Reverse Pyramid Training: Does It Actually Work?

Quick answer: Reverse pyramid training puts your heaviest, lowest-rep set first, then cuts about 10 percent off the load on each following set as reps rise. It works, but the direct research finds it no better than ascending pyramids or straight sets. Its real advantages are time efficiency and a progression decision that comes down to one set.

Traditional pyramid sets ask you to do something slightly odd. You lift the lightest weight while you are freshest, then attempt your heaviest set last, after three sets of accumulated fatigue.

Reverse pyramid training flips that. Warm up, hit your hardest set immediately, then peel weight off and chase reps on the way down.

The internet is fairly convinced this is the superior method. The research is not. Both of those statements deserve more detail than they usually get.

What is reverse pyramid training?

After your warm-ups, the first working set is the heaviest and has the lowest rep target. Every set after that drops load and adds reps.

A typical bench press session looks like this:

Compare that to an ascending pyramid, which starts at 80kg for 10 and finishes at 100kg for 4, and to straight sets, which use one load for every set.

The method is much older than the internet suggests. It was published in 1951 as the Oxford technique, when Zinovieff reversed the then-standard DeLorme protocol specifically because patients were too fatigued to complete the heaviest set when it came last.

In lifting circles it was popularised decades later by Martin Berkhan of Leangains, and spread further through Andy Morgan at RippedBody.

The standard prescription

Berkhan's version is 2 to 3 working sets per exercise with roughly 10 percent knocked off between them. His load-to-rep mapping is worth memorising:

• Drop 5 percent and expect about the same reps.
• Drop 10 percent and expect 1 to 2 extra reps.
• Drop 15 percent and expect 3 to 4 extra reps.

Rest is long, at least 2 to 3 minutes, and longer after the top set. Berkhan suggests up to 5 to 10 minutes between heavy squat or deadlift sets.

Advocates differ on the exact ladder. RippedBody uses 10 to 15 percent drops with a 4 to 8 rep top set; Legion uses about 10 percent with a 4-6 / 6-8 / 8-10 structure. The popular 6/8/10 version is a middle reading, not a canonical one.

How does progression work on RPT?

Double progression on the top set, and nothing else.

Pick a rep bracket for your top set, say 4 to 6. Work within it until you hit the top of the range, then add weight and drop back to the bottom. Berkhan's rule is to add either 2.5 percent or one rep each session.

The back-off sets progress independently. When set two starts feeling easy, add weight to set two. It is not chained to the top set.

This is genuinely one of RPT's better features. Your entire progression decision each session collapses to one question: did the top set clear the top of the range? Our guide to double progression covers the general pattern.

Is reverse pyramid training better than straight sets?

No, reverse pyramid training is not better than straight sets, and this is the part almost every article leaves out.

Start with the only head-to-head trial of ascending against descending loading. A 2003 randomised study in the American Journal of Physical Medicine and Rehabilitation put 50 subjects through nine weeks of three sets of ten, three times a week. One group ran DeLorme (ascending, 50 then 75 then 100 percent of 10RM), the other ran Oxford (descending, the reverse).

The ascending group added 71.9kg to their 10RM. The descending group added 67.5kg. The difference was not statistically significant, and the authors concluded both improve strength with equivalent efficacy.

It was knee extensions in a rehab population over nine weeks, so it is not the last word. It is still the only direct test, and it is null.

The acute comparison is also null

A 2024 study in the International Journal of Strength and Conditioning had 18 well-trained men bench press under three matched conditions: constant load at 75 percent, an ascending pyramid, and a descending pyramid.

Total volume load came out essentially identical, at 2,285kg constant, 2,218kg ascending and 2,290kg descending. Once relative intensity was matched, rep performance did not differ either.

The researchers expected the descending condition to produce more reps and more volume. It did not.

Two findings did separate. Set-level RPE was lower in the descending condition, 7 against 7.5, so it felt easier. And peak barbell velocity was significantly lower during the descending condition, which makes RPT a poor fit if you are training for power.

Pyramids in general do not beat constant loading

A 2017 study in the European Journal of Applied Physiology equated volume across 12 weeks in trained men and found muscle cross-sectional area rose 7.6 percent with traditional sets and 7.5 percent with a pyramid scheme. Leg press one rep max rose 25.9 percent in both.

A 2023 review in the Journal of Bodywork and Movement Therapies pooled 15 studies and concluded pyramid training was not superior for acute responses, strength or hypertrophy, and that the choice between schemes could come down to periodisation, motivation or personal preference.

So why do people swear by it?

Because "not superior" is not the same as "not useful", and RPT has three real arguments in its favour.

You do your hardest work fresh. This is the founding logic, and it is the same reason Zinovieff reversed DeLorme's protocol in 1951. A 2021 meta-analysis in the European Journal of Sport Science covering 11 studies found strength gains were largest in exercises performed at the start of a session, with multi-joint strength favouring going first (effect size 0.32). Note carefully that this is about exercise order within a session, not set order within an exercise, so it supports the principle rather than testing RPT itself.

The varied rep ranges cost you nothing. Because sets are taken close to failure, the 4-6 / 6-8 / 8-10 ladder is not a problem for growth. A 2017 meta-analysis in the Journal of Strength and Conditioning Research found hypertrophy was similar across loads when sets went to failure, while heavier loads had the edge for one rep max strength. The heavy top set is where the strength benefit sits.

