Kenny is right.
Have a look at this:
View attachment 5044
That is something I made from the IPF classic world records. First you have the weight class, then you have the records, and last you have percentage of the top lift.
If you have some calculations, I'd love to see them. It would be best if you got them from a single federation so the judging is constant. And of course, a federation with a big sample size.
So the headers are
(Weight class kgs) (Sq) (BP) (DL) correct?
Assuming that's correct, it's very clear just how much the DL improves as the weight classes go up right?
The slope calculation is pretty simple:
((Max of weight class X) - (Max of weight class Y))/((Weight class X) - (Weight class Y))
If you take weight classes 59 and 66 kilos, then:
Squat slope = (250-240)/(66-59) = 10/7 kilos added to the squat for an added kilogram of BW
Bench slope = (189-171)/(66-59) = 18/7 kilos added to the bench for an added kilogram of BW
DL slope = (285-271)/(66-59) = 14/7 kilos added to the DL for an added kilogram of BW
In this case, the BP benefited most, with DL second and squat third.
You can go a bit longer, like 83 vs 59 weight class:
Squat slope = (298-240)/(83-59) = 2.4 kilos
added to the squat for an added kilogram of BW
Bench slope = (209-171)/(83-59) = 1.6 kilos
added to the bench for an added kilogram of BW
DL slope = (326-271)/(83-59) = 2.3 kilos
added to the DL for an added kilogram of BW
This case DL comes at a very close second, with bench far behind.
It occurs to me I might not be understanding at all what you're both saying, but from both data I provided and you provided, it's pretty clear that the DL benefits just as much (if not generally a bit more) from BW gain.
Perhaps I'm misunderstanding your point?
EDIT: On an added note, you can plot all of that data onto a regression program like wolframalpha. He will determine the best fit function for the data (which will most definitely be some higher exponential function, or perhaps some logarithmic function). The linearization is just useful for a rough estimate. Once you have that function, you can take its derivative. That will tell you exactly
how much added BW will add to your max for every one of the lifts, at each weight class. I am very confident if you did that, the derivative for the DL would be generally the same value as the other two. Meaning, at any given weight class, additions of weight will manifest in max additions very similar to the bench and squat. We can make that rough prediction based on the slopes above we took but you might find other interesting things too.
We can talk more about this later if you want as I understand it might be a little bit complicated but something tells me you'd personally find it pretty interesting ^__^