Wilks is obsolete: what replaced it and why

A 60 kilogram lifter totalling 400 kilograms and a 120 kilogram lifter totalling 700 have both done something impressive. Deciding which is more impressive requires a formula, and which formula you use changes the answer.

Why coefficients exist at all

Strength scales with body size, but not proportionally. Muscle cross-sectional area grows roughly with the square of a linear dimension while body mass grows with the cube, so a lifter twice as heavy is not twice as strong. Absolute total therefore favours heavier lifters, and simple bodyweight ratios over-correct and favour the very light.

Scoring coefficients exist to sit between those two failures. Each is a curve fitted to competition data, mapping a total and a bodyweight to a single comparable number.

Because they are fits to data, they inherit the data. A formula built on results from one era and one set of federations will misrepresent lifters outside that sample — which is precisely what happened to the formula most calculators still default to.

The Wilks era

The Wilks coefficient, published in 1994, was the global standard for roughly a quarter-century. It was a genuine improvement on what preceded it and it made cross-class comparison routine in a way it had not been before. ³

Two facts about how it was built explain most of what followed. The polynomials were fitted to equipped results from 1987 to 1994 — meaning the reference data is squat suits and bench shirts, and it predates the raw lifting that dominates the sport today. And Wilks intended the formula to be revised every two to five years as the data moved. It never was. A rebalanced Wilks-2 finally appeared in March 2020, twenty-six years later, with entirely different coefficients. ²

So the drift people complain about is not a flaw in the maths. It is a curve fitted to one era of the sport, left in place across another.

It is also worth saying plainly that no major federation still uses it. The IPF scores on GL Points; USA Powerlifting scores on DOTS. Both moved from published lookup charts to formulas computed from exact bodyweight. Wilks survives in conversation and in calculators rather than in competition, which is a reasonable argument for knowing what it is and a poor one for steering by it.

One practical caveat if you are comparing against an official result. USA Powerlifting adjusts for age using two systems, on the assumption that maximal strength potential peaks between 24 and 39 and falls away either side: Foster coefficients for competitors aged 14 to 23, and McCulloch coefficients for those 40 and over. Both are the same for men and women, and both multiply the total.

The adjustment is substantial at the edges, and our calculator reports open-category scores only — so if you are under 24 or over 39, your official placing will differ from what you see here, by a wide margin at older ages.

Two things are worth knowing about how USA Powerlifting actually uses these numbers, because both are easy to get wrong.

Placings are decided by age division, not by age-adjusted score. The current rulebook awards the top finishers in each division, and divisions are defined by age — youth, teen, junior, open, and masters 1 through 7, running to 110 years old with no upper limit. Age coefficients are a best-lifter and cross-division comparison device, not the mechanism that decides who wins your class.

And DOTS is explicitly a placeholder. The rulebook records that USA Powerlifting approved developing a formula superior to DOTS at its 2024 meeting, with work running through 2025 and possibly beyond, and states that DOTS continues to be used until such a time as a new formula is approved. So the scale this calculator reports as current is, by its own governing body’s account, on its way out. That is the third replacement in this story, and it will not be the last.

There is a wrinkle worth knowing if you go looking for the age numbers yourself. The McCulloch values that circulate on calculator sites and forums — a 60-year-old multiplier of 1.340, a 75-year-old’s of 1.835 — match USA Powerlifting’s 2021 document and an older edition of the WRPF’s table. USAPL’s more recent document carries materially different and less generous figures: 1.194 at 60, against 1.340. That is roughly a twelve per cent difference in a lifter’s score.

We have not implemented either, and the reason is the discrepancy itself. Adopting the popular numbers would flatter masters lifters relative to their actual placings, and adopting the newer ones on the strength of a cached document would be guessing at which is in force. If you are a masters or junior lifter, get your coefficient from your federation rather than from any calculator, including ours.

Its weakness is at the extremes — though the picture is more particular than the shorthand suggests. The formal validation, published in 1999, found no systematic bias in the total for either sex, which is the number this calculator scores. What it did find was a bias against heavier lifters in the deadlift specifically, and a bias favouring intermediate weight classes in the women’s squat. ⁴ It also noted, in passing and rather remarkably for a formula already governing the sport, that Wilks was not based on published data.

That validation rested on 30 men and 27 women per lift — world record holders and the top two from the 1996 and 1997 World Championships. A narrow, elite sample, which is worth holding in mind before treating “validated” as settling much.

Analyses over the years found that the formula does not treat very light and very heavy lifters even-handedly, and as the sport grew — with more lifters at both ends of the weight range and far more recorded data — the mismatch became harder to ignore.

This is not a scandal. It is the ordinary lifecycle of a statistical fit: it described its data well, the data changed, and it needed refitting.

DOTS

DOTS was published in 2019 by Tim Konertz, built on modern competition data with the explicit aim of correcting the bias at the extremes. USPA, WRPF and several other federations adopted it from 2020, and it has become the default in much of the sport outside the IPF.

Structurally it resembles Wilks — a fourth-order polynomial in bodyweight, scaled so the resulting numbers land in a familiar range. That similarity is deliberate and useful: at mid bodyweights DOTS and Wilks produce very similar scores, so lifters around 80 to 90 kilograms barely notice the change. The divergence appears where it should, at the light and heavy ends.

IPF GL Points

The International Powerlifting Federation took a different route. IPF GL Points — Goodlift points — became the federation’s official system on 1 May 2020, replacing the earlier IPF Points formula.

Two things distinguish it. First, the mathematical form is different: an exponential rather than a polynomial. Second, it uses separate coefficient sets for each combination of sex, equipment category and event, so classic and equipped lifting are scored on their own curves rather than sharing one.

The scale is also different, and this catches people out. IPF GL points are calibrated so an elite result approaches 100, while DOTS and Wilks both run in the hundreds. A GL of 72 and a DOTS of 352 can describe exactly the same performance. Comparing a number from one system against another is meaningless.

The bottom line

If you compete under the IPF, use IPF GL, because that is what your federation scores you on. If you compete elsewhere, or you are simply tracking your own progress, DOTS is the better modern default.

Keep Wilks in view for one reason only: a great many older records, gym leaderboards and personal benchmarks are still expressed in it, and you will want to compare like with like. Our calculator shows all three at once for exactly that reason — with Wilks labelled legacy, rather than presented as the standard the way most incumbent tools still do.