Pyramid training
Pyramid training changes load and repetitions across successive sets of the same exercise. In an ascending pyramid, the weight generally rises while the repetition target falls. In a reverse pyramid, the heaviest working set comes first after an appropriate warm-up, followed by lighter sets with more repetitions. A full or double pyramid goes up and then back down.
For example, an ascending sequence might move from a lighter set of twelve repetitions to moderate sets of eight and six. Those numbers illustrate the shape; they are not a standard prescription. The actual training load, rep range, number of sets, rest periods, and effort determine the demand. If the early lighter sets are taken close to failure, they may fatigue the lifter before the heaviest set rather than merely prepare for it.
This is why warm-up sets and pyramid working sets should be distinguished. A warm-up set prepares the movement and is deliberately kept below a demanding effort. An ascending pyramid counts its early sets as working sets when they are intended to provide training stimulus. In a reverse pyramid, lighter sets performed after the top set overlap with the idea of back-off sets, although not every back-off arrangement follows a pyramid.
No pyramid shape has a special adaptation independent of the work performed. A 2026 ACSM overview of resistance-training evidence found that set structure did not consistently change training outcomes across the available reviews. Pyramid training can still be a practical way to expose a lifter to several loads or repetition ranges within one exercise, but a straight-set arrangement can work as well when volume, effort, and progression are appropriate.
Choose the direction according to the priority. An ascending pyramid provides a gradual transition toward heavier loads but can reduce performance at the top if earlier sets are too demanding. A reverse pyramid places the strength-priority set before accumulated work, yet requires a separate warm-up and sensible load reductions. Record each set individually, because averaging the pyramid can hide where performance changed.