Barbell row: how to log it honestly and progress the load

A barbell row is a horizontal pull done standing, hinged at the hips with the torso leaning well forward, the bar pulled from straight arms to the lower ribs and lowered under control. It is one of the easiest lifts in the gym to cheat without noticing. A little more hip drive, a little more torso lift, and the load climbs while the back does the same work. For a 75 kg lifter rowing 80 kg, eight reps at RPE 8 is worth an estimated 108.3 kg for one, computed with the model behind the one-rep max calculator. Ten reps ground out at RPE 10 are worth the same 108.3 kg, which is the whole argument for rating every set.

Every figure below is compute.mjs output, the script that runs the app's own formulas, or carries a citation to the study it came from.

What the barbell row trains

In the app's exercise library the barbell bent-over row is a compound horizontal pull. The back is the primary muscle, and the biceps and rear delts are secondary. Of the 14 muscles the app sets weekly targets for, those are the three the lift touches.

The best direct look at the lift comes from a small study of seven healthy men who did three rowing variations while their trunk muscles and spine were measured [1]. The standing bent-over row produced large activation symmetrically across the back, and also the largest lumbar spine load of the three [1]. The inverted row, a bodyweight pull under a fixed bar, produced the highest latissimus dorsi activation with less load on the lower back [1].

Read that as a trade. The barbell row lets you load the whole back heavily, and the lower back pays for holding the hinge. A row that gets heavier by turning into a half deadlift adds load to the spine without adding work for the back.

How to log it

A barbell row is a load lift: not per-side, not bodyweight plus load, one line per set with the weight on the bar. The sets on the way up are warm-ups. On a 6 to 10 rep pull every working set is a normal set at the same load. There is no top set, and the app reads all of them when it decides the next weight.

Rate every working set with RPE, where 10 means no reps left, 9 means one and 8 means two. The scale was built around reps in reserve so that load can be adjusted from set to set to match what a lifter actually has that day [2]. Its authors suggest it may be more precise than a general effort rating close to failure [2], which is where the last reps of a row live.

The rating matters more here than on most lifts. A set logged without RPE is treated as RPE 10, the assumption every classic formula makes, so 80 kg for eight with the rating blank estimates 101.8 kg instead of 108.3 kg. At RPE 9 the same eight reps estimate 105 kg.

The honest part is what the rating describes. Rate the reps you could have done in the same form as rep one: same torso angle, same bar path, no heave. Reps that only happen because the hips joined in are not reps in reserve.

Three working sets and three warm-ups, run through the app's volume code, give the back 3 sets and the biceps and rear delts 1.5 each. Secondary muscles count half a set. The warm-ups add nothing anywhere, as the post on whether warm-up sets count as volume explains.

The warm-up ramp for 80 kg

The app builds the ramp from the planned weight of the first working set. For 80 kg, compute.mjs warmup --kg 80 returns three sets.

The working weight is a 25 and a 5 on each side, from the plate calculator. Move the working weight up one step to 82.5 kg and the ramp gains a fourth set, a 70 kg double at 85 per cent. Lifters who rowed after squats or deadlifts can drop the bar set: the hinge is already warm and only the pull needs grooving.

What 80 kg for eight at RPE 8 is worth

Eight reps with two in reserve means 80 kg was a 10-rep load that day. The model reads the reps-by-RPE chart along its diagonal, reps plus reps in reserve, and lands on an estimated max of 108.3 kg. Epley says 101.3 kg and Brzycki 99.3 kg, because both treat the eighth rep as the last one available. The estimated 1RM guide covers how the chart is read.

Nobody should test a one-rep max on a row. The number is a reading of strength from session to session. The ladder from compute.mjs ladder --kg 80 --reps 8 --rpe 8 gives the heaviest load for each rep count at RPE 10 on that day.

