Load-Velocity Profiling for Weighted Pull-Ups: A Practical Build Guide

You added 20 kg to the belt, hit a clean set of 5, and the next session that same 20 kg felt like a different exercise entirely. Bodyweight itself is a moving variable in weighted pull-ups — sleep, hydration, and even how much you ate that morning shift the load your lats have to move before a single plate touches the chain. Percentage-based pull-up programming built around a single tested max ignores all of that. A load-velocity profile does not. It gives you a personal curve that turns bar speed into a live readout of how much weight you are actually capable of moving today, and it takes about 20 minutes to build once.

This guide walks through a one-session testing protocol built specifically for the vertical pulling pattern, shows what a realistic profile looks like, and covers the mistakes that quietly wreck the data — most of which happen in the first three warm-up sets, before anyone realizes something is wrong.

Load-velocity profiling was built and validated on barbell lifts — squat, bench, deadlift — where the external load is the entire resistance. Weighted pull-ups add a wrinkle: your own bodyweight is the base load, and only the added weight is precisely known. A 70 kg lifter doing a pull-up with a 20 kg vest is moving roughly 90 kg total, but almost every velocity-based training article treats 20 kg as the number to plot. That mismatch is a big part of why lifters copy-paste a barbell squat protocol onto pull-ups and end up with a profile that does not predict anything useful.

Sánchez-Moreno et al. (2017) addressed exactly this in a study of 24 resistance-trained men performing weighted pull-ups across a range of external loads from bodyweight up to loads producing near-maximal effort. They found that mean propulsive velocity declined in a strong, near-linear relationship with relative load (percentage of pull-up 1RM, r = 0.96), and — more usefully for daily training — that velocity loss within a set tracked the percentage of reps completed relative to maximum with high precision (r² > 0.95). The practical takeaway: velocity works just as well for autoregulating pull-ups as it does for squats, but only if you build the profile against total system load (bodyweight plus vest/belt load), not external load alone.

A second study worth knowing about here is Muñoz-López et al. (2017), who examined load-, force-, and power-velocity relationships in a prone pull-up (horizontal row) exercise using a linear velocity transducer on 21 trained subjects. They confirmed the same near-linear load-velocity pattern seen in barbell lifts (r = 0.98) and reported that peak velocity at zero external load averaged around 1.6–1.8 m/s for their sample, dropping to roughly 0.2–0.3 m/s near one-rep maximum loads — numbers that map surprisingly closely onto what we see in vertical pull-up testing, even though the exercise geometry differs. Both studies used sample sizes under 25, all male, all resistance-trained — so treat the absolute velocity numbers as a starting reference, not gospel, especially if you are newer to strength training or training as a woman, where anthropometric and strength-to-mass differences shift the curve.

You need a velocity sensor that clips to a belt or vest (not the bar, since the bar does not move) — this is the single most common setup error people make when adapting barbell VBT gear to pull-ups. Here is the protocol we recommend, built for a single 25–30 minute session:

1. Warm up thoroughly. 5 minutes of scapular activation and band pull-aparts, then 2 sets of 5 bodyweight pull-ups, resting 90 seconds.
2. Rep 1 — bodyweight only. Perform 2 reps at maximal intended velocity. Record mean propulsive velocity. This anchors the top of your curve.
3. Add load in fixed increments. Add 5 kg (or 10–15% of your estimated pull-up max, whichever is smaller) and perform 2 reps. Rest 2–3 minutes.
4. Repeat for 4–5 total loads, working up toward a load where velocity drops to roughly 0.20–0.25 m/s — this is close to your true 1RM zone, so the last load should feel hard but the reps must stay clean (full lockout, no kipping).
5. Stop the session once velocity on a new load is below 0.15 m/s or technical breakdown appears (excessive leg drive, incomplete range of motion). Do not push to true failure — you are profiling, not testing a 1RM directly.
6. Record total system load (bodyweight + external load) against mean velocity for every load, not just the external weight.

A 70 kg lifter with a target working max around 25 kg added load would typically test at total loads of roughly 70, 75, 82, 90, and 95 kg across the session — five data points is the minimum for a usable curve, and six is better if your schedule allows the extra 5 minutes.

