Warm-up sets and plate math, in plain language

Why this program ramps up instead of jumping straight to your top set, the actual ramp this app computes, and how to load a bar correctly by hand — in pounds or kilos, down to the fractional plate.

Never planned a warm-up or loaded a bar off a written number before? Start with why this program ramps up at all, then the ramp itself. Why the ramp exists · The ramp itself · How long to actually rest · How long a session takes, and what to cut

Just need to load the bar for a specific number right now? Skip straight to the arithmetic. Plate math · The fractional-plate case

Why this program ramps up instead of jumping straight to your top set

Everything below assumes you already have a real working weight to ramp up to. If you don't yet, see finding your starting numbers first — this page picks up once that number exists. Skipping or rushing the ramp below is one of several genuinely common early mistakes; see the common early mistakes page for the others.

"Get the blood flowing" is the usual answer, and it undersells what's actually going on here. Raising muscle temperature does make a muscle contract and relax faster, which is real, but the broader claim — that a generic warm-up prevents injury — is genuinely contested; controlled studies on it disagree. That's not the reason this program builds a real ramp for every session.

If you also do dedicated mobility work from the drills on the mobility page, the timing of that work matters too, in a way that's easy to get backward — see when to actually do that work before stacking a long static stretch in front of the ramp below.

The actual reason is specific to what this program asks of you: every working set is a small number of reps at a real percentage of your best effort — squat and bench work sets of five on novice, and by the time the advanced phase's taper arrives, the second of its two taper sessions is a single rep at 100% of your training max, with no lighter reps in the set to "settle in" on the way there. A twelve-rep bodybuilding set warms itself up — the first few reps of that set are submaximal by definition. A top set of five, three, or one has no such reps to spare; the first rep at your working weight is also the first time today your body has tried to move that much weight, and it has to be right immediately, not eventually.

The ramp is where that gets rehearsed instead of gambled on. Each ascending set confirms your setup, bar path, and bracing still feel right at a heavier load than the last one, and lets your nervous system recruit progressively more muscle rather than being asked for a near-maximal contraction cold. Skip the ramp on a near-max day and you've combined two unknowns into one all-or-nothing rep: whether today's technique is actually on, and whether the weight is actually manageable. The ramp is what separates those two questions and answers the first one before it costs you the second.

The ramp this app actually computes

This isn't a generic percentage chart applied the same way to every lift. The app's warm-up engine treats squat, bench press, and overhead press as one family and builds a short empty-bar block, then a graded climb of ascending sets with the rep count shrinking as the weight climbs, landing the last warm-up rung at ninety percent of your working weight — rounded to a normal-feeling number, then matched to the nearest total your actual rack can build. How many rungs the ramp gets scales with how far the weight has to travel: a small jump from the bar gets a short ramp, a big one gets a longer one, within a floor of three sets and a ceiling of seven.

Run the exact same lift, engine, and logic against a kilo rack and a kilo working weight and the result isn't a unit conversion of the table above — it's a separately optimized ramp for the plates a kilo rack actually has, fractional plate included where it happens to be the best fit:

Any rung you don't like isn't fixed — tap it and change the weight or the rep count, and the app snaps your override to the nearest total your own plates can build, recomputing every plate change downstream of it. A bar-only set is the one thing you can't edit, since there's nothing to load differently. If a target weight genuinely can't be built exactly from your plates — a percentage-based prescription lands between two loadable totals more often than you'd think, more on that below — the app tells you the nearest weight on either side rather than silently rounding and letting you find out on the platform.

This ramp, and deadlift's own version just below, are built for squat, bench press, overhead press, and deadlift specifically. Neither one extends to a chin-up, a row, or anything else you might add on top — see warming up for accessory work for the real, code-level answer on whether that's a gap.

Deadlift warms up differently, on purpose

The deadlift starts on the floor, not out of a rack, and this program's ramp treats that as a real difference, not a detail: there's no empty-bar block at all. The first deadlift warm-up set of the day is already loaded — the heaviest total your plates can build at or under half your working weight, preferring a real full-size plate on each side over smaller ones stacked up, since a full-size plate is what actually sets the bar at the same height off the floor you'll pull it from at the top. A bar with no plates on it doesn't rehearse that; a loaded bar does, even a light one.

