Velocity-Based Training: velocity-loss threshold ≤25% superior to >25%
Meta-analysis
Key Findings
10 studies, 308 resistance-trained young men; VL as a categorical (≤25% vs >25%) and continuous variable
Low-Mod VL (≤25%) is superior to Mod-High VL (>25%) for gains in strength and athletic measures
The Low-Mod group trained with less VL (16.1 vs 39.8%) and LESS volume (212 vs 384 repetitions) — yet with a better result
Between-group: Low-Mod VL more effective for performance against low loads (ES=0.31, p=0.01) and moderate/high loads (ES=0.21, p=0.07)
Within-group ES: strength Low-Mod 0.79-2.39 vs Mod-High 0.59-1.91; athleticism Low-Mod 0.35-0.59 vs Mod-High 0.05-0.36
Adaptations DECLINE as the VL threshold rises — especially for light loads (R²=0.73, p=0.01), local endurance (R²=0.93, p=0.04) and sprint (R²=0.61, p=0.06)
Practical Application
Meta-analysis of 10 studies in 308 resistance-trained young men: stopping sets at ≤25% velocity loss produced larger strength (ES 0.79-2.39 vs 0.59-1.91) and athletic (0.35-0.59 vs 0.05-0.36) gains than >25% velocity loss, with 212 vs 384 repetitions of training volume; adaptations fell as the velocity-loss threshold rose. End the working set well before failure — better and more efficient.
Not from this paper; the base's own practice: without an encoder, gauge the ~25% velocity-loss point by the loss of speed/movement crispness, roughly 2-4 reps in reserve
For hybrid/HYROX: less intra-set fatigue = less interference with endurance (see REF-512)
Inference, not tested: for masters, the same or better stimulus with less volume plausibly means faster recovery
Limitations
Trained young men only (n=308) — direct transfer to masters/women requires caution
"Athletic" outcomes are heterogeneous; the moderate/high-load between-group effect (p=0.07) and the sprint relationship (p=0.06) were not significant
The optimal VL threshold may depend on the goal (strength vs hypertrophy vs endurance)
How this entry was checked and changed
Content correction 2026-09-27
Lead bullet "end at ~20-25% VL" -> the meta-analysis's own numbers (≤25% vs >25%, ES ranges, 212 vs 384 reps); "2-4 RIR" labelled as the base's own practice.
Added the between-group effect sizes and the p-values the abstract reports (including the non-significant ones); Tier "mixed" -> Tier 2-3, Sport "strength, sprint" -> "strength".