Testing a true one-rep max tells you exactly one thing: what you can lift once, on that day, with fresh legs, a good warm-up, and enough psych-up to grind out a single ugly rep. It does not tell you much about the weeks of training that got you there, and it is a poor way to track progress week to week. True maxes are rare events by design. They carry a real fall-off in form under fatigue, they reward technique and nervous-system readiness as much as raw muscle, and testing one every week just burns your best training days finding a number instead of building the strength that number is supposed to reflect.

Estimated 1RM (e1RM) is the practical answer. Instead of testing a max, you estimate what it would be from a normal working set, the kind you're already doing at the end of a hard week. Do 5 reps at 100 kg for your top set, and the app can back-calculate roughly what you could do for one rep, without you actually attempting one.

The formula, and why there are two

Zirv uses two formulas depending on rep range, because a single formula does not hold up across the whole spectrum:

1 rep: no formula needed. The lifted weight IS the 1RM.
2 to 10 reps: the Brzycki formula, 1RM = weight x (36 / (37 - reps)).
11+ reps: the Epley formula, 1RM = weight x (1 + reps / 30).

Brzycki is more accurate at low reps, which is where most people actually test strength. But Brzycki has a known flaw: past around 10 reps, it starts overestimating badly, projecting maxes nobody could actually lift. Epley holds up better at higher reps, so Zirv switches formulas there. The switch is seamless because the two formulas produce identical results at exactly 10 reps, so there's no jump in the chart when a set happens to land right on that boundary.

Worked example: 100 kg for 5 reps gives 100 x (36 / 32), which is 112.5 kg estimated. 100 kg for 3 reps gives about 105.9 kg. Same weight, fewer reps, lower estimate, which matches what your body already knows: 3 hard reps is a smaller effort than 5.

How this catches PRs you would otherwise miss

A personal record is not just "the heaviest weight you've ever moved." Zirv compares e1RM, not raw weight, so a new lift beats your old PR if its estimated 1RM is higher, even if the weight on the bar is lower. That means 100 kg for 3 reps beats an old record of 100 kg for 1 rep, because 3 clean reps at that weight represents more strength than one. If you only tracked raw weight lifted, that set would look identical to your last max and the improvement would go unlogged.

For bodyweight exercises where there is no external load to plug into a formula, pull-ups being the obvious case, Zirv compares by reps instead: more reps is the record. And if you later add weight to that same movement, a weighted record beats a bodyweight-only one, since adding load is a strictly harder version of the same lift.

Read the trend, not the point

The value of e1RM is not any single number, it's what happens when you plot it across a training block. A single reading is noisy: sleep, stress, and how warmed up you happened to be can all shift a set by a few kilos either way. Zirv charts the e1RM trend over time, both as a sparkline on the dashboard and as a longer progression chart per exercise, so what you're actually looking at is the shape of the line, not one data point on it. A bench press that's been trending up for six weeks tells you something real about your training. A single top set that's a few kilos higher than last time might just be a good day.

This is also how e1RM connects back to plan design. Weight targets in a training plan can be derived from percentages of estimated 1RM rather than fixed numbers, so as your e1RM moves, the working weights the plan calls for can move with it, instead of staying pinned to a number from months ago.

The honest limitations

These formulas are population averages, not laws of physics. They were built by fitting curves to lifters in general, and any individual can deviate from that curve based on limb length, technique, and which specific lift is being tested. They tend to be least reliable at the extremes: very low reps where a formula barely matters anyway, and very high reps where small errors in rep-counting or form get magnified. And a formula tuned on a barbell squat does not necessarily transfer cleanly to something like a cable row.

None of that makes e1RM useless. It makes it what it's designed to be: a good enough estimate, repeated often, that turns into a real trend line. That's more useful for tracking progressive overload than an occasional true max ever was, because it's built from the sets you were already doing, not a separate event you have to schedule around.

The takeaway

A true 1RM test is a rare, high-fatigue event that tells you about one day. Estimated 1RM is a number you get for free from every hard working set, accurate enough in the 2-to-10-rep range where most training happens, and precise enough at the boundary to switch formulas without anyone noticing. Track the trend it produces, not any single reading, and you get a strength signal that updates every session instead of every few months.