The job no raw total can do
Put a 60kg lifter and a 120kg lifter next to each other with the same 250kg total, and everyone in the room knows the smaller lifter did something more impressive. Nobody can quite say by how much. That single unanswered question is the reason powerlifting has spent fifty years building, discarding, and rebuilding formulas that turn a total into a bodyweight-adjusted score.
The math has to fight biology. Bigger bodies carry more muscle, so heavier lifters move more weight in absolute terms. But strength does not climb in a straight line with mass. A body twice as heavy isn't twice as strong, because muscle force tracks cross-sectional area while bodyweight tracks volume. Geometric scaling puts absolute strength at roughly the two-thirds power of bodyweight, an observation that goes back to studies of lifting records in the 1950s. Any fair scoring system has to bend that curve back flat, so a featherweight and a superheavyweight land on the same ruler.
Getting that bend right is hard, and every attempt has been an argument about the shape of one line.
Before Wilks, federations each picked their own curve
The 1970s and 1980s were a patchwork. American federations leaned on the Schwartz formula for men and the Malone formula for women, two separate curves stitched together by sex. Elsewhere the O'Carroll formula and the Siff formula had their own followings. Some meets used the Glossbrenner formula, which was itself a blended average of two earlier systems because no single one satisfied everyone.
The practical result was chaos for anyone trying to compare across meets. A lifter's "best-lifter" placing could flip depending on which federation was scoring, because the formulas disagreed most exactly where it mattered, at the light and heavy ends of the field. Two athletes could each walk away convinced the math had robbed them, and both could be pointing at a real flaw in whichever curve had been used that day.
What Wilks standardized in 1994
Robert Wilks, then running Powerlifting Australia, published a coefficient set in 1994 that consolidated the mess into one formula per sex. The structure was a fifth-order polynomial: take your bodyweight in kilograms, run it through a five-term equation to produce a coefficient, and multiply your total by it. One curve for men, one for women, applied everywhere the same way.
The Wilks coefficient caught on fast and became the International Powerlifting Federation's official scoring method for years. Part of why it stuck was that it held up under scrutiny. A 1999 validation study by Vanderburgh and Batterham tested whether the Wilks score still favored certain bodyweights and found it carried less bodyweight bias than the formulas it replaced. For the first time the sport had a single, defensible, internationally shared number.
That number quietly became the culture. "What's your Wilks" turned into gym shorthand for relative strength, printed on meet sheets and argued over in forums. For a couple of decades it looked like a solved problem.
Why one formula was never going to be the end of it
A scoring curve is only as good as the data it was fit to, and the data kept growing. By the late 2010s there were far more meet results on record than existed in 1994, and the larger sample showed the Wilks curve was not perfectly neutral after all. It slightly favored lifters in the middle of the bodyweight range and was harsher at the tails, the same weakness the older formulas had, just smaller.
Two responses arrived almost at once, and they went in different directions.
The IPF built its own replacement. In 2019 it stopped scoring its own championships with Wilks and switched to a points system of its own design, refined into IPF GL Points the following year. Independently, a formula called DOTS emerged around 2019, fit to a broad modern dataset, and got picked up by USAPL and USPA as their scoring standard. Robert Wilks himself was not idle either, releasing an updated coefficient set, often called Wilks 2020, to correct the drift in his original.
So within about two years the sport went from one shared formula to four live ones, each claiming to be the fairer bend of the same line.
Three curves, not three versions of one
The tempting way to read this history is as a version number ticking up, Wilks 1.0 to Wilks 2.0 to DOTS to IPF GL. That framing is wrong, and it hides the interesting part. These systems do not share a shape. They are genuinely different mathematical animals fit to the same problem.
| System | Introduced | Curve shape | Primary adopters |
|---|---|---|---|
| Wilks | 1994 | Fifth-order polynomial | Historical IPF standard |
| Wilks 2020 | 2020 | Re-fit fifth-order polynomial | Powerlifting Australia |
| DOTS | ~2019 | Fourth-order polynomial | USAPL, USPA |
| IPF GL Points | 2020 | Saturating exponential | IPF |
Wilks and its 2020 update both run a five-term polynomial of your bodyweight. DOTS trims that to a four-term polynomial fit to newer data, which is why it and Wilks return numbers that are close but never identical for the same lifter. IPF GL Points throws the polynomial out entirely. Its curve is an exponential that saturates, flattening as bodyweight climbs, so it treats the heavyweight end of the field with a fundamentally different logic than a polynomial does.
That is why asking "which one is the newer Wilks" misses the point. DOTS is a leaner polynomial. IPF GL is not a polynomial at all. They disagree not because one is more up to date, but because their designers made different bets about the true shape of strength across bodyweight, and there's no single fact that settles which bet is right.
Where every version still struggles
For all the rebuilding, one weakness survives every formula: the bodyweight extremes. A curve is only as trustworthy as the data that anchors it, and there are far fewer 52kg and 140kg lifters in the record books than there are 80kg lifters. Out at those tails the curves are least constrained, so small design choices swing the coefficient more than they should. A polynomial in particular can wiggle in regions where the data is thin, producing a score that looks precise to a decimal place but rests on very little.
This is the practical trap for anyone reading a score. The number arrives with a decimal, which reads as authority, but that last digit means far more for an 80kg lifter sitting in the dense middle of the data than for a 145kg lifter out where the curve is guessing. It is also a large part of why the IPF reached for a saturating exponential and why DOTS was refit at all: both were attempts to make the heavy end behave. Neither fully solved it.
Watch what happens to a single superheavyweight. Take a 145kg lifter with a 900kg total. Run that through the Wilks curve and you get a coefficient near the flat, guessing region of a fifth-order polynomial, where a few kilograms of bodyweight barely nudge the result. Run the same numbers through DOTS, a shorter polynomial fit to different data, and through IPF GL's exponential, which is deliberately engineered to keep flattening past 120kg, and the three systems can place that lifter several points apart from one another. The lifter didn't change. The total didn't change. Only the assumption about how strength scales at 145kg changed, and each formula made a different one.
A lifter near the tails who compares their Wilks against their DOTS will usually see the two numbers drift apart like that, and the gap is not noise. It is the formulas openly disagreeing in exactly the region where none of them has enough data to be confident.
The honest way to read any of these scores, then, is as a well-reasoned estimate rather than a physical constant. The history is not a story of one right formula finally being found. It is a story of a sport repeatedly deciding that "close enough, and fairer than last time" was worth another rewrite.
If you want to watch that disagreement play out on your own numbers, Coefflift scores your total under both Wilks and DOTS at once and expands the full worked formula for each, so you can see precisely where and why the two curves part ways.