How much weight to start with on a new exercise is the one question in this collection with no formula behind it — and the honest answer is that no app can give you a number until you have lifted something. Zyra Workouts does not pretend otherwise: an exercise with no history shows no history where the recommendation would be. This page explains why, and how to run the first session so the second one has a number.
Why no formula can tell you
Every progression rule in the app is a function of your last logged set. The double progression engine reads the weight and reps you did, compares them with the target, and returns the next load. The estimated 1RM is computed from a set you completed. Both need a set to exist first, and a set needs a weight. The loop has to be started by hand.
Could a starting weight be guessed from something else — body weight, another lift, age? Only badly. The spread between two people of the same body weight on the same exercise is enormous, and the spread between two exercises for the same person is not much smaller: a lifter's row and their overhead press bear no fixed ratio. Any number printed on a screen would carry a false precision the app cannot back up. The source code is explicit about the choice: *No history → no recommendation.*
What the app can say is the shape of a sensible first session, and that is what its starting-weight guidance says in words rather than kilograms: pick a conservative weight that lets you hit the target reps at the prescribed RIR, then iterate.
In plain terms: the first weight is your guess; everything after it is the app's arithmetic on that guess.
The conservative first session
The goal of the first session is not a good set. It is a good *log entry* — one honest data point the engine can work from. That changes what "the right weight" means.
| If you start… | First set looks like | What the engine does next time | Cost |
|---|---|---|---|
| clearly light | top of the range with 3–4 reps to spare | above the ceiling → +1 step, reps to repMin | one step per session until the reps say stop |
| about right | inside the range at the target RIR | hold; climb the reps from here | none |
| too heavy | below the bottom of the range | back-off −1 step, reps to repMin | a missed range and a set that taught little |
The asymmetry is the whole argument. Starting light costs you a session or two of stepping up — 2.5 kg at a time on a barbell, 2 kg on a dumbbell or stack — while the reps do the talking. Starting heavy costs a back-off, and it can cost more than that if the exercise is new to your technique. A light first set is also the set on which your RIR rating is most likely to be honest, because nothing about it is a grind.
A practical way to choose the guess, with no numbers attached because none would be true for you:
- Take the empty bar, the lightest dumbbell or the first plate on the stack and do the movement for the top of the rep range. This is a form set; it also tells you roughly how far the real weight is from here.
- Add load in the equipment's own step until a set of the top-of-range reps feels like the RIR target plus about two — if the program says RIR 2, stop when a set feels like RIR 4.
- Log that set as a working set, with the weight, full reps and the RIR you actually felt. Warm-up sets are excluded from the engine's view; the working set is what it will read.
One session is enough. The next time the exercise comes up, the engine has a last set to compare with the target and shows a recommendation — with a blue wand if it is a hold or a step up.
In plain terms: find a weight that is honestly a bit too easy, do the set properly, and write down what it was.
Let progression find the number
From the second session on, the app's rule does the searching, and it searches quickly. Suppose the target is 8–12 @ RIR 2 (ceiling round(10) + 2 = 12) on a barbell with 1.25 kg plates:
- Session 1: you guess
40 kgand get 14 easy reps → above the ceiling → recommendation42.5 kg × 8. - Session 2:
42.5 kg × 13→ still above →45 kg × 8. - Session 3:
45 kg × 11 @ RIR 2→ inside the range → hold at45 kg; climb the reps.
Three sessions from a deliberately light guess to a weight that sits inside the range at the target effort, with every step earned by reps you actually did. The e1RM series over those sessions (40 × 14 → Epley 58.7 kg; 42.5 × 13 → 60.9 kg; 45 × 11 → Brzycki 62.3 kg) rises gently and honestly, and the plateau logic on the dashboard is not confused by a wild first number.
Two situations where the search needs a hand. If the equipment increment is large relative to the load — a dumbbell rack that jumps 10 → 12.5 → 15 kg — the app's 2 kg step will sometimes name a weight your rack does not have; take the nearest one and log what you lifted. And if you have real history on a close variant (a dumbbell bench press when you know your barbell bench press), that is a better prior than a bare guess — but it is still a guess, and the light-first rule still applies.
This is a method, not a trial result: it describes how the app behaves on a new exercise and the cheapest way to feed it a first data point.
The short version
No formula picks a first weight, and Zyra Workouts shows "no history" rather than inventing one. Start clearly light — top of the rep range with reps to spare — log the set honestly with its RIR, and let the progression rule step the load up in equipment-sized increments over the next few sessions. Light costs a couple of steps; heavy costs a back-off. Zyra Workouts is in development and not yet on the App Store; the rules quoted here are read from its source code and may change before release. Training guidance, never medical advice. Educational overview only — not medical advice.