5/3/1 or 8×3? The Program Is Decided by Progression, Not Sets

2026-08-19

Two popular ways to structure a session look like they differ in sets and reps. The real difference is what decides when the load goes up – the calendar or your performance.

When programs get compared, sets and reps are what catch the eye first. 5/3/1 does three main working sets; a density scheme like 8 × 3 does eight. The numbers say less than they appear to. The more important difference is what triggers a load increase – and that difference determines which model suits whom.

Three reps is not a load

Start with what misleads in both. A rep count tells you nothing about the load until you know how close to failure the set is taken and how many sets there are.

According to the RPE chart, three reps taken to failure is about 92% of max. Leaving two reps in reserve, at RPE 8, still puts you near 87%. As a single set, three reps is a heavy proposition.

Eight sets of three is doing something else entirely. There the load is set by the eighth set, not the first. The opening set feels easy – and it has to, or the sixth won't happen. The same load travels from light to near-failure across the session, which is why talking about a "three-rep load" is meaningless unless you also state how many sets and how much rest.

5/3/1 in brief

Jim Wendler's 5/3/1 is built on a training max: 90% of your true max, with every percentage calculated from that number. The four-week cycle looks like this:

The plus sign means the last set is taken for as many reps as you can get. That set is the program's measuring instrument: progress shows up as reps, not as load. After each cycle the training max goes up by 2.5 kg on upper-body lifts and 5 kg on lower-body lifts, regardless of how the sessions felt. The 5/3/1 calculator works out every load in the cycle at once.

The training max has one consequence many lifters never notice. The week-three top set is 95% of the training max – but since the training max is 90% of the real one, that is only about 85.5% of your actual maximum. The program never goes near a true max. The AMRAP set does that job without a max attempt.

Total main-set volume in week three is nine reps, plus whatever the AMRAP adds.

8 × 3 with one-minute rests

The density model works differently. Reps, rest and target sets are fixed: three reps, one minute of rest after every set, eight sets as the goal. You keep going as long as three reps keep going up. When the third rep fails, the exercise is done for the day.

The load stays the same the following week. Only once all eight sets are completed does the weight go up – by a smaller step on upper-body lifts than on lower-body ones.

Structurally this is double progression where the advancing variable is the number of sets, with load following behind. Five sets in the first week, six in the second, eight in the third, and the fourth week has earned more weight. A bad day breaks nothing: it shows up as fewer sets rather than as a failed rep.

Prilepin's chart reveals the load

Here Prilepin's chart exposes something the program itself never states. Eight sets of three is 24 reps. In the 70–80% zone the chart's recommended total is 12–24, so 24 sits exactly at the ceiling. In the 80–90% zone the ceiling is 20 – and 24 clearly exceeds it.

The program therefore chooses its own load. Try 8 × 3 at 85% and the eighth set will not happen on one-minute rests; the structure pushes you back into the 70–80% band. That is a self-correcting feature rather than a flaw. It also means there is no point guessing high on the starting weight: better to begin too light and progress than to stall at six sets for a month.

For comparison, week three of 5/3/1 spans 67.5% to 85.5% of a true max, and nine main-set reps fit comfortably inside every zone. The two programs are not competing for the same job: one accumulates volume at moderate loads, the other touches heavy loads briefly.

Two ways to decide on progression

This is the whole difference in one paragraph. 5/3/1 raises the load on a schedule: the cycle ends, the weight goes up. Performance is measured by the AMRAP set, and if that starts fading, the training max is reset. Feedback arrives on a delay, in cycle-sized chunks.

The density model raises the load on performance: when the work is done, the weight goes up. Feedback arrives within the session itself, and the program adapts automatically to a bad night's sleep or a stressful week.

Both logics work, but they fail differently. A schedule-driven model can carry the load past your actual condition, at which point the AMRAP sets shrink and a reset is due. A performance-driven model can grind on the same weight for a long time, because nothing forces progress – and since it has no built-in deload week, fatigue can accumulate unnoticed.

Which suits whom

The schedule-driven model fits when training is regular and you want to know now what you'll be lifting next month. It is also easier to plan around a fixed date.

The density model fits when life is irregular, when recovery varies, or when you want quality volume without heavy individual sets. The one-minute rest also makes it a conditioning stimulus – good or bad depending on what you're after.

Nothing stops you from using both: a schedule-driven model on the main lift and a density model on a second lift in the same session.

The most common mistake is the same in both

The starting load is set too high. In 5/3/1 that happens when the training max is set to a true max rather than 90% of it – which makes the week-three top set 95% instead of 85.5%, and the AMRAP sets die within a couple of cycles. In the density model it happens when the load is chosen by how three reps feel in the first set.

The remedy is the same in both cases: work out the starting load with the one-rep max calculator, take the percentages from the percentage chart, and err on the light side. These programs are designed to progress over weeks, not to prove anything on day one.

Summary

Sets and reps are the most visible part of a program and rarely the most important. 5/3/1 and 8 × 3 differ above all in who decides when the load goes up: the calendar, or yesterday's session. Pick the logic that survives your own week – the loads can be calculated for either with the same tools.