Variable resistance topped 2026 squat rankings for 1RM (one-rep max), while APRE (load adjustment) topped jump height; four weeks of heavy lifting left 20 m sprint times untouched in players under 20 whose strength, jump height and strength deficit moved.

An autoregulation review tracked one-repetition maximum (1RM), the heaviest load an athlete can lift once, and nothing else. A 2026 network meta-analysis—a comparison of several methods using direct and indirect evidence—ranked methods on squat 1RM and countermovement jump (CMJ), a jump after a quick downward dip. A four-week football block measured strength, jump, 20 m sprint and strength deficit, the gap between force at a maximum lift and force at a lighter load. The 1RM findings do not establish effects on CMJ, sprint time or match running, and the sprint finding does not tell you anything about the squat or jump.

Autoregulation handles readiness

Autoregulation adjusts load, repetitions or sets to how an athlete performs or feels on the day. Larsen, Kristiansen and van den Tillaar reviewed 14 studies covering 30 training groups and 356 participants, and every autoregulation protocol they included raised one-repetition maximum. Those protocols spanned repetitions-in-reserve ratings (how many more repetitions an athlete thinks they can complete), flexible daily undulation (changing the work from day to day), velocity targets (prescribed lifting speeds) and velocity-loss methods (stopping after speed drops by a set amount). ¹

Every study in that review measured 1RM. So the review backs autoregulation as a route to maximal strength, and it leaves sprinting and football performance untested.

Repetitions-in-reserve ratings lose accuracy in untrained lifters and in high-repetition sets. Bartolomei and colleagues tested recovery-based autoregulation in 2024 over 10 weeks with 24 trained men. The autoregulated group did more total volume. Neither group came out ahead on squat or bench 1RM, power, lean mass or measured muscle thickness. ²

VRT led for squat 1RM and APRE led for jump

VRT ranked first for squat 1RM, but APRE ranked first for the countermovement jump. Bao et al. pooled 27 randomized trials with 694 healthy, active participants in a 2026 network meta-analysis, comparing APRE (adjusting load from performance), velocity-based training (VBT, using lifting speed to set or adjust load), RPE (rating of perceived exertion, how hard the set feels) and RIR (repetitions in reserve, the reps left before failure) autoregulation, variable resistance training (VRT, changing resistance through the lift) and fixed percentage training (PBT, prescribing a set percentage of 1RM). They measured two outcomes, back-squat 1RM and countermovement jump, and reported SUCRA scores, percentages summarizing how highly each method ranked within each outcome. The table reports SMD (standardized mean difference, a way to compare effects on different scales) and 95% CI (confidence interval, the range showing the estimate’s uncertainty). ³

Variable resistance leads the network for squat 1RM. APRE leads it for the jump. The analysis pooled direct and indirect comparisons, ran from age 13 to 35, and had few head-to-head trials between methods to work from. Neither of the two outcomes it measured was a sprint.

Strength and jump improved while sprint stayed flat

The four-week block improved relative strength, peak force and CMJ but did not change 20 m sprint time. Loturco et al. followed 35 male under-20 players from two professional clubs through a four-week in-season block. One squad of 18 did 12 sessions of half-squats and jump squats at fixed loads from 30% to 80% of 1RM. The other 17 carried on with their normal training and matches. ⁴

Relative strength and relative peak force express those measures relative to body mass. The group-by-time effect size describes the difference in change between the two groups. A p value estimates how surprising that difference would be if there were no real effect; smaller values indicate stronger evidence against that explanation.

In this study, the lighter load was 40% of 1RM. The gap widened, which means the players were not necessarily putting a greater share of their peak force into the lighter lift. Nobody timed their match running, so the finding says nothing about whether they slowed down on the pitch.

The study ran one programme, compared no methods, and the two teams did not play the same amount of football. It cannot crown an optimal method or pin down a cause.

Squat, jump, sprint and strength deficit

VRT, APRE and VBT all outperformed fixed percentages for squat 1RM and CMJ in the 2026 network meta-analysis. For sprint and lighter-load force, the football study found no resistance method that produced a gain. In the table, SDef is shorthand for strength deficit, the heavy-to-light force gap.

Autoregulation improved 1RM across every protocol in the review, but Bartolomei’s recovery-based version added volume without adding strength, and no study here timed an autoregulated group over a sprint.

The evidence is a maximal-strength review, one trial in trained men, a 2026 network meta-analysis and four weeks with 35 male under-20 players. None of it establishes an optimal programme for adult professionals, for women, for particular positions, or for match performance.