What Velocity Loss % Should You Stop a Set At? Strength vs Hypertrophy Cutoffs

Stop the set when bar speed has dropped a set percentage from your fastest rep: about 20% if you're chasing strength and power, and 25–40% if you're chasing size, with 40% and beyond marking true failure and under 10% reserved for explosive, technique-focused work. These aren't arbitrary numbers. They come from velocity-loss research that tracked how much intra-set slowdown different training goals actually need.

The catch is the instrument. Those percentages were measured with a barbell-mounted transducer that costs more than most gym memberships, and almost nobody reading this owns one. The thresholds still hold, but how you apply them changes when your sensor is a wrist, a phone camera, or just your own eyes. The research first, then how to use it on hardware you already own.

What does velocity loss actually mean?

Velocity loss is the percentage drop in your rep speed from the fastest rep of a set to the rep in front of you. Your first clean rep is the reference; every rep after it gets a little slower as the muscle fatigues, and the size of that slowdown is a direct, measurable readout of how deep into the set you are.

The logic is simple and well-established: as you accumulate fatigue, the muscle produces less force, so the bar moves slower against the same load. A barbell that's still flying at rep 8 tells you you're nowhere near failure. A barbell crawling at half its starting speed tells you you're done. That's why velocity loss is one of the few failure cues you can put an objective number on instead of a feeling. The physiology behind the slowdown is its own topic, and I cover it in why reps slow down at the end of a set.

A worked example makes it concrete. Say your first rep on a squat moves at 0.80 m/s. A 20% loss means the bar has slowed to 0.64 m/s; a 30% loss is 0.56 m/s; a 40% loss is 0.48 m/s. The cutoff is always relative to that first fast rep, never an absolute speed, which is why the same rule travels across loads and days: 0.80 m/s might be rep 1 at one load on a fresh Monday and rep 1 at a lighter load on a tired Friday, and a 20% cutoff lands you at the same effort either way.

What are the research velocity loss thresholds?

The literature clusters into 4 practical bands, and each maps to a different training outcome.

The systematic review by Jukic and colleagues anchors the edges of this table. In their analysis, terminating a squat set at a 20% velocity loss typically left you having completed roughly 50% of the reps you could have done, while a 40–50% loss meant you'd taken the set to, or very near, muscular failure. GymAware's velocity-loss reference frames it the same way: about 20% for strength, 30% or more for hypertrophy, under 10% for explosive work, and 40%-plus as the marker of extreme fatigue, where effort deteriorates and bar speed suddenly drops.

For strength specifically, lower is friendlier. The Jukic review concluded that the amount of velocity loss experienced during training did not reliably change strength or muscular-endurance gains, but higher velocity loss negatively affected jump, sprint, and bar speed against submaximal loads. So keeping intra-set fatigue low to moderate is the more efficient strategy and clearly superior for athletic qualities. You don't need to grind to grow stronger; you need quality reps and enough of them.

Treat the bands as regions rather than lines. The studies compared conditions like 20% against 40%, and the practical difference between neighbouring cutoffs is small; the difference between the bottom of the table and the top is not.

Why is velocity loss better than a fixed rep count?

Because your "10 reps" today is not the same set as your "10 reps" yesterday. Fatigue varies daily, and a fixed rep count ignores that, while velocity loss adjusts to whatever shape you're in. A fixed scheme says "do 3×10 no matter what." A velocity-loss rule says "stop each set when the bar has slowed 20%," which lands you at the same relative effort whether you slept well, ate enough, and showed up fresh, or you're dragging after a bad week.

This is the core appeal of autoregulation. On a strong day, a 20% cutoff might let you grind out 12 reps before the bar slows that much; on a beaten-up day it might cut you off at 6. Both sets carry the same physiological dose. A static rep target would have you under-stimulate on the good day and over-fatigue on the bad one, which is how junk volume creeps in. I dug into that trap in junk volume: more reps past the point of useful slowdown buy fatigue, not growth.

How close to failure do these thresholds put you?

A 20% loss leaves several reps in the tank; a 25–40% loss puts you within a couple of reps; and 40%-plus is functionally failure. That's the bridge between velocity loss and the more familiar reps-in-reserve language most lifters already use, which I break down in reps in reserve explained.

This matters because hitting failure isn't the holy grail people think it is. The proximity-to-failure meta-analysis by Refalo and colleagues found a non-linear, plateauing relationship: muscle growth jumped a lot moving from low velocity loss to moderate (effect sizes of about 0.20 to 0.39), but barely improved moving from moderate to high (0.39 to 0.42), and there was no statistically significant difference between high (over 25%) and moderate (20–25%) velocity-loss conditions for hypertrophy. Training all the way to momentary failure showed no reliable advantage over stopping short. The takeaway: get close enough and accumulate volume; chasing the last sliver of velocity loss costs you a lot of fatigue for almost no extra muscle. If you want the full argument, see how close to failure should you train.

