Why guessing heavy on day one is the single most common way to sabotage this program before it starts, a real method for picking a conservative number instead, and what changes if you already have some lifting history.
Never done these lifts before? Read why starting light matters, then the actual method for finding a number. Why starting heavy backfires · The work-up method · For a true beginner
Returning to lifting, or coming from a different style? Your old numbers might not be the right numbers to enter here. Using a number you already have
This isn't generic caution. It's a direct consequence of how this program's novice progression actually works: every lift adds a small, fixed amount of weight every time you succeed, at roughly double the normal rate for as long as that lift has never once missed a rep. That faster rate exists to be spent while the bar is still trivially light for you — it's not a permanent feature, it's a head start that ends the instant you miss, even the very first time it happens, for any reason, on any set.
Start a lift at what turns out to be its real limit already, and that head start gets spent before you ever benefit from it. Miss a rep on day one and the lift's increment permanently drops to its slower, steady rate right there — not after a string of good sessions, immediately. Miss again at the same weight next time you train it — likely, if it really was already your limit — and that's two misses in a row, which is this program's own trigger to reset: the working weight drops to 90% of the number you were failing at, and you rebuild from there at the now-permanently-slower rate. Two full sessions in, you'd have no added weight, no banked fast-rate progress, and a rebuild from a number that was already too ambitious.
Start conservative instead and none of that math changes in your favor by accident — it changes because you arranged it to. A lift that keeps completing its sets keeps its faster rate the entire time, which means more sessions of free, guaranteed progress before you ever reach a weight that asks something real of you. Starting light costs nothing here; the program is built to find your real ceiling for you, session by session, far more reliably than a guess on day one can.
Starting too heavy is the single most common mistake, but it isn't the only one — see the common early mistakes page for the others, all of them habits rather than technique faults.
The same judgment call — find the number you're certain of, not the number that tests a limit — comes back on a meet platform as picking your opener attempt; see the first-meet logistics page for how that plays out on competition day specifically.
You don't need to test anything close to a real maximum to find a reasonable starting number — the approach below is the well-known method associated with Mark Rippetoe's Starting Strength, and it works because it never asks you to find out where you fail. It asks you to find where you're still comfortable, and stop there.
That stopping point does double duty: it's both the answer to "what's my starting weight" and, for that first session, your actual warm-up ramp — the same ascending, shrinking-effort climb described on the warm-up and plate math page, just run live instead of planned in advance because the destination wasn't known yet. Every session after this one, the app already knows your working weight and builds that same style of ramp for you automatically, without you having to repeat this process.
If you'd rather not make the judgment call at all, that's a completely valid choice, not a worse one: enter the empty bar as your starting number for every lift and let the faster early rate carry you up from there. This is this program's own enrollment default whenever a starting weight isn't supplied, precisely because a light start costs nothing and a heavy guess costs the reset cycle described above.
You have two equally legitimate options, not a right one and a wrong one. Either run the work-up above once per lift, the first time you train it, and enter whatever number you land on. Or skip the decision entirely, start every lift at the empty bar, and let the faster early-progression rate find your real number for you over your first handful of sessions — slower to reach a meaningful weight, but there's no judgment call to get wrong. Most true beginners are better served by leaning toward the second option: your real working weight this early is almost always lower than it feels like it should be, and this program's whole design assumes exactly that.
Learning the movement pattern matters more than the number on the bar for your first few sessions on any lift — see each lift's own technique page (squat, bench press, deadlift, overhead press) for what "right" actually looks like before you're loading it up.
If you have a real, current, comfortable working weight for a clean set of five on a lift — trained regularly and recently, not a number from years ago or a personal record you once touched for one hard rep — start there instead of at the empty bar. There's no benefit to re-discovering ground you already hold.
