Velocity loss & fatigue
The cleanest single number for "how much fatigue is in this set", and the autoregulation metric that holds up best with real athletes.
Velocity loss is the percentage drop in mean concentric velocity from the fastest rep of a set to the current rep. Cross a chosen cutoff and the set ends — regardless of how many reps the program said to do.
It works because it’s the closest thing we have to a direct read of intra-set fatigue. RPE tries to estimate it after the fact. Reps-in-reserve tries to predict it before. Velocity loss measures it as it happens.
Why it works better than RPE
- Objective. No subjective rating, no scale calibration drift between athletes.
- Real-time. You see the cutoff coming on rep 5 of the set, not in the journal review afterwards.
- Comparable across sessions. Two sets to the same v-loss target are matched for fatigue regardless of load, reps, or how the athlete felt that day.
Common cutoffs by goal
| Goal | Typical cutoff | Why |
|---|---|---|
| Power / speed | 5–10% | Stop while reps are still fast and fresh |
| Strength | 10–20% | Productive volume without grinding |
| Hypertrophy | 20–30% | Mechanical tension over fresh velocity |
| Capacity / GPP | 30%+ | Buffer-burning sets, accept some grinder reps |
These are starting points, not laws. Calibrate to your athlete. A powerlifter on a peaking week probably wants 10% even on hypertrophy days; a beginner doing 3×8 squats might safely take 30% v-loss every session.
What it changes about training
Two consequences fall out of using v-loss as the cutoff:
- Volume self-regulates. A fresh athlete gets more reps to the same cutoff than a fatigued one. The training stress equalises across days that would otherwise look very different in the journal.
- Overreaching shows up early. When yesterday’s cutoff came after 6 reps and today’s came after 3 at the same load, you have hard evidence that the athlete needs to back off — before performance drops or injury risk climbs.
Looking for the protocol? Picking a cutoff for a specific block, calibrating it across athletes, and integrating it into a program template is its own write-up.