How to Use Your 1RM in 5/3/1 (Without Screwing It Up)

5/3/1 doesn’t run off your true one-rep max — it runs off 90% of it, called your training max. Take your estimated 1RM, multiply by 0.9, and every working weight in Jim Wendler’s program is a percentage of that number. Get this one step right and the whole program works; get it wrong and you’ll stall in six weeks.

What is a training max and why does Wendler insist on 90%?

531
531

Your training max (TM) is a deliberately conservative number: 90% of your best estimated or tested 1RM. If your squat max is 300 lb, your training max is 270 lb, and every set you do for the next month is calculated from 270 — not 300.

Wendler’s reasoning is brutally practical. Most lifters overestimate what they can handle day in and day out. A true 1RM is a peak performance on a perfect day; training off it means grinding at your absolute ceiling every session, which is how you stall, get hurt, or burn out. The 90% training max builds in a buffer — you start a little light, progress a little longer, and the weights catch up to you instead of crushing you.

(Honest footnote: in his later books Wendler introduced some templates that start at 85% instead of 90%. The classic program — and the one 99% of people mean by “5/3/1” — is 90%. If you’re brand new to the program, 90% is the right start.)

I’ve watched lifters “fix” 5/3/1 by bumping their training max to 95% or 100% because 90% “felt too easy” in week one. Every single one stalled by week six. The program is designed to feel easy at the start. That’s the point.

How is a 5/3/1 cycle structured?

Each 4-week cycle hits every main lift (squat, bench, deadlift, overhead press) once per week. The rep schemes:

  • Week 1 — 3×5: 65% × 5, 75% × 5, 85% × 5+ (last set is AMRAP — as many reps as possible)
  • Week 2 — 3×3: 70% × 3, 80% × 3, 90% × 3+
  • Week 3 — 5/3/1: 75% × 5, 85% × 3, 95% × 1+
  • Week 4 — Deload: 40% × 5, 50% × 5, 60% × 5 (no AMRAP)

All percentages are of your training max, not your true max. The “+” sets are where the magic happens — pushing those top sets for extra reps is what drives progress and tells you when your training max needs to go up.

Worked example: 300 lb squat max

Estimated 1RM: 300 lb. Training max: 300 × 0.9 = 270 lb. Here’s the full cycle:

WeekSet 1Set 2Set 3
Week 1 (5/5/5+)175 × 5200 × 5230 × 5+
Week 2 (3/3/3+)190 × 3215 × 3245 × 3+
Week 3 (5/3/1+)200 × 5230 × 3255 × 1+
Week 4 (deload)110 × 5135 × 5160 × 5

(Exact math: 270 × 65% = 175.5, 75% = 202.5, 85% = 229.5, 70% = 189, 80% = 216, 90% = 243, 95% = 256.5, 40% = 108, 50% = 135, 60% = 162 — rounded to the nearest loadable 5 lb.)

Notice week 3’s top set: 255 × 1+. That’s 85% of your true max — a weight you should own. If you hit it for 3–4 reps, your training max is honest. If you barely grind out 1, your training max was set too high.

When do you recalculate your training max?

After every 4-week cycle, add weight to the training max — not the true max:

  • Upper body lifts (bench, overhead press): +5 lb per cycle
  • Lower body lifts (squat, deadlift): +10 lb per cycle

That’s it. No retesting, no recalculating from scratch. You just add the increment and run the next cycle. This slow, boring progression is the entire secret of 5/3/1 — it works for years because it never asks you to make a jump you can’t sustain.

Every 3–4 cycles, sanity-check: if your “+” sets are consistently hitting big rep numbers (8+ on week 3’s top set), your training max is too low and you can bump it a little extra. If you’re missing reps, it was too high — reset it back 10% and rebuild. Wendler calls this “two steps forward, one step back,” and it’s better than stalling for three months out of pride.

Common mistakes that kill 5/3/1

Mistake 1: Using your true max instead of 90%. The single most common way people break this program. Your ego says 300; the program says 270. Listen to the program.

Mistake 2: Skipping the deload. Week 4 feels like a waste of time until you’ve skipped three of them and your elbows are screaming. The deload is where the fatigue you’ve accumulated for three weeks dissipates. Don’t be clever; deload.

Mistake 3: Not pushing the “+” sets. Those AMRAP sets aren’t optional decoration — they’re the autoregulation built into the program. Leaving 3 reps in the tank on every “+” set is leaving progress on the table. Push them (with good form), and let the reps tell you when to adjust.

Mistake 4: Adding extra max-effort work. 5/3/1 is a submaximal program by design. Bolting heavy singles onto every session turns it into a different program — one that doesn’t work. Trust the percentages.

Frequently Asked Questions

Can I use an estimated 1RM (not a tested max) for 5/3/1?

Absolutely — that’s the recommended approach. Plug a recent 3–5 rep set into a 1RM calculator, take 90% of the estimate, and start there. It’s safer than testing a true max just to start a program, and the 90% buffer absorbs any estimation error.

What if 90% feels too easy?

It’s supposed to — at first. The “+” sets let you push when you feel strong, and the monthly 5/10 lb increases compound fast. A training max that feels easy in week 1 feels exactly right by week 9. Starting too heavy is the #1 reason people stall on this program.

Do I need to retest my 1RM between cycles?

No. Just add 5 lb (upper) or 10 lb (lower) to the training max each cycle. Retest only if something major changes — long layoff, big bodyweight change, or switching the lift variation (e.g., low-bar to high-bar squat).

Can I run 5/3/1 for overhead press?

Yes — it’s one of the four core lifts in the original program, using the same percentages and progression (+5 lb per cycle like bench). Use a strict press training max, not a push-press number.

What RIR should the “+” sets be?

Push them to 1–2 RIR (reps in reserve) — hard but not to absolute failure, especially on squat and deadlift where form breakdown under fatigue is the real risk. Leaving the last ugly rep in the tank is smart training, not weakness.

Sources

Similar Posts

Leave a Reply

Your email address will not be published. Required fields are marked *