Double progression: the simplest way to keep adding weight

Double progression is a rule for one lift. Hold the load, add reps inside a fixed range until every working set reaches the top of it, then add the smallest step the bar allows and start again at the bottom. For a barbell lift the app's step is 2.5 kg, computed with the app's model. A dumbbell lift steps by 2 kg and a machine by 1 kg. Run on a barbell row for three sets of 8 to 12 at RPE 8, the rule turns 60 kg for 8 into 62.5 kg for 8 at the same effort within six weeks, which moves the estimated one-rep max from 81.2 kg to 84.6 kg.

That is the whole method. The rest is the detail that makes it hold up: the range and the step, why the ceiling only counts at the same RPE, which sets have to reach it, and what the rule does when a session goes badly. Every number below is computed with the app's model or cited to the study it comes from.

Why double progression works

Progressive overload has two cheap levers, reps and load, and double progression pulls them in turn. The research says the order barely matters for muscle. Plotkin and colleagues gave 43 lifters with at least a year of lower-body training eight weeks of four sets of four exercises, twice a week [1]. One group added load inside a fixed rep range and the other added reps at a fixed load. Rectus femoris growth modestly favoured the rep group, by 2.8 mm summed across sites, squat strength slightly favoured the load group, by 2.0 kg, and the authors called both differences of questionable practical significance [1].

Chaves and colleagues ran the same comparison inside one body. In 39 untrained adults, one leg progressed the leg extension by load and the other by reps for ten weeks [2]. The load leg's one-rep max went from 52.9 to 69.1 kg and the rep leg's from 51.7 to 66.8 kg. Vastus lateralis cross-section went from 21.3 to 23.5 cm² and from 21.1 to 23.4 cm², and neither difference was significant [2].

What double progression adds to either lever alone is a stopping rule. Reps alone drift toward sets of twenty, and load alone stalls the first week the bar refuses to move. Alternating them keeps every set inside a range you chose, and where that range sits matters less than lifters expect. A meta-analysis of 21 studies with every set taken to failure found hypertrophy similar whether loads were above or below 60 per cent of 1RM, while one-rep max strength favoured the heavier loads [3]. Pick the range for the lift, and keep the sets honest.

Pick the range by the lift, and the step by the bar

A range needs to be at least two reps wide, or there is nowhere to progress before the load moves. In the app, the Auto scheme reads a program exercise with a range that wide as a Double exercise and one with a fixed rep count as a Load exercise, so under Auto 3 sets of 8 to 12 runs as double progression and 3 sets of 5 does not. Eight to twelve suits rows, dumbbell presses, pulldowns and most machines. Five to eight or six to ten suits squats, deadlifts and the bench press, where the twelfth rep of a set is usually a different lift from the first.

The research does not force a range on you. Helms and colleagues put strength, hypertrophy and muscular endurance training in the same band of effort, sets finished near failure or a few reps short of it, at RPE 7 to 10 [4]. They noted that 6 to 12 reps may save time per session rather than grow more muscle [4]. The range is a choice about the lift and your week.

The step is the smallest load the equipment allows. The app's default for a barbell compound lift is 2.5 kg, for a dumbbell compound 2 kg and for a machine 1 kg, all computed with the app's model. On a 20 kg bar, 62.5 kg loads as one 20 kg plate and one 1.25 kg plate a side, which the plate calculator will confirm for any target. When the step is a large share of the load, as 2.5 kg is on a light overhead press, the app asks you to bank reps beyond the range before it writes the increase.

The ceiling only counts at the same RPE

RPE is a 1 to 10 rating of how hard a set was, read as reps in reserve: RPE 10 means no reps left, RPE 9 means one left, RPE 8 means two, and so on down the scale [5]. Double progression needs it because the ceiling only counts if the effort stayed put. Twelve reps at RPE 8 on 60 kg is worth 92.4 kg for one, computed with the app's model. Twelve at RPE 9 is worth 90.3 kg and twelve at RPE 10 is worth 88.2 kg. That last number is exactly what 60 kg for ten at RPE 8 is worth, so a ceiling reached by grinding is a set of ten wearing a bigger number.

A set logged without an RPE is treated as RPE 10, the assumption every classic formula makes. 60 kg for eight with no rating estimates 76.3 kg, against 81.2 kg for the same set at RPE 8. Leave the field blank and every set reads as a max, which is why a stalled lift in a log without RPE is often a lift that never stalled. The one-rep max calculator shows the same conversion for any set.

Ratings get better the closer a set ends to failure. Helms and colleagues report that experienced squatters' RPE scores grew more consistent as the load rose, with standard deviations of 1.18, 0.97, 0.92 and 0.32 at 60, 75, 90 and 100 per cent of 1RM, and they advise novices to practise recording reps in reserve before progressing on it alone [4]. A ceiling at RPE 8 sits two reps from failure, close enough that the rating means something and far enough that the next two sets survive. The same review warns that training to failure can cost performance on the sets that follow [4].

Closer to failure is still the direction for growth. A series of meta-regressions found a negligible relationship between estimated reps in reserve and strength gain, while hypertrophy improved as sets ended closer to failure [6]. The authors estimated reps in reserve from each study's description and call the analysis exploratory, so the shape of that curve is not settled [6]. The practical reading is narrow: set the ceiling at RPE 8 or 9, not RPE 5, and log it so the rule can see it.

Six weeks on a barbell row, in numbers

Take a 90 kg lifter rowing three sets of 8 to 12 at RPE 8 on the Double scheme with its default trigger, every set reaching the ceiling. The load starts at 60 kg. Every estimate in the table is for the best set of the session, computed with the app's model.

