The 3-Test PR Protocol: A Defensible 1RM in 6-9 Weeks

TL;DR

A single 1RM test is unreliable. Form fatigue, daily readiness, technique drift on the heaviest attempt, and grinding reps that violate form all inflate or deflate the number. The three-test protocol takes three separate tests across 6-9 weeks at different rep ranges and triangulates the results, producing a confidence band rather than a single point estimate. This is the methodology Body Unbound’s free template (207 upvotes on Reddit) baked in, and we’ve adopted it for Coach Sheet’s Lifting Records tab.

Why one test isn’t enough

A traditional max test asks the lifter to attempt 1 rep at the heaviest possible load. Three problems:

1. Day-to-day variance. True 1RM fluctuates 3-5% based on sleep, food, mood, accumulated fatigue. Catch a lifter on a bad day, undershoot. Catch them on a good day, you set a target they can’t replicate next week.
2. Form degrades at maximum loads. Most lifters’ technique drifts on the heaviest set of the day. The 1RM you measure is partly “best lift with valid form” and partly “best lift period”, and the line between blurs at the top.
3. Grinder reps inflate the estimate. A rep that took 4 seconds and required visible compensation isn’t a clean 1RM. It’s a 1RM-with-asterisk. But coaches and lifters tend to record it as a clean 1RM anyway.

A single test produces a number you have low confidence in. Three tests, structured carefully, produce three numbers you can compare.

The protocol

Across 6-9 weeks (typically the back end of a strength block, 2 weeks apart):

Test 1, 5 reps RPE 8 (week 1 of test phase) The lifter performs a top set of 5 reps stopping at RPE 8 (2 reps in reserve). Estimate 1RM via median of Epley/Brzycki/Lombardi with RIR adjustment.

Example: 100 kg × 5 reps RPE 8 → estimated 1RM ≈ 119 kg (median of three formulas).

Test 2, 3 reps RPE 9 (week 3-4 of test phase) Top set of 3 reps at RPE 9 (1 rep in reserve). Same median calculation.

Example: 110 kg × 3 reps RPE 9 → estimated 1RM ≈ 123 kg.

Test 3, 1 rep RPE 9-10 (week 5-7 of test phase) True near-max attempt. RPE 9 if a coach is conservative about technique; RPE 10 if the lifter is experienced and the form holds.

Example: 120 kg × 1 rep RPE 10 → estimated 1RM = 120 kg.

The three estimates form a confidence band: ~119, ~123, ~120 kg. Take the median (120 kg) as the working 1RM, or the mean (120.7 kg) if you want to be slightly more aggressive. The width of the band (4 kg in this example) signals confidence: tight band = reliable, wide band = something inconsistent across the tests.

Why three tests at different rep ranges

The three formulas (Epley, Brzycki, Lombardi) have different accuracy profiles at different rep ranges. Brzycki is most accurate at 1-5 reps; Epley is general-purpose; Lombardi handles higher reps better. Testing at 5, 3, and 1 reps lets each formula contribute its most-accurate prediction:

• 5-rep test favors Lombardi
• 3-rep test favors Brzycki and Epley equally
• 1-rep test is itself the 1RM (no formula needed)

Triangulating the three reduces formula bias.

When the band is wide

A confidence band wider than 5% of the lifter’s 1RM means something inconsistent across the tests. Possible causes:

Form degraded at the heavy 1-rep attempt. The lifter compensated, the 1-rep number is inflated. Repeat the 1-rep test 1-2 weeks later with stricter form.

The 5-rep test was conservative. The lifter rated RPE 8 but it was RPE 6.5-7. The estimate undershoots. Re-rate honestly next time.

Daily variance. One of the three tests fell on a particularly bad or good day. Compare with the lifter’s logged sleep, stress, food. Re-test the outlier.

Technique improved between tests. A novice lifter improving technique mid-block can produce genuinely higher 1RMs at later tests. This isn’t a confidence-band problem; it’s actual progress. Take the latest test as the working 1RM.

When to retest

For most lifters, test the main lifts (squat, bench, deadlift, OHP, row) once per mesocycle (every 6-12 weeks). For a powerlifter approaching a meet, test every 4-6 weeks during the peaking block. For novices, every 4-8 weeks because their 1RM is moving fast.

What “main lifts” means here

The protocol works for any compound lift the lifter trains regularly: barbell squat (any variation), deadlift (conventional or sumo), bench press, overhead press, barbell row, weighted chin-up. It does not work for:

• Accessory lifts (curls, lat raises). Too few reps in the working range
• Bodyweight movements that aren’t loadable
• Short-range partial reps
• Anything the lifter trains less than once per week (no recent rep history to predict from)

Logging it

Coach Sheet’s Lifting Records tab has a structured layout for the three tests per main lift:

The “Confidence” column flags the band width, prompting the trainer to decide whether the working 1RM is solid enough to drive next mesocycle’s loads or needs a retest.

Trade-offs

The protocol takes 6-9 weeks. A trainer who needs an estimated 1RM today for a new client doesn’t have that time. For first-week prescriptions, use a single working set (5-8 reps RPE 7-8) and the median of the three formulas. Update with the proper protocol over the next 8 weeks.

Three tests across 8 weeks also means 8 weeks of uncertainty in your loading prescriptions, which is fine for hypertrophy programs but suboptimal for powerlifting peaking. For peaking blocks, front-load the protocol: test 1 in week 1, test 2 in week 3, test 3 in week 5. Tighter timeline, slightly less daily-variance smoothing.

Why we ship this

We adopted the three-test protocol from Body Unbound’s free template (which we cite in our manifesto and across the site). It’s the most defensible way to estimate 1RM with a working trainer’s typical equipment (just a barbell and a session, no velocity meter, no force plate). The methodology is publicly available, well-documented, and embeddable in a spreadsheet workflow with zero extra technology. That’s the whole reason it fits Coach Sheet.