Double progression: how it works and how to run it
Double progression adds a rep until you top out your range, then adds weight and resets the reps. Here is the rule, the edge cases, and a worked example.
Double progression answers one question with one rule: what goes on the bar next session. Hold the weight and add a rep until you reach the top of your rep range, then add the smallest plate jump you own and start again at the bottom. That question is where most programs quietly leave you on your own, and this rule keeps answering it for years.
What is double progression?
Double progression means you progress two variables, reps and weight, but never at the same time. You pick a rep range instead of a rep count. You hold the weight until you reach the top of the range on every set. Then you add weight and drop back to the bottom of the range.
The name is the method. Two things go up, in strict order, one at a time.
Compare it to what most people do instead. Linear progression adds weight every session at a fixed rep count, which works beautifully for about four months and then falls off a cliff. Percentage programs hand you 82.5 percent of a max you tested in March and no longer have. Both need you to know something you do not know. Double progression only needs you to know what happened in the last session.
Why add reps before weight?
Because reps are the smaller increment.
The smallest plate jump most gyms have is 5 lb, which is 3.7 percent of a 135 lb bench press. That is a large step to ask for in one week. One extra rep at the same weight is a much smaller step, and there are three or four of them between plate jumps. So the rep range is the ramp: you climb it gradually, and the plate jump lands only when you have built enough room to absorb it.
This is also why the range should be wide enough to matter. A range of five to six gives you one rep of runway. Five to eight gives you three, which is enough for the jump to feel earned rather than optimistic.
What counts as a successful session?
Every working set at or above the target, all at one weight.
That last clause does the real work. Three sets of eight at 135 lb is a success. Eight, eight, six is not. Eight at 135 and eight at 125 is not either, because the sets were not at one weight and there is no single number to progress from. Anything less than all sets at one weight leaves you arguing with yourself about partial credit, and the argument is always won by the version that adds weight sooner than you earned it.
A rep in reserve is fine. You do not need to grind every set to failure for this to work, and the research on training to failure suggests you get most of the adaptation without it, at a much lower fatigue cost. What matters is that the target was met on every set.
When do you add weight, and how much?
When the top of the range is met on every set. Then you add one jump and reset the target to the bottom of the range.
The jump should be the smallest one you can actually load:
- 5 lb (2.5 kg) on barbell compounds: squat, bench, deadlift, row, overhead press.
- 2.5 lb (1.25 kg) on isolation work: curls, lateral raises, triceps, rear delts.
- 10 lb (5 kg) on leg press and other machines where the stack has no finer setting.
Kilogram lifters should use 2.5 kg rather than a converted 2.27 kg. Progression should land on numbers you can actually put on a bar.
Here is a full cycle on a bench press with a five to eight range and three sets:
| Session | Weight | Target | What you did | Next session |
|---|---|---|---|---|
| 1 | 135 lb | 5 | 5, 5, 5 | 135 lb for 6 |
| 2 | 135 lb | 6 | 6, 6, 6 | 135 lb for 7 |
| 3 | 135 lb | 7 | 7, 7, 6 | 135 lb for 7 again |
| 4 | 135 lb | 7 | 7, 7, 7 | 135 lb for 8 |
| 5 | 135 lb | 8 | 8, 8, 8 | 140 lb for 5 |
| 6 | 140 lb | 5 | 5, 5, 5 | 140 lb for 6 |
Six sessions, one plate jump, and no decision made at any point. Session 3 is the important one: the miss cost nothing except a repeat.
What do you do after a missed set?
Nothing. The target stays where it is and you try it again next session.
There is no deload, no percentage back-off, no reset to the bottom of the range. A missed set is information about one day, and one day is usually about sleep or lunch rather than about your program. Log what you actually got, honestly, and repeat the target.
The honest log matters more than it sounds. If you round eight, eight, six up to three sets of eight, you will add weight you have not earned, miss the next session too, and conclude that double progression does not work. It works. The record has to be true.
How fast should this go?
Slower than you want, and you can choose how much slower.
At one successful session per bump, a five to eight range takes four clean sessions to earn a plate jump. On a four-day split where each lift comes around once a week, that is about a month per jump. That is the honest pace for an intermediate lifter, and it adds up to roughly 60 lb a year on a main lift, which is a great year.
If you are recovering badly, or coming back from a layoff, you can require two or three successful sessions before each bump instead of one. The climb gets slower and the sessions get more repeatable. This is the correct adjustment when progress starts costing you sessions.
What double progression does not solve
It is a progression rule, not a program. It has nothing to say about:
- Exercise selection. It will happily progress a badly chosen movement.
- Volume. Three sets or five sets, it applies the same rule to both.
- Fatigue over months. If everything stalls at once, that is a recovery problem, and no rule about reps will fix it.
- The true plateau. When one lift stops for two or three months while the rest keep moving, the answer is a variation of the movement or a change in set count, not a smaller plate.
It solves the decision, which is the part that actually gets skipped.
How Compound runs it
Compound applies this rule automatically to every lift in the program, which is what the app is for. Each exercise carries its own progression, so your bench and your row climb at their own paces without you tracking either.
You tap a circle for each set you complete. If all your sets hit the target at one weight, the app raises the target by a rep for next time. When you top out the range it adds the jump for that exercise and resets the reps. If you got fewer reps than the target, you long-press the circle and log the real number, and the target repeats.
You can also tell it to wait: two to five successful sessions per increase instead of one, if you would rather climb slowly. The rule is the same. The bookkeeping is not yours any more.
Common questions
What rep range should I use for double progression?
Five to eight reps for heavy compound lifts, eight to twelve for isolation work, twelve to twenty for small muscles and high-fatigue movements like calves or lateral raises. The range matters less than staying in it long enough to climb through it. Any range works if you actually add the rep.
Does double progression work for bodyweight exercises?
Yes, up to the top of the range. Pull-ups and dips climb reps exactly the same way, but when you top out there is no plate to add. At that point you either add external load with a belt, or hold at the top of the range and add a set.
What if I hit the target on some sets but not all of them?
That is not a successful session, and the target stays where it is. Partial credit is what turns a progression scheme into a spreadsheet argument. All sets at the target on one weight, or the target repeats.
How is this different from linear progression?
Linear progression adds weight every session at a fixed rep count, which works for a beginner and then stops abruptly. Double progression adds reps between weight jumps, so each jump is earned across several sessions and the same rule still applies two years in.
How long should one weight take?
On a five to eight range, four successful sessions takes you from the bottom to the top, so a jump every four to six weeks per lift is normal once you are past the beginner phase. If you are adding weight every week on a compound lift, you are still in the beginner phase and should enjoy it.