Read it at RPE 8 instead and the 6 to 10 range runs from 85.1 kg for six to 73.6 kg for ten, from the same model. That spread is why one load can serve a whole rep range: at 80 kg, six reps leave plenty in reserve and ten get close to the edge.

How double progression moves the bar

The app has seven progression schemes: Auto, Double, Top set + back-off, Load, Reps, RIR and Manual. For three sets of 6 to 10, Double fits. The rule holds the load, chases one more rep each session, and adds the step once every working set reaches the ceiling at the target RPE. For a barbell lift that step is 2.5 kg, computed with the app's model. The post on double progression covers the rule in full.

At 80 kg the block looks like this. Sets of 7, 6, 6 hold the load. Sets of 10, 9, 8 still hold it. Sets of 10, 10, 10 at RPE 8 earn 82.5 kg, and the reps reset to six. The first session at the new load reads lower on the estimate, and that is expected: 82.5 kg for six at RPE 8 is worth 105 kg against 117.6 kg for the ten at 80 kg that earned it. Like with like, 80 kg for six at RPE 8 was worth 101.8 kg, so the six at 82.5 kg is real progress.

The app's next-weight model shows what the rating does to a single logged set. Against a target of eight reps at RPE 8, 80 kg for eight at RPE 8 moves to 82.5 kg. The same set at RPE 9 holds 80 kg. At RPE 10 it eases to 75 kg, because a set that hard at the target reps means the load ran ahead of the lifter.

A planned deload week never emits an increase, and a load increase never leaks through on a cut. More than 30 days off resets the lift: the same 80 kg for eight after 35 days away comes back as a 75 kg suggestion.

Where rows inflate

Rows inflate through form, and the log cannot see form. It can only see what you type, so the defence is a rule you set before the set starts.

The common version is torso angle. As the load rises the lifter stands taller, the bar path shortens and the pull turns partly into a shrug. Ten reps at 80 kg logged at RPE 10 because the last reps needed a heave are worth 108.3 kg, the same as eight clean reps at RPE 8. Log the heaved set as ten at RPE 8 and it reads 117.6 kg, strength that never happened. Double progression then adds load on a ceiling the back did not reach.

Hip drive is the second. A row started with a bounce out of the hinge moves more weight and loads the lower back, the region the bent-over row already loads hardest of the variations measured [1]. A rep that needed the hips counts as the end of the set.

Range is the third. A bar that stops short of the ribs is a partial rep. Pick a touch point, the lower ribs or the top of the belt, and a rep that misses it does not count.

How many weekly sets of back work?

The research gives a direction, not a number for rows. A meta-analysis of 15 studies found each extra weekly set linked to about 0.37 per cent more growth, though splitting volume into fewer than 5, 5 to 9 and 10 or more sets per muscle showed only a trend [3]. A later review of trained young men suggested 12 to 20 weekly sets per muscle group as a possible standard and called for more research before it becomes one [4].

Counting sets works as a volume measure when each set goes to or near failure in a range of about 6 to 20 or more reps [5]. Three rows at RPE 8 qualify. Three rows logged at RPE 8 that were really RPE 5 do not, which is one more reason the rating has to be honest. The app's weekly targets cover 14 muscles in plain English, and the post on how many sets per week per muscle covers the research in full. Rows, pulldowns and pull-ups all add to the same back total.

What changes when you swap it?

A dumbbell row changes the step. The app's increment for a dumbbell compound lift is 2 kg against 2.5 kg for a barbell, computed with the app's model, so a rack that jumps in bigger steps leaves the rule adding reps for longer between loads. One arm at a time also means a bench to brace on and less load on the lower back.

An inverted row trades load for back-friendliness. It produced the highest lat activation and a lower spine load than the bent-over row in the same study [1]. A chest-supported row removes the hinge entirely, so the lower back stops being the limit and the reps stop being inflated by hip drive.

Whatever you swap in starts its own history. The 108.3 kg estimate stays with the barbell row, and a new lift begins with a new first session rather than borrowing the old one's numbers.