Top set and back-off sets: how to program the drop and the reps

A squat of 120 kg for five at RPE 8 gives back-off sets of 105 kg. That is 88 per cent of the top set, rounded down to a weight you can load, and it is the default the app uses for a top set and back-off scheme. Do two back-off sets at the top of your rep range, and let the whole block move when the top set does.

The rest of this guide explains each part of that sentence: why one heavy set and then lighter ones, how big the drop should be, how many back-offs to stack, and what happens to all of it next session.

What a top set and back-off scheme is for

A top set is the heaviest working set of the day for a lift, done first, at a target effort. A back-off set is a working set done after it at a lower load. The top set is where strength shows up. The back-offs are where you pile up hard sets you can recover from.

The two jobs pull in different directions. A heavy five at RPE 8 is a good strength stimulus, but three or four of them in a row leave you grinding by the last one. Drop the load a little and the same lifter can do several more sets with clean reps and a few reps in reserve. RPE here is the reps-in-reserve version of the scale: RPE 8 means you could have done two more reps [1].

The scheme also gives you a daily reading. The top set is a measured effort at a known load, so it tells you how strong you are today before you commit to the volume. That is the main argument for letting the top set, not a fixed percentage of an old max, set the back-off load.

Research on RPE-based loading backs the idea without proving it is better. In an eight-week study of 21 trained men, a group that chose loads by RPE and adjusted the next set by 2 per cent for every half point off target gained 17.1 kg on the squat, against 13.9 kg for a group on fixed percentages of 1RM [2]. The difference was not statistically significant. Both approaches worked. RPE simply did no worse, and it adapts on bad days.

How big should the drop be?

The useful range sits around 80 to 90 per cent of the top set, which is the span the table below works through. A smaller drop keeps the back-offs heavy and close to the top set's reps. A bigger drop lets you do more reps or more sets at the same effort. Neither is wrong. They trade load for volume.

The most direct test of that trade is a three-week study of 12 nationally qualified powerlifters. Each session opened with a top set at a target RPE, then back-off sets of the same reps at a lower load, stopped once a set reached the target RPE again [3]. The researchers ran a 2, 4 and 6 per cent drop in different weeks. Combined weekly relative volume load across squat, bench and deadlift rose from 74.6 in the 2 per cent week to 88.4 at 4 per cent and 114.4 at 6 per cent [3]. A bigger drop bought more sets before effort caught up.

Those were powerlifters doing back-offs for the same reps as the top set, so the drops were small. A lifter training mostly for size usually wants more reps on the back-offs than on the top set, and that takes a bigger drop. The table below shows what the app's estimated 1RM model says each drop allows for our 120 kg lifter. Their set of five at RPE 8 works out to an estimated 1RM of 148 kg.

The reps column is the highest rep count whose RPE 8 load on the same chart is at or above the back-off load. For eight reps at RPE 8 the chart gives 109.4 kg, and for nine it gives 104.6 kg, so 105 kg sits just inside eight. The same lifter doing 105 kg for only five reps would be easier than RPE 6, since the chart puts five reps at RPE 6 at 112.8 kg. That is too easy to count as a hard set for most goals.

So match the drop to the reps you want. If your back-offs are fives like the top set, 88 to 90 per cent keeps them honest. If you want eights or tens, use 80 to 85 per cent. The one-rep max calculator runs the same model if you want to check a different top set.

How many back-off sets, and which reps

Two back-off sets is the default in the app, and a reasonable start for a main lift. Add a third or fourth when the lift is the main source of weekly sets for that muscle and your recovery holds up. Fewer than two rarely justifies the scheme over straight sets.

The case for more sets is about the week, not the session. A meta-analysis of 15 studies found each additional weekly set was linked to about 0.37 per cent more muscle growth, with shrinking returns as volume climbs [4]. Back-offs are an easy place to add those sets because they cost less than another top set. Our guide to weekly sets per muscle covers where the total should land.

Reps on the back-offs follow the rep range on the lift. The app suggests the top of your range for each back-off. A squat programmed at five to eight reps shows a top set target of five and back-off suggestions of eight. You can override any of them, and the next suggestion uses what you actually logged.

An alternative from the research is to stop by effort instead of by count. Do back-offs until one reaches your target RPE, then end the lift [3]. It works well, but it needs honest RPE ratings, and lifters new to the scale tend to stop too early. A fixed count is easier to plan and to compare week to week.