The rep counts also stay closer to your working reps the whole way up instead of starting high and shrinking fast, matching the general coaching consensus for this lift specifically: warm up deadlift with rep counts similar to what you're about to pull, not a heavy single built on top of a set of ten. Run the same 140 kg working weight through a kilo rack and the app produces its own independent ramp the same way — 70, 95, 112.5, and 125 kg on the way to a 140 kg working set — again not a conversion of the pounds numbers, a separate real optimization for a kilo bar and kilo plates.

Plate math: what actually goes on the bar

The arithmetic behind every table above is the same arithmetic you'd do by hand, and it's worth knowing even if the app always does it for you. Start from the total weight you want on the bar, subtract the bar itself (and your collars, if you use any that add real weight — most clip collars don't), and divide what's left by two — that's what goes on each side, not the whole remaining weight on one side.

1. Subtract the bar (and collars) from your target total. — A standard bar is 45 lb / 20 kg. Most training collars add close to nothing — check yours if you use real weighted ones.
2. Divide what’s left by two — that’s your per-side target. — This is the number the rest of the math works from, not the total weight again.
3. Load the biggest plate that still fits under your per-side target, then repeat with what’s left. — This gets you to the same total the app would build, almost always with the same number of plates per side — it’s also just the fastest way to load a bar by hand.
4. Stop when nothing’s left, or when the smallest plate you own is bigger than what’s left. — A gap you can’t close with the plates you have is the fractional-plate case below, not a mistake in the math.

Worked example, standard pounds rack, no collars: a 225 lb squat target. 225 minus the 45 lb bar leaves 180, divided by two is 90 lb per side. The biggest plate that fits under 90 is 45 — load it, 45 left. The biggest plate that still fits under 45 is 45 again — load it, nothing left. Two 45s per side, bar included, is 225 — matching the working set in the first table above exactly, because it's the same search the app runs.

When your target doesn’t divide evenly

Standard plates bottom out at 2.5 lb or 1.25 kg — real jumps, and on their own they set a floor for the smallest total change you can make on the bar: twice your smallest plate, 5 lb or 2.5 kg. Most of this program's own increments clear that floor comfortably. Squat and deadlift's steady novice jump is 5 lb / 2.5 kg — per side, 2.5 lb or 1.25 kg, exactly one standard small plate, no fractional plate required.

Two real places in this program don't clear that floor. Overhead press's own jump steps down to 2.5 lb / 1.25 kg the moment it resets for the first time, and intermediate's own Test Day increment for bench and press is the same 2.5 lb / 1.25 kg. Both are smaller than a standard plate's own minimum jump. Per side, that's 1.25 lb — which is exactly one plate from the fractional plate set on the equipment page, the 1.25 lb pair. Without it, the only way to hit this program's own prescribed number is to round up to the nearest 2.5 lb plate and quietly take a bigger jump than the program actually called for.

A kilo lifter hits a sharper version of the same wall here: this specific increment (1.25 kg total, so 0.625 kg per side) is finer than even a standard rack's own smallest 1.25 kg plate can resolve on its own — true microloading, finer than a quarter-kilo, is what actually closes that specific gap on a kilo bar.

The same gap shows up from the other direction on a percentage-based prescription, not just a fixed increment. Say a training max of 275 lb feeds the advanced phase's own Intensification heavy day, prescribed at 88% of training max — 242 lb exactly. On a standard rack, that's simply not buildable: the nearest totals are 240 lb below it and 245 lb above it, and the app will show you both rather than pick one silently. Add one 1 lb plate per side to what a standard rack already builds at 240 (35 + 35 + 25 + 2.5 per side) and the total lands on 242 exactly — the same Micro Gainz set closing a gap a flat percentage calculation opened, not just a small fixed increment.

How long to actually rest between working sets

"Rest until you are ready, two to five minutes" is the one line this program gives for the actual pause between working sets, and nothing on this site has explained where that range comes from or when to lean toward one end of it rather than the other — worth fixing the same way the ramp above already gets a real explanation instead of a bare instruction.

Put together, that's a real answer for why this program's own instruction is a range rather than one fixed number, and for which end of it to actually lean toward. A true beginner's very first sessions — the empty bar, or a weight still trivially light — sit closest to Grgic et al.'s own "untrained" finding above, so there's no real cost to resting nearer the shorter end of the range while a set is still comfortable. As a lift gets heavier and genuinely tests you — every set later in novice, and essentially every Test Day or Heavy day set from intermediate onward — the same review's own "resistance-trained" finding takes over, and the honest answer moves toward the fuller 3 to 5 minutes instead. That's a real, checkable case where a source still can't be over-read: Grgic et al.'s own review is explicit that resting under 60 seconds still produces real, "robust" strength gains — cutting rest short isn't ruinous, it's a real, measured cost, not a total loss.

