts are actually moving.
• Track rep records at fixed weights. If 107.5 kg × 5 becomes 107.5 kg × 8 over a few cycles, you are stronger.
• Track estimated 1RM from comparable sets. Use the same rep range each time; an estimate from a triple and one from a set of 10 are not comparable.
• Test a single occasionally, if you want one. Many lifters never need to. If you do, test when rested, with spotters or safety arms, and ideally with a coach watching.
• Compare like with like. If you want to see where a competition-standard lift sits among other lifters, read strength standards explained first.

Common mistakes

• Using your 1RM where the program says TM. Every set becomes too heavy.
• Setting the TM from a gym max with high squats or bounced benches. The TM inherits the inflation.
• Raising the TM after a cycle full of grinders. Hold it, or lower it.
• Re-rounding the TM every cycle. Store the exact value; round only the bar weight.
• Mixing templates. Using one template's TM percentage with another template's set percentages produces a program nobody wrote. Label any change as your modification.
• Treating an estimate as an attempt target. An estimate is a planning number, not a weight you should expect to lift on demand.

Pillars Check

Workout. A correct denominator keeps the program's intended difficulty. Choose a template, use its TM rule, and progress by its schedule. Our Workout pillar and the 5/3/1 guide cover structure; Starting Strength vs StrongLifts covers programs that do not use percentages.

Diet. TM increases assume you are recovering and, for many lifters, gaining or at least keeping muscle. If your sets feel heavier each cycle while your bodyweight drops, look at food intake before blaming the TM. The Diet pillar has practical targets.

Recovery. Tested maxes and estimates from tired days both mislead. Adults should regularly get seven or more hours of sleep, per the AASM and Sleep Research Society consensus. Set or review your TM after a normal week, not after a run of short nights. See the Recovery pillar.

FAQ

Sources

• Wendler J. The 5/3/1 Philosophy for Beginners, 2011. Supports the 80–85% TM, erring on the light side, percentages based on the TM, and not raising the TM after grinding reps.
• Wendler J. 5/3/1 Beach Body (Performance Based) Challenge. Supports the 90% TM of best gym lift, the three-week percentages and the TM increases.
• Lefever C (GZCL). GZCL Applications and Adaptations, 2016. Supports the T1 restart at 85% of a newly tested 5RM.
• Brzycki M. Strength testing: predicting a one-rep max from reps-to-fatigue. J Phys Educ Recreat Dance. 1993;64(1):88–90. doi:10.1080/07303084.1993.10606684. Source of the Brzycki formula.
• Epley B. Poundage Chart. Boyd Epley Workout. Lincoln, NE: Body Enterprises; 1985 (as cited in Richens and Cleather 2014). Source of the Epley formula.
• Reynolds JM, Gordon TJ, Robergs RA. Prediction of one repetition maximum strength from multiple repetition maximum testing and anthropometry. J Strength Cond Res. 2006;20(3):584–592. doi:10.1519/R-15304.1. Supports using no more than 10 reps and 5RM data predicting best.
• Mayhew JL, Prinster JL, Ware JS, Zimmer DL, Arabas JR, Bemben MG. Muscular endurance repetitions to predict bench press strength in men of different training levels. J Sports Med Phys Fitness. 1995;35(2):108–113. PubMed 7500624. Supports worse accuracy above 10 reps.
• LeSuer DA, McCormick JH, Mayhew JL, Wasserstein RL, Arnold MD. The accuracy of prediction equations for estimating 1-RM performance in the bench press, squat, and deadlift. J Strength Cond Res. 1997;11(4):211–213. doi:10.1519/00124278-199711000-00001. Supports the deadlift underestimate.
• Nuzzo JL, Pinto MD, Nosaka K, Steele J. Maximal number of repetitions at percentages of the one repetition maximum. Sports Med. 2024;54(2):303–321. doi:10.1007/s40279-023-01937-7. Supports individual and exercise differences in reps at a given %1RM.
• Zourdos MC, et al. Novel resistance training-specific rating of perceived exertion scale measuring repetitions in reserve. J Strength Cond Res. 2016;30(1):267–275. doi:10.1519/JSC.0000000000001049. Supports experience differences in rating maximal efforts.
• Watson NF, et al. Recommended amount of sleep for a healthy adult. Sleep. 2015;38(6):843–844. doi:10.5665/sleep.4716. Supports seven or more hours of sleep for adults.