Worked example: one session and the next

Here is a full squat session for a lifter whose rep range is five to eight, target RPE 8, with the default two back-offs.

The warm-up ramp is what the app's warm-up generator gives for 120 kg. Warm-ups never count toward volume, so the session has three working sets of squat, not seven. On a 20 kg bar, 105 kg is a 25, a 15 and a 2.5 kg plate each side, and the plate calculator will lay out any other load.

The back-off load is worked out from the heaviest set you logged in that session. If the top set felt heavy and you took 117.5 kg instead, the back-off suggestions follow it down. Nothing is locked to last week.

Next session depends on the top set's RPE. If 120 kg for five came in at RPE 8, right on target, the model suggests 122.5 kg for the next top set and back-offs of 107.5 kg. If it came in at RPE 9, one point harder than planned, the suggestion holds the top set at 120 kg and the back-offs at 105 kg. A top set that badly overshoots its target is treated as a grinder: the load holds, and the back-off is not recalculated down from one bad day.

If the top set is logged without RPE, the model counts reps instead. Hit the top of the rep range on the top set and the load goes up by one step, which is 2.5 kg on a barbell lift. The back-offs then move with it, because they are always a share of the top set.

Three more rules can stop an increase. On a cut, a load increase never goes through. On a deload week, the scheme never emits an increase either. And three sessions in a row without progress flag the lift as stalled by default, which is your cue to change something rather than repeat it a fourth time.

Mistakes that stall the scheme

The most common one is chasing the top set and letting the back-offs decay. A lifter adds 2.5 kg a week to the top set at RPE 9 or 10, then does the back-offs half-heartedly because there is nothing left. The top set is supposed to land at your target RPE. If it lands two points over, it is a test, and the next session should hold.

A second mistake is fixing the back-off load to an old max. If you write "back-offs at 100 kg" into a plan and keep it while the top set climbs, the drop grows each week until the back-offs stop being hard sets. Tie the back-off to the top set, as a percentage, and it stays in the right place.

The third is logging the back-offs as warm-ups or leaving their role blank. A back-off is a working set. It counts toward weekly sets per muscle and toward tonnage, and it should be marked as one. A warm-up logged as a working set does the opposite damage and inflates your volume.

Skipping RPE on the top set causes its own stall. Without it, the model treats the set as a max, which drags the estimated 1RM down and can hold back the next weight. Our post on what RPE a top set should be covers how hard to push it and how to rate it.

Finally, lifters often make the drop too small for the reps they want. Back-offs at 90 per cent for eight reps will push most lifters past RPE 8 on the second set. Check the table above, or run your own numbers, and choose the drop that fits the reps.

When straight sets are the better choice

A top set and back-off scheme earns its keep on lifts where you have a clear strength goal and a stable technique: squat, bench press, deadlift, overhead press, a row. On accessories and machines, the top set adds a heavy effort you do not need. Straight sets with double progression are simpler there and progress just as well.

It also suits lifters who already rate RPE with some accuracy. If you are new to the scale, start with straight sets for a block and practise rating them. Then move your main lifts to a top set and back-offs once your RPE numbers match what the bar is telling you.

Questions

What percentage should back-off sets be?

A drop to 80 to 90 per cent of the top set covers the useful range, and the app's default is 88 per cent. Use the higher end when back-offs match the top set's reps, and the lower end when you want more reps on the back-offs. For a 120 kg top set of five at RPE 8, 85 per cent gives 102.5 kg, which the estimated 1RM model puts at about nine reps before RPE 8.

Do back-off sets count toward weekly volume?

Yes. A back-off set is a working set, so it counts toward weekly sets per muscle and toward tonnage, the same as the top set. Warm-ups never count. That is why a session of four warm-ups, one top set and two back-offs counts as three sets for the muscles that lift trains. Mark each set's role when you log it so the count stays right.

Should back-off sets go to failure?

Usually not. The point of the drop is volume you can recover from, so stop back-offs one to three reps short of failure, which is RPE 7 to 9 on the reps-in-reserve scale [1]. Taking the last back-off to failure now and then is fine on safer lifts. On squats and deadlifts, hold a rep or two in reserve so technique does not break down late in a heavy session.