Resolve's own built-in rest timer already reflects this range directly, not just in this page's prose: its presets run 1, 1.5, 2, 3, and 5 minutes, and a fresh timer opens at 3 minutes by default — the middle of the range the research above actually supports for a set like this program's own, not a number picked for looking round.

How long a real session actually takes, and what to cut first when you’re short on time

Everything needed to answer "will this actually fit in my lunch break" already exists on this page and inside this app: the real sets and reps each Day A and Day B prescribes, the real warm-up ramp the engine above computes, and the real research on rest length just above this section. Nobody had added them up into an actual number, or said what to trim first once the honest answer is "not quite" — worth fixing the same way the bare "two to five minutes" line got fixed above.

Take a typical novice Day A: squat, bench press, and deadlift, 3 × 5 on squat and bench, 1 × 5 on deadlift — seven working sets total, every single Day A, per the novice phase page. Each lift gets its own ramp first, sized by how far its working weight has to travel, inside the floor-of-three, ceiling-of-seven rungs the ramp section above already states. Squat's own worked example above, at 225 lb, needed six warm-up sets before its first working rep; deadlift's own worked example, at 315 lb, needed five. Bench press runs the same squat/press-family ramp logic, but a novice's bench working weight usually sits much closer to the empty bar than squat's own — a proportionally shorter jump for the same ramp logic to size — so a typical early-to-mid-novice bench ramp lands close to the ramp's own floor: two empty-bar sets plus three more rungs, five warm-up sets in a representative case.

Twenty-three sets means twenty-two gaps between them, and they aren't the same kind of gap. Only four sit between two real working sets of the same lift — two on squat, two on bench, none on deadlift, since one working set has nothing to rest between — and those four are exactly what the research cited above actually measured. The other eighteen are warm-up-to-warm-up, the last warm-up into the first working set, or the walk from one lift's station to the next. None of the three studies above say anything about that kind of gap, because a warm-up set is submaximal by construction — the entire reason the ramp exists in the first place, per the very first section on this page — so there's no finding that moving briskly through it costs anything a study has actually measured.

Put a real number on both kinds of gap and the totals tell a clear story. The eighteen warm-up-and-transition gaps need enough time to actually change plates and get set up again, not to fully recover — call it 40 seconds each in practice, a practical convention here, not a citation — about 12 minutes total, the same either way. The reps themselves add relatively little on top of that: twenty-three sets of five reps or fewer, even at a generous 20 seconds of actual bar time each, comes to about eight minutes total. The real variable is the four working-set gaps, and it's the exact range this page already covers above. A true beginner, still sitting at Grgic et al.'s own "untrained" end near 90 seconds, adds about six minutes there — call it roughly 25 minutes for a light early session, start to finish. A lifter far enough in that sets are genuinely heavy, leaning toward the fuller three-to-five-minute end the same review recommends once you're resistance-trained, adds sixteen to twenty minutes instead — call it roughly 40 for a heavy, later novice session. Same three lifts, same seven working sets, and rest is the entire reason the clock time nearly doubles. Day B follows the same arithmetic: overhead press's own novice working weight typically runs even lighter than bench's (see the fractional-plates section above), never heavier, so it doesn't push the total up.

That breakdown is also the honest floor a disciplined lifter running a timer actually gets, not an average of how a real session goes — chatting, waiting for a rack, or checking a phone between sets all add real time on top, and nothing above accounts for any of it. It's still the right place to start if the real question is what to cut, because it shows exactly where the minutes are actually going: not evenly, and not where intuition usually points first.

That second row doesn't apply to a genuine novice, and it's worth saying so directly rather than implying a cut that doesn't exist: the novice session generates exactly its three prescribed lifts and nothing else, so there's no accessory volume sitting on a novice's own session to trim in the first place — see accessory work for why. For a time-crunched novice, the first two real levers are the warm-up-and-transition rest above and, only while a session's own working sets are still light, the short end of the working-set rest range — in that order, not the reverse.