The hardware tax and the wrist alternative

To measure velocity loss the textbook way, you bolt a linear position transducer (LPT) to the bar, a tethered device that reads displacement directly and is treated as the gold standard. The problem is obvious the moment you price one: a quality LPT or barbell tracker runs from a few hundred dollars to over $1,000, lives in the bar path, and is built for a coached weight room, not a crowded commercial gym where you're supersetting cables and dumbbells.

There's a cheaper sensor that works for a lot of lifting: an inertial measurement unit (IMU), the same accelerometer-plus-gyroscope package inside your phone and watch. A systematic review of IMUs for barbell velocity found that several of the devices tested can be considered valid and reliable against linear transducers, with strong agreement: correlations as high as 0.90 to 0.97 in good setups. The honest limitation is one of physics, not of that review: accelerometers measure acceleration, so estimating distance means integrating twice, and any small error accumulates into drift that grows quadratically with time. That's why a sensor that isn't bolted to the bar gives you a directional proxy rather than a lab-grade barbell readout.

Can you use these thresholds from a wrist?

You can use the shape of the signal; you can't copy the number. The thresholds in the table were measured on the bar. A wrist sits at the end of a different lever, travels a different arc through the same rep, and on some exercises (a leg press, a leg extension) is not moving with the load at all. So the percentage a wrist reports is its own number, not the study's, and it will differ by exercise and setup.

That doesn't make wrist velocity useless; it makes it something you read relative to yourself rather than against a published cutoff. A validation of barbell velocity measured with an Apple Watch found the wrist-worn placement performed well, though slightly less precisely than mounting the same watch on the bar, and the authors called the results "very promising" for putting velocity-based training into mainstream wearables. So you can't read the literature's exact "20% for strength" off your watch, but you can trust what the signal is telling you: that your reps are slowing, by how much relative to your own earlier reps, and when that slowdown suddenly steepens.

Velocity is also not a universal cutoff on its own: in a study of nearly 3,000 measurements in trained lifters, bar velocity explained only about 30% of the variance in perceived reps-in-reserve, and the relationship shifted by exercise, load, and set number. Velocity is a powerful second opinion. It is not an oracle, and any honest tool should tell you that.

3 ways to read velocity loss without a barbell tracker

1. Time the concentric on video. Velocity loss is a drop in mean velocity, and over the same range of motion mean velocity is just distance divided by time. Film the set on your phone, and compare how long the lifting phase takes on rep 1 against the last rep. If rep 1 takes 0.8 s and the last rep takes 1.2 s, the last rep moved at about two-thirds the speed: roughly 33% velocity loss. This assumes the range of motion didn't shrink, so watch the depth too.
2. Wear a wrist sensor and read it against yourself. Ignore the textbook percentage. Take a few sets to honest failure on a movement that's safe to fail, note what the slowdown reads at the end, and use that personal ceiling as your reference from then on.
3. Use your eyes. Rep speed doesn't fall evenly. It holds, then dips, then falls off a cliff. The rep that is suddenly much slower than the one before it is the visible version of the 25–40% band: you're within a couple of reps of the bar actually stopping.

A practical stop-rule you can run this week

You don't need a lab to apply this. Here's a concrete protocol that works with a barbell tracker, a budget IMU, a phone camera, or just your own attention:

1. Pick your goal and your band. Strength and power: aim to stop around the point where bar speed has clearly dropped but reps are still crisp (the ~20% zone). Hypertrophy: let the set run deeper, into obvious grinding (the 25–40% zone), but don't chase the last ugly rep.
2. Set a reference on rep one or two. Your fastest clean rep is your baseline. Everything is measured relative to that, not to some absolute number from a study.
3. Watch for the inflection, not a fixed count. The moment the bar speed visibly falls off a cliff, the rep that's suddenly much slower than the one before, that's your signal you're entering the failure zone. For strength, that's your cue to rack it.
4. If you're using a wrist sensor, learn your own numbers. The wrist isn't on the bar, so its percentage isn't the study's. Note the slowdown on a few sets you take to honest failure; that personal ceiling becomes your reference, and every later set is read against it.
5. Cross-check against feel. On a set where you stopped at your velocity cutoff, ask honestly how many reps you had left. Trained lifters who think they're at failure usually have a couple still in the tank, and beginners often have far more, which is exactly the gap velocity is there to close. Pairing an objective signal with your own estimate tightens that judgment fast; if you're not sure where your sets actually land, start with how to know if you're training hard enough.
6. Use it where it's strongest. Velocity loss is most reliable on multi-rep sets (6–15 reps). On heavy singles and doubles there aren't enough reps for a clean slowdown signal; use RPE and feel there instead.

Doing this on the watch you already wear

Riven is the Apple Watch app that scores muscle failure. It reads your wrist motion and stays quiet during the set; about three seconds after you rack the weight you get the exercise, the rep count and a 0-100 score. Reps slow down as a muscle nears failure, and that is what the score is built on. No barbell clip, no camera, no extra hardware. It shows you which set was the real one.