Week 3 is where the trigger matters. The best set reached 12, so a best-set trigger would have added the 2.5 kg there, two weeks earlier. Under the default the rule waits for the other two sets, and in week 4 it waits again, because 12, 12, 11 is one rep short on one set and averages under 12. The lifter spent those two weeks adding reps to the second and third sets, which is work the best-set trigger would have skipped.

Week 6 looks like a step backwards and is not. The estimate falls from 92.4 kg to 88.4 kg because the model rates twelve reps with two in reserve above nine reps with two in reserve, whatever the load. The honest comparison is the same reps at the same RPE: 60 kg for 8 at RPE 8 in week 1 was worth 81.2 kg, and 62.5 kg for 8 at RPE 8 in week 6 is worth 84.6 kg. The post on progressive overload for hypertrophy walks through why the estimate moves that way on every load step.

The model also says why the reps come back quickly after a step. From 60 kg for 12 at RPE 8, the rep-max ladder puts the 8-rep max at 72.6 kg and the load for 8 at RPE 8 at 68.3 kg, both computed with the app's model. A step to 62.5 kg is well inside that. Double progression under-steps on purpose, and the reps it hands back are the margin.

Which sets have to hit the top?

The app offers three triggers for a Double exercise, and the one you pick changes how fast the load moves.

• Every set reaches the ceiling. The default. The load moves least often and the back sets do the most work.
• The sets average the ceiling. In the table, 12, 12, 11 does not fire and 13, 12, 11 does. The Auto scheme uses this trigger when its pace is set to aggressive.
• The best set reaches the ceiling. The load moves soonest, and the second and third sets will miss the floor more often at the new weight.

Pick every set for lifts where the third set does the growing: rows, curls, leg extensions, anything where set one is never the hard one. Pick best set for a lift you run for one or two working sets, where the first set is the honest one. Whichever you pick, write it into the exercise before the block starts, so the decision is made at a desk and not under the bar.

When the reps do not come

A bad session does not move the load down. If every working set misses the floor of the range once, the rule holds the weight and calls it a possible off day. If every set misses the floor again at the same load, it eases the weight down by one step, to 57.5 kg in the example, and sets the target back to the ceiling. One miss on one set changes nothing at all. The rule reads the whole session.

Three sessions at the same load without progress flags a stall by default, with one exception. If the estimated one-rep max has been rising across those sessions because the reps have climbed at a fixed load, the rule does not call it a stall. That is what double progression is supposed to look like. A planned deload week never emits an increase, however good the sets were. More than 30 days off a lift resets it rather than resuming. On a cut a load increase never goes through, and the reps and the RPE carry the progress instead.

The rule also notices when the load was too light. If every set beats the ceiling by two or more reps, 14, 14, 14 on an 8 to 12 range, the next session jumps two steps, to 65 kg, and resets to 8. The numbers back that up: 60 kg for 14 at RPE 8 is worth 96.6 kg, flagged as a rough estimate, and 65 kg for 8 at RPE 8 is worth 88.0 kg, so even the double step leaves reps in hand. All of this runs on the reps and RPE logged on every working set, which is what the home page means by automatic progressive overload.

Mistakes that look like double progression

Reaching the ceiling at RPE 10. Twelve reps to failure on 60 kg is worth 88.2 kg, the same as ten at RPE 8, and the load step that follows lands on a lifter who has not got stronger. The ceiling is twelve at the RPE you wrote down, and if the RPE drifts up, the set is not there yet.

Changing the trigger when the back sets lag. Switching from every set to best set in week 3 because the second set is stuck at 10 is a different rule, not progress. If the back sets never catch up, the range is too wide for that lift or the rest between sets is too short.

Widening the range instead of stepping down. Turning 8 to 12 into 8 to 15 when the load will not move keeps the number on the bar at the cost of the thing the range was for. Take the step down the rule offers and rebuild.

Taking two steps because one feels small. A 5 kg jump on a 60 kg row is the move the rule reserves for a session where every set beat the ceiling by two. Taken after a plain 12, 12, 12 it spends the margin the rule just built.

Logging the reps and not the effort. Without RPE the model reads every set as a max, so progress inside the range is invisible and the first week at a new load looks like a regression. One rating a set is the price of the rule working.

Questions

What rep range should you use for double progression?

Any range at least two reps wide works, and the research does not favour one for muscle: hypertrophy was similar above and below 60 per cent of 1RM across 21 studies with sets taken to failure [3]. Use 8 to 12 on rows, dumbbell presses and machines, and 5 to 8 or 6 to 10 on squats, deadlifts and the bench press, where the step is a smaller share of the load.

Should every set reach the top of the range before adding weight?

By default, yes. Waiting for every set keeps the second and third sets doing real work and means the new load will hold the floor of the range. The best-set trigger moves the load sooner but leaves the back sets to catch up at a heavier weight. Choose one before the block starts and keep it for the whole block.

How much weight do you add in double progression?

The smallest step your equipment allows. The app's defaults, computed with its model, are 2.5 kg on a barbell compound lift, 2 kg on a dumbbell compound and 1 kg on a machine. When that step is a large share of the load the app asks you to bank reps beyond the range first. A two-step jump is reserved for a session where every set beat the ceiling by two or more reps.

Does double progression work for strength as well as size?

Yes, with a lower range. One-rep max strength favoured heavier loads across the 21 studies Schoenfeld and colleagues pooled, while hypertrophy did not care [3]. Run 3 to 5 or 4 to 6 on the main lifts, keep the ceiling at RPE 8 or 9, and let the estimated one-rep max on your working sets, not a test day, tell you the load is moving.