Worked example: 405 lb for 5 reps
The seven formulas return 456, 456, 461, 472, 473, 476 and 482 lb. Median 472 lb, spread 26 lb. On the deadlift specifically, treat that as a floor rather than a centre — see below.
Why deadlift estimates run low
Deadlift rep sets degrade faster than any other lift's, and not because the prime movers give out. Grip fails, the lower back fatigues, and each rep starts from a dead stop with no stretch reflex to help. A set of ten deadlifts therefore represents a smaller fraction of your true single than a set of ten bench presses does — which means the formulas, fitted mostly on bench data, tend to under-estimate a deadlift max.
The practical version: the higher the rep count you entered, the more likely your real single is above the number on this page. From sets of two or three the estimate is reasonable. From a set of ten, expect the truth to sit above the top of the range.
Touch-and-go versus dead stop matters too. Bouncing reps off the floor produces a higher rep count from the same strength, and therefore an inflated estimate. If your sets are touch-and-go and your single would be from a dead stop, take the low end.
Why rep sets under-state a deadlift single
Of the four main lifts, the deadlift is the one where a rep set tells you least about your single, and it is worth understanding why rather than just applying a correction.
A set of ten deadlifts usually ends because your grip failed, or because your lower back could no longer hold position, or because you ran out of breath between reps. None of those is the prime movers giving out. A single is over in three seconds and none of those limits has time to arrive. So the rep set is measuring a different thing, and the seven regressions — mostly fitted on bench press data, where the failure mode is far cleaner — do not know that.
The practical rule: use sets of five or fewer for a deadlift estimate. If all you have is a set of ten, expect the true single to sit above the median this page returns, and treat the top of the range as the more likely figure rather than the middle of it.
Methodology
The same seven published regressions used across this site — Epley, Brzycki, Lombardi, O'Conner, Wathan, Lander and Mayhew — with the median reported and the full range shown. Details and the source of each equation are on how these calculators work.
Limitations, stated plainly
None of these equations was fitted on deadlift data specifically, which is stated here rather than buried. The estimate is a starting point for programming percentages, not a number to attempt cold. Warm up thoroughly and stop a set when your back position changes, not when the bar stops moving.
Touch-and-go and dead-stop sets are also not the same set, and no formula can see which one you did. A touch-and-go set of five uses the stretch reflex on every rep after the first and is worth meaningfully more reps than the same weight from a dead stop. If you want an estimate you can programme from, reset the bar between reps — that is the version which resembles the lift you are estimating.
Attempting a maximal deadlift is where a large share of serious gym injuries happen, and it is the lift with the least margin for a technical error under load. The estimate exists precisely so that you do not have to test one. This is a fitness estimate, not medical advice, and not a substitute for a professional who can actually see you.