Training Max vs 1RM: when to use which

A Training Max (TM) is roughly 90% of your true 1RM, a programming convention popularized by Jim Wendler's 5/3/1. The point of TM isn't precision — it's sustainability. A true 1RM drifts week to week with sleep, stress, and groove; TM doesn't, because it's anchored 10% below your best day. The clean per-context split: TM for week-to-week programming. 1RM for peaking attempts and meet-attempt selection.

Why 90% specifically. The number stuck because it's high enough to drive strength gains under standard intensity ramps and low enough to leave headroom for a bad Tuesday without missing reps. If you want to derive a TM, start from an estimated 1RM via the calculator and run percentages through the percentage-of-1RM calculator. Both tools live in the calculators hub.

What Training Max actually means

A Training Max is the number you program off, not the number you can actually lift on your best day. The standard 5/3/1 setup takes 90% of a recent estimated or tested 1RM and locks that as TM for the training cycle. Working sets — top sets, AMRAP sets, supplemental volume — get computed as percentages of TM, not of the true 1RM.

The other 10% is structural slack. Sleep was bad. The warm-up groove was off. A coworker dropped a stressful problem in your lap an hour before training. Any of that can knock 5% off a heavy single without telling you in advance. TM is the strength-training equivalent of leaving slack in your build pipeline — the headroom isn't pessimism, it's how the program survives a bad Tuesday. Lift with no slack and the first off-day stalls the cycle.

The case for TM: sustainability over multi-month blocks

Wendler's argument for TM is straightforward: missed reps cascade. Miss a top set on Monday and the next session's opener feels mentally heavier, the AMRAP set gets one rep shorter, the supplemental volume drops a set. Over a 12-week block that pattern compounds. A TM set 10% below true 1RM produces top sets that you hit, week after week, even on days where you wouldn't have hit the 1RM-derived equivalent. The programmed PRs accumulate; the missed-rep cascades don't.

The honest take: programming off true 1RM is what beginners and program-skeptics do, and it works for a few weeks. It stops working the first week the lifter is under-recovered, and from there the program is in salvage mode rather than building. TM is the convention that turns “the program works when conditions are perfect” into “the program works when conditions are average.”

The case for 1RM: peaking and meet-attempt selection

The actual 1RM matters in two contexts. First, peaking blocks: when the goal is to express maximal strength on a known date, the calculator's peaking band runs 85–97.5% of 1RM (sets of 1–3, long rest, low total volume). The percentages engine caps programmed peaking at 97.5%— the last 2.5% is reserved for an actual test or meet attempt, not a programmed working set. Programming the peak off TM instead of 1RM under-shoots the meet-day intensity by ~10% and leaves the lifter un-rehearsed at the loads they'll actually attempt.

Second, meet-attempt selection. A powerlifting opener is conventionally ~91–93% of true 1RM, the second attempt ~95–97%, and the third the actual PR attempt at 100%+ of historical best. Those numbers are all computed off the lifter's known 1RM, not off TM — picking attempts off TM produces a wildly under-loaded opener and surrenders the totaled weight a meet is scored on.

The directional take: don't run a 1RM test off the top of a program block; run it as the OUTPUT of a block. Test fresh, after a deload, with the program's strength gains baked in. Then the new 1RM resets TM for the next cycle, and the engine has a clean number to project the next 90% off.

TM doesn't update every week

The 5/3/1 convention is to reset TM after a training cycle — typically 3–4 weeks — not after every PR. Either by adding a fixed increment (5 lb on upper-body lifts, 10 lb on lower-body, in Wendler's stock setup) or by re-deriving TM from a fresh estimated 1RM via the end-of-cycle AMRAP set. Either path produces a stable, slowly-rising TM that programmed top sets land on.

What this rule prevents: chasing a single good day's number into next week's programming. A lifter who hits a 5RM PR on a great Wednesday and immediately bumps TM up the next Monday tends to miss the new top sets, get demoralized, and stall. The PR was a point-estimate; TM is a moving average. Update on cycle boundaries, not on rep-day spikes — that's the whole structural argument against the “every-PR-counts” instinct.