Everything above about the wrist still applies. Velocity explains only part of perceived effort, so treat any wrist reading, app or not, as a second opinion that sharpens your judgment rather than replacing it. But "an objective second opinion that beats guessing" is a high bar, because guessing is what almost everyone in the gym is doing. To see how the wrist stacks up against a dedicated barbell device, I compared the approaches in velocity-based training without a device.

FAQ

What velocity loss percentage should I stop a set at?

About 20% if your goal is strength and power, and 25–40% if your goal is muscle growth. Under 10% is for explosive, speed-focused work, and 40% or more means you're at or very near failure. These figures come from velocity-loss research using barbell linear transducers; if you're reading velocity off a wrist sensor, treat the percentage as your own reference and compare sets against each other rather than copying the study figures.

Is 40% velocity loss the same as muscular failure?

Close to it. In the velocity-loss literature, a 40–50% drop in bar speed corresponds to taking the set to, or very near, momentary muscular failure, while a 20% loss leaves you at about 50% of your possible reps. Worth knowing: research on proximity to failure shows that grinding all the way to failure doesn't reliably build more muscle than stopping at a moderate velocity loss, so 40%-plus is rarely the most efficient place to stop.

Is 20% or 30% velocity loss better for building muscle?

The evidence says the difference is small. The Refalo meta-analysis found hypertrophy improved a lot going from low to moderate velocity loss but barely at all from moderate to high, with no significant difference between conditions over 25% and those at 20–25%. Either works if you accumulate enough volume; the higher cutoff simply costs more fatigue per set, which matters if you train the same muscle again in 2 or 3 days.

Why not just count reps instead of tracking velocity?

Because a fixed rep count ignores daily fatigue. The same "10 reps" lands at a different proximity to failure depending on sleep, stress, nutrition, and how many sets you've already done. Velocity loss adjusts automatically: stopping at a 20% slowdown gives you the same relative effort whether you're fresh or beaten up, which is the whole advantage of autoregulation over a static program.

How can I estimate velocity loss without any sensor?

Time the lifting phase. Over the same range of motion, mean velocity is distance divided by time, so if rep 1's concentric takes 0.8 s and your last rep's takes 1.2 s, the last rep moved at about two-thirds the speed, roughly 33% velocity loss. A phone video is precise enough for this. As a rule of thumb, a rep that takes half again as long as rep 1 is past the 30% mark, and a rep that takes twice as long is at 50%, which is failure territory.

Can an Apple Watch actually measure velocity loss?

It can measure a proxy for it. A validation study found barbell velocity measured with an Apple Watch was valid against a reference device, with the wrist placement performing slightly less precisely than mounting the watch on the bar. Accelerometers drift when estimating distance, so treat the wrist as a second opinion on how much your reps slowed relative to your own earlier reps, not a lab-grade barbell readout. Riven is the Apple Watch app that scores muscle failure; it scores each set the moment you rack it.

Does velocity loss work on every exercise?

It works best on multi-rep sets of compound lifts where rep speed is consistent and you get a clean slowdown: think 6 to 15 reps on presses, squats, rows, and pulls. On heavy singles and doubles there aren't enough reps to read a reliable trend, and the velocity-to-RIR relationship also shifts by exercise and load, so on those use RPE and feel instead. Velocity is one strong signal among several, not a universal cutoff.

Sources

• Jukic, I. et al. (2023), The Acute and Chronic Effects of Implementing Velocity Loss Thresholds During Resistance Training: A Systematic Review, Meta-Analysis, and Critical Evaluation of the Literature, Sports Medicine — https://pmc.ncbi.nlm.nih.gov/articles/PMC9807551/
• Refalo, M.C. et al. (2023), Influence of Resistance Training Proximity-to-Failure on Skeletal Muscle Hypertrophy: A Systematic Review with Meta-analysis, Sports Medicine — https://pmc.ncbi.nlm.nih.gov/articles/PMC9935748/
• Paulsen, G. et al. (2025), Exercise type, training load, velocity loss threshold, and sets affect the relationship between lifting velocity and perceived repetitions in reserve in strength-trained individuals, PeerJ — https://pmc.ncbi.nlm.nih.gov/articles/PMC12360324/
• Steele, J. et al. (2017), Ability to predict repetitions to momentary failure is not perfectly accurate, though improves with resistance training experience, PeerJ — https://pmc.ncbi.nlm.nih.gov/articles/PMC5712461/
• Clemente, F.M. et al. (2021), Validity and Reliability of the Inertial Measurement Unit for Barbell Velocity Assessments: A Systematic Review, Sensors — https://pmc.ncbi.nlm.nih.gov/articles/PMC8038306/
• Velocity-Based Strength Training: The Validity and Personal Monitoring of Barbell Velocity with the Apple Watch — https://pmc.ncbi.nlm.nih.gov/articles/PMC10383699/
• Inertial measurement unit (integration drift), Wikipedia — https://en.wikipedia.org/wiki/Inertial_measurement_unit
• GymAware, Understanding Velocity Loss — https://gymaware.com/understanding-velocity-loss/