Below is a representative profile from a 78 kg intermediate lifter (roughly 3 years of structured pulling work) tested using the protocol above. Total system load includes bodyweight plus any added vest/belt weight.

The curve for this lifter is close to linear between 0.21 m/s and 1.0 m/s, which is typical, but it flattens noticeably above 1.1 m/s — bodyweight-only pull-ups have a velocity ceiling that added-load reps do not show, because there is a limit to how fast a human can pull their own mass through that range of motion regardless of intent. This is one reason we do not recommend anchoring your entire curve on the bodyweight data point alone; use the 3–4 loaded points as the backbone of the relationship.

Once you have five or six points, plotting them (even in a basic spreadsheet) lets you estimate a load for any target velocity. If this lifter wants to work at 0.45 m/s for a strength-focused block, interpolating between the 98 kg and 106 kg points suggests roughly 101–102 kg total load, or about 23–24 kg of added weight — a number the lifter would never have landed on by feel alone on a day when the nervous system was running 5–8% below normal.

Most bad pull-up profiles are not caused by bad math — they are caused by inconsistent reps at the testing stage. Three mistakes show up over and over:

Kipping or leg-driving on max-velocity reps

Any hip snap or leg swing artificially inflates velocity readings at the exact loads where you need the cleanest data — usually the mid-range loads where lifters start to feel the weight and unconsciously recruit momentum. If your torso is swinging more than a few centimeters at the top of the rep, that data point should be discarded and retested with a 60-second pause to reset stillness before pulling.

Skipping the bodyweight anchor point

Some lifters jump straight to loaded sets to save time. Without the bodyweight data point, you lose the top of the curve and every velocity zone estimate downstream shifts. It only costs two reps — do not skip it.

Testing fatigued or under-recovered

Building a profile the day after a heavy back session, or first thing in the morning without a proper warm-up, will shift your entire curve downward. The resulting profile is not wrong exactly — it is a profile of your fatigued state, not your baseline. Test on a day that represents a normal training state, and note the date so you know when the data needs refreshing.

Using external load instead of total system load

This is the error we see most often when people import a barbell VBT spreadsheet template. If you plot 20 kg vest weight against 0.50 m/s without accounting for your 75 kg bodyweight, your percentages will be meaningless the moment you compare sessions where bodyweight fluctuated by even 1–2 kg from water weight or a big meal.

Once the profile exists, weekly programming becomes a lookup rather than a guess. A simple three-tier approach works well for most lifters running a 4–8 week pulling block:

• Strength-speed days (0.65–0.85 m/s): Lower added load, 3–5 reps per set, 3–4 sets, full recovery between sets (2–3 minutes). Good for weeks 1–2 of a block or as a recovery-adjacent session after a heavy week.
• Strength days (0.40–0.55 m/s): The bulk of a hypertrophy-and-strength block sits here. 3–5 reps, 4–5 sets, stop the set once velocity drops 20–25% from the first rep.
• Max-strength days (0.20–0.30 m/s): Singles and doubles, longer rest (3–5 minutes), reserved for the final 1–2 weeks before a testing session or competition.

On any given day, do a quick 2-rep check at a known load before committing to the planned session. If velocity comes in more than 10% below what your profile predicts for that load, drop the planned added weight by roughly 5–10% rather than grinding through sets that were programmed for a version of you that showed up better rested. This single adjustment step is what separates autoregulated pull-up training from a spreadsheet that just happens to have velocity numbers on it.

A load-velocity profile is a snapshot, not a permanent fixture. Retest under these conditions:

1. Every 4–6 weeks during a dedicated strength block, timed for a day when you are reasonably fresh — not the day after a max-effort session.
2. After a bodyweight change of more than 2–3 kg, since total system load shifts even when your strength has not changed at all. This is common after a diet phase or an off-season bulk.
3. Following any layoff longer than 10–14 days, since detraining affects the velocity curve unevenly — max-strength reps typically slow down faster than speed-strength reps.
4. When your day-to-day velocity readings consistently miss the profile's predictions by more than about 8% across several sessions, which usually signals the curve is stale rather than that you are simply having a bad week.

Full retesting only takes the same 20–25 minutes as the original session, and most lifters find it easier the second time because the loading increments can be tailored to the previous curve instead of guessed from scratch.