When 90% is the wrong percentage

90% is a default, not a constant. Two cases bend it. Advanced lifters running long blocks — especially powerlifters running 8–12 week peaking cycles — often drop TM to 85% to absorb more grinding work without burning out by week 6. Wendler himself has discussed an 85% TM in subsequent editions of the program for exactly this reason. Novice lifters with low week-to-week variability and high recovery margin can sometimes hold TM at 92–95% without missing reps, but the payoff is small and the downside (a missed cycle) is real.

The diagnostic: if you're missing reps in week 3 of every cycle, your TM is too high. Drop it 5% and rerun. That's a StrengthMath rule of thumb, not a Wendler-published threshold — but it's the cleanest signal a self-coaching lifter gets that the programming reference is set wrong. Don't patch a missed week with a longer warm-up or extra caffeine; patch it by dropping the number the program is built on.

The full chain: 1RM input → TM → percentage table

The end-to-end programming workflow is three steps, and every step is a number the calculators on this site produce:

1. Estimate or test 1RM. Run a 3–5 rep set through the 1RM calculator; for the per-formula breakdown, see the best-1RM-formula guide.
2. Take 90% as TM. Multiply 1RM × 0.90; round to a sensible plate-loadable number.
3. Run percentages off TM. Plug TM into the percentage-of-1RM calculator as if it were your 1RM. Working sets, supplemental volume, and AMRAP top sets all derive from this table.

The worked example. A lifter with an estimated 1RM of 100 kg:

• TM = 100 × 0.90 = 90 kg.
• 5/3/1 working set at 85% TM = 90 × 0.85 = 76.5 kg. Compare against 85% of true 1RM = 85 kg. That 8.5 kg gap is the headroom — the same prescription, anchored to a sustainable reference.
• After a 3-week cycle, the end-of-cycle AMRAP set produces a clean 105 kg estimated 1RM. TM resets to 105 × 0.90 = 94.5 kg for the next block, and the percentage table shifts accordingly.

The directional take: program off TM. Test off 1RM. Update TM at cycle boundaries. The whole reason 5/3/1 outlasted most of its contemporary programs is that this three-step structure absorbs the week-to-week noise that wrecks 1RM-based programming — not because 90% is a magic number.

Common questions

What is a Training Max?

A Training Max (TM) is a programming reference set at roughly 90% of your true 1RM, popularized by Jim Wendler's 5/3/1 program. The other 10% is headroom — week-to-week variability from sleep, stress, technique, and groove. You program working sets off TM (not 1RM) so a single rough day doesn't blow up the program.

Should I program off my TM or my 1RM?

TM for sustained week-to-week programming. 1RM for peaking attempts, max-test sessions, and meet-attempt selection (1st, 2nd, 3rd attempts at a powerlifting meet are all chosen relative to a known true 1RM). The percentages calculator's peaking band caps at 97.5% — that's the ceiling for programmed peaking; an actual 1RM test goes higher and should be the OUTPUT of a block, not the input.

When does the Training Max update?

After a training cycle, not after every PR. Wendler's convention is to reset TM after a 3–4 week cycle by adding a small fixed increment (often 5 lb upper-body, 10 lb lower-body) — or by re-deriving from a fresh estimated 1RM. Updating TM every week chases a single good day's number into next week's programming and produces an inflate-then-stall pattern.

What if 90% is too high for me?

Drop to 85%. If you're missing reps in week 3 of every cycle, your TM is too high — that's StrengthMath's rule of thumb, not a published threshold. Advanced lifters running long blocks often drop TM to 85% to absorb more grinding work; novice lifters with low week-to-week variability can sometimes run higher. The number is a knob, not a constant.

Where to next

Once TM is set, the next decision is what percentage of TM to actually train at — strength, hypertrophy, peaking, and deload all map to different bands. The percentage-of-1RM-for-strength guide covers the NSCA traditional 80–95% band and the ACSM 2026 ≥80% threshold; the percentage table on the calculator works equally well driven by TM as a stand-in for 1RM. For the broader framing of how the engine pins reliability bands and why the methodology page draws a hard line between sourced and StrengthMath-methodology numbers, the methodology page is the canonical reference.