How the estimate works
Every formula above does the same thing: it takes a set that went close to failure and extrapolates back to a single rep. They all agree that a weight you can move five times is roughly 85 to 90 percent of what you could move once, and they disagree about the exact figure because each one was fitted to a different group of lifters on a different exercise.
At 100 kg for 5 reps the seven land between 112.5 kg and 119 kg, which the tool shows as 112.5 to 120 because it rounds every weight to something a gym can actually load. That 6.5 kg spread is the honest error bar on the whole method. It is small enough to program from and large enough that treating any single formula as the truth is a mistake.
The spread widens fast as reps rise. At 5 reps the seven sit within 6.5 kg of each other; at 12 reps they span 128 kg to 144 kg. That is why the reps field is capped and why the tool warns you above ten.
Which formula to use
Use the average, and program off the low end. The tool shows both. If the seven say 112.5 to 119, write the week off 112.5 and let the client beat it; the cost of a percentage that is four percent too heavy is a missed set and a dented client, and the cost of one that is four percent too light is nothing.
If you want to pick one, the ordering is predictable rather than mysterious. Brzycki reads lowest from two to five reps. From six to ten reps O'Conner is the lowest. Mayhew reads highest through the low reps and Wathan takes over around eight. Above ten reps Lombardi collapses toward the original weight while Brzycki and Lander run away upward, which is another way of saying nobody should be estimating a max off a set of fifteen.
A single rep is not estimated at all. If the client did one rep, that weight is the max, so the tool returns it unchanged rather than running it through an equation that would report 103 kg for a 100 kg single.
Turning the number into a week
The estimate is only useful because it makes percentages loadable. The table above is the NSCA training load chart with the arithmetic already done: 5 x 3 at 85 percent stops being a note in your phone and becomes a weight, a rep count and the plates for each side.
A worked week off a 115 kg estimated squat. Heavy day: 5 x 3 at 87 percent, which is 100 kg. Volume day: 4 x 6 at 75 percent, which is 87.5 kg. Third session, technique: 3 x 5 at 70 percent, which is 80 kg. Every number rounds to something the gym can actually load, because a prescription the client cannot put on the bar is not a prescription.
Percentages drift out of date faster than trainers expect. A beginner adding 2.5 kg a session is 10 kg stronger inside a month, so a table built off a max from six weeks ago is prescribing a warm-up as a working set.
The warm-up ramp
The ramp above is the one most strength coaches write by hand: the empty bar for eight, then roughly 40, 60 and 80 percent of the working weight for five, three and two, then the first working set. Reps come down as the weight goes up, so nothing in the ramp costs the client a working set.
Heavier work wants more rungs, not more reps. A client working at 90 percent needs five or six ramp sets; a client doing three sets of twelve on a leg curl needs one. Rest between the ramp sets is short, about 60 to 90 seconds, until the last two, which want the same rest as the working sets.
Retesting without maxing out
Most clients never need to attempt a true single. The estimate is accurate enough, and testing a real max costs a session, adds risk on the lifts where form breaks first, and tells you something you could have inferred.
Retest by repeating the same set. Take the exercise the client is being programmed on, work up to a set of three to five at an effort of about RPE 8, log the weight and reps, and put those two numbers back into this calculator. If the estimate moved, the percentages move with it. Every six to eight weeks, or at the end of a block, is enough.
Keep the conditions the same or you are measuring the conditions. Same exercise, same bar, same rough time of day, same position in the session. A max estimated from a fresh first set and one estimated after twenty sets of legs are not comparable numbers.
Where the estimate breaks
The formulas were built on barbell lifts, mostly the bench press and the squat. A dumbbell press estimates per hand and the two sides rarely match. Machines with a stack in fixed increments cannot be loaded to a percentage anyway. On those, prescribe by reps and effort rather than pretending the arithmetic transfers.
It also assumes the set genuinely went close to failure. A set of eight stopped at RPE 6 is a set of eight with three left in reserve, and running it through a formula reports a max the client does not have. If you are not sure how hard the set was, ask how many more they could have done before you use the number. If you know the RPE, the RPE calculator accounts for the reps left in reserve.
Technique is the other limit. A max estimated from reps that were rounding and rebounding is a number about the rounding, not about the muscle.
Keeping the number current
A calculator gives you one estimate for one lift on one day, and then you own the problem of keeping it up to date across every client you coach.
Onyx Coach stores a 1RM per exercise per client, and you write sets as a percentage of it. "5 x 3 at 85%" fills in the kilos from that client's own number, so when you update the max after a good session, every prescribed load in the block moves with it and you do not retype a single weight. The client logs weight, reps, RPE and rest per set from their phone in the gym, so the set you would put back into this calculator is already in their history.