Two situations still call for the work-up method above rather than a number you already have in mind. Both are about picking a number here, on this page, before you're actively running the program — if you're already mid-program and the question is instead "I was off for several days to a couple of weeks, what do I do with the working weight I already have," that's a shorter layoff than either row below assumes, and coming back after time off on the sleep and recovery basics page answers that specific case instead.
| Your situation | Why your old number doesn’t transfer directly |
|---|---|
| Your recent training used a different rep range or lift variant | A heavy single, a high-rep bodybuilding set, high-bar or sumo work, or push-press strength doesn’t tell you what a comfortable set of five on this program’s specific lift feels like — the work-up finds that directly instead of asking you to guess a conversion. |
| It’s been many months or longer since you trained consistently | Detraining is real, but previously-trained lifters typically regain lost strength faster than a true beginner builds it the first time — starting meaningfully lighter than your old number costs you very little time and removes the guesswork entirely. |
One more real option if your training history is both current and heavy enough: this program supports enrolling directly into the advanced phase with a declared one-rep max for squat, bench press, and deadlift, skipping novice and intermediate entirely. That's a real fit for an experienced lifter with an actual training history to draw on, not a shortcut for anyone else — see the advanced phase's direct-entry section for exactly what it requires.
Two of the four lifts are worth calling out specifically, because the natural instinct — "press should be close to what I bench" or "deadlift and squat should be similar" — is measurably wrong for most lifters, not just a matter of individual variation. Real, aggregated lift data from Strength Level's strength standards makes the gap concrete: at a beginner level and a fixed bodyweight, overhead press standards run to roughly 60% of bench press, and deadlift standards run to roughly 15-20% above squat — a consistent pattern that holds at both their "beginner" and "novice" tiers, not a fluke of one data point.
Overhead press: expect a much lighter number than bench, and don't fight it. Pressing a bar overhead is a smaller-muscle, harder-leverage movement than pressing it off your chest, and the two lifts simply don't scale together the way bench and squat's general "bigger lift, bigger number" intuition suggests. If your press starting weight looks disproportionately light next to your bench number, that's very likely correct, not a sign you under-tested it.
Deadlift: it's normal to start heavier than your squat. Part of the reason is already built into this program's own design — deadlift's per-session increment (+10 lb / 5 kg) is double squat's (+5 lb / 2.5 kg), reflecting how much more total muscle mass the deadlift's hip-hinge pattern involves. The rest is about what the lift demands of a body that hasn't trained it yet: a deadlift is one controlled pull in one direction from a dead stop, with no eccentric phase to control on the way down and nothing to balance on your back or overhead — a squat and a press both ask for real-time balance and control through a full range of motion under load. Fewer coordination demands for an untrained lifter is a big part of why deadlift so often "feels" strong before any real training has happened to it specifically. Don't force your deadlift starting number down to match your squat just because they're often talked about together — let the work-up method above tell you the real number for each lift independently.
Say a work-up session on squat goes: empty bar for two sets of five, then 85 lb, then 110 lb for a triple, then 130 lb for a single — every one of those clean, controlled, and clearly not close to hard. The next jump would take real effort to judge honestly, so that's the stopping point: 145 lb becomes the number entered as the starting working weight, for three sets of five.
| Stage | Total (lb) | Plates per side | Reps |
|---|---|---|---|
| Warm-up | 45 (empty bar) | — | 5 × 2 sets |
| Warm-up | 85 | 10 + 10 | 5 |
| Warm-up | 110 | 2.5 + 5 + 25 | 3 |
| Warm-up | 130 | 2.5 + 5 + 35 | 1 |
| Working | 145 | 25 + 25 | 5 × 3 sets |
Notice this ramp has four rungs where the warm-up and plate math page's own 225 lb example needed five — exactly the rule described there: how many rungs a ramp gets scales with how far the weight has to travel from the bar, and 145 lb is a smaller jump than 225 lb. The work-up you just ran to find 145 lb, and the ramp the app builds to 145 lb every session afterward, are the same climb for the same reason — this page finds the destination once; that page explains the climb you'll repeat every time from here on.