When should you add weight at the gym? The exact rule
Add weight too soon and you trade technique for injury risk. Add it too late and you plateau for months. The good news: there is a concrete, evidence-backed rule for knowing exactly when it is time. It is called double progression and it fits in two sentences.
The rule in two sentences
Every exercise in your program has a rep range (say, bench press, 4 sets of 6 to 8). First you progress REPS inside the range; the day you hit the top of the range on EVERY set while leaving 1 or 2 reps in reserve, you add weight next session and drop back to the bottom of the range.
That is it. New weight, build the reps back up, repeat. That is why it is called double progression: reps progress first, then load.
What “reps in reserve” means
Reps in reserve (RIR) is how many MORE reps you could have done with solid technique when your set ended. You finished your set of 8 and had 2 more in you: that is RIR 2.
Why not train to failure every time? Because the evidence says you do not need to: stopping 1 to 3 reps short of failure produces the same hypertrophy as grinding to failure, with meaningfully less fatigue (Refalo et al., 2024). Strength progresses well across a wide RIR range too (Robinson et al., 2024). Failure charges a lot and pays the same: bad trade.
This matters for the rule: the top of the range only counts if you got there with 1 or 2 in reserve. Eight agonizing half-reps are not eight reps.
How much to add
It depends on the lift:
- Lower-body and back compounds (squat, deadlift, leg press, rows): 2.5 to 5 kg, about 5%.
- Upper body and isolation work (presses, curls, extensions, raises): 1 to 2.5 kg, about 2.5%.
The logic is simple: bigger muscles tolerate bigger jumps. If your gym lacks small plates, the smallest available jump rules, and sometimes it pays to spend one extra session at the top of the range before taking it.
The day matters: do not add weight on a bad day
Here is the detail almost everyone ignores: the rule tells you WHEN you earned the jump, but the day to attempt it is decided by your recovery. Four hours of sleep, sore everywhere, foul mood: today is not record day, even if you hit the top of the range last session.
This is not a hunch either: methods that adjust load to how you show up that day (autoregulation) beat fixed percentages for strength gains in recent meta-analyses (Zhang et al., 2025). The practical version: on a good day, go for the jump; on a normal day, match last session; on a bad day, back off a little and protect technique. On a bad day, maintaining IS progress.
A worked example
Bench press, 4 sets of 6 to 8, at 60 kg:
- Week 1: 60 kg × 6, 6, 6, 6. Bottom of the range, normal.
- Week 2: 60 kg × 7, 7, 7, 6. Reps moved up.
- Week 3: 60 kg × 8, 8, 8, 8 with 2 in reserve. Top of the range on every set: you earned the jump.
- Week 4: 61.5 kg × 6, 6, 6, 6. New weight, bottom of the range, build again.
That kilo and a half looks tiny. It compounds to ~19 kg of progress per year on a single lift, without plateauing and without breaking yourself.
The three mistakes that kill progression
- Ego jumps: you hit the top on one set (not all of them) and add weight anyway. Result: half-reps with more load and zero extra stimulus.
- Not logging: without a record there is no rule to apply. If you do not know what you did last session, you are guessing.
- Jumping on a bad day: you earned the jump on Monday, show up wrecked on Thursday and try it anyway. The jump does not expire; save it for a good day.