It is fast. Two or three working sets on one or two compounds, and you are done.

The honest case against

• The back-off sets are polluted by fatigue. Sets two and three happen under accumulated fatigue from the top set, so their load-to-rep relationship tells you very little. Only the top set is a clean signal.
• Volume is low. RippedBody lists insufficient volume as a genuine long-term limitation.
• It is not for beginners. Near-maximal top sets before technique is grooved is both an injury risk and a data quality problem.
• It costs peak velocity, per the 2024 study above.
• It does not extend past one or two lifts. Every credible source restricts RPT to the first compound or two, with straight sets for the rest.

Why is RPT so awkward to log?

Here is the structural problem: within a single exercise, every set has a different load, a different rep target, and its own progression state.

There is no 3x8 @ 100kg shorthand for an RPT bench press. There are three separate load and rep pairs, of which exactly one decides whether you progress.

That breaks most tracking setups in specific ways:

• One rep target per exercise is not enough. Set one targets 4 to 6 while set three targets 8 to 10.
• The top set has to be identifiable. Compare top set to top set across sessions. Comparing total tonnage gives the wrong answer, because a poor top set followed by generous back-offs can produce more total volume than a great session.
• Back-off sets progress on their own schedule, so each set needs its own history.
• Warm-ups must stay out of the working-set data, since RPT uses a long ramp. Our guide on whether to log warm-up sets covers that distinction.
• Rest differs by set, so one global timer per exercise is wrong.

A written entry that actually works looks like this:

Bench (RPT)
100kg x5   top set, RIR 1
90kg x7
80kg x9

Working out your back-off loads is easier if you know your numbers; plug your best set into the one rep max calculator and take the percentages from there.

A simple starting plan

1. Apply RPT to one or two compounds only. Your first pressing movement and your first pulling movement. Everything else stays straight sets.
2. Warm up properly, then go. Ramp to your top set with singles and doubles that do not accumulate fatigue.
3. Pick a bracket and a drop. A 4 to 6 rep top set with 10 percent drops is the standard starting point.
4. Rest longer than feels necessary, at least 3 minutes after the top set and up to 5 on squats and deadlifts.
5. Judge the session on the top set alone. Hit 6, add 2.5 percent next time and go back to 4. Nothing else changes your decision.

Keeping a per-set log without slowing down

If you write your sessions into a notes app, RPT is easy to record. Three lines, three weights, done. No app on earth logs it faster than typing it.

The friction arrives when you want to know whether the top set moved. That answer lives across twelve weeks of notes, and pulling it out means scrolling past every back-off set you ever did.

Gym Note Plus reads the plain text you already write and builds the history from it:

• Every set kept separately, so three different weights on one exercise stay three different weights.
• Per-exercise trends that let you compare this week's top set against every previous one.
• No forced templates, so a scheme with a different target on every set does not need to be squeezed into one.

Paste a few RPT sessions into the free notes to workout translator and see what it makes of them.

Reverse pyramid training is not a better stimulus. It is a simpler decision: one set to beat, one number to write down, and everything after it is accessory work.

Frequently Asked Questions

Is reverse pyramid training effective?

Reverse pyramid training is effective, but not more effective than the alternatives. The only direct trial comparing ascending and descending loading, published in 2003, found 10RM gains of 71.9kg and 67.5kg respectively with no statistically significant difference. Reverse pyramid training is a reasonable way to organise sets rather than a superior training stimulus.

Is reverse pyramid training better than straight sets?

No published trial shows reverse pyramid training beating straight sets for hypertrophy in trained lifters. A 2024 study matching volume across constant load, ascending and descending conditions found near-identical total volume load and no difference in repetition performance. Reverse pyramid training wins on time efficiency and simplicity of progression, not on results.

How much weight should you drop between reverse pyramid sets?

The standard drop is about 10 percent per set. Martin Berkhan's guidance maps drops to expected reps: 5 percent yields roughly the same reps, 10 percent adds 1 to 2 reps, and 15 percent adds 3 to 4 reps. Choose the drop that lands each set inside its intended rep bracket and keep it consistent between sessions.

Is reverse pyramid training good for beginners?

Reverse pyramid training is not recommended for beginners. It opens with a near-maximal set before technique is reliable, which raises injury risk and produces noisy data, since a novice's top set varies with sleep, nerves and gym conditions. Beginners progress faster on straight sets with a consistent load across all sets.

How long should you rest between reverse pyramid sets?

Rest at least 2 to 3 minutes between reverse pyramid sets, and longer after the top set. Berkhan recommends up to 5 to 10 minutes between heavy squat or deadlift sets. Because only the top set is near-maximal, back-off sets need enough recovery to hit their rep targets without becoming maximal efforts themselves.

Final Takeaways

• The direct evidence is null. Ascending and descending loading produced equivalent strength gains in the one trial that compared them head to head.
• The real benefits are practical, not physiological: less time, fewer sets, and one clear number to beat each session.
• Only the top set is a clean signal. Back-off sets are fatigue-affected, so do not read progress into them.
• It costs peak velocity, so pick something else if power is the goal.
• Use it on one or two compounds, then run straight sets for the rest of the session.
• If you are weighing set schemes, our guides to rep ranges and training to failure cover the neighbouring decisions.