Not a physicist or engineer here, but trying to follow anyway. I looked up Hooke's law, as mentioned by @oldgoat, but I was not up to the math. (I can calculate a tip, and after that I pretty much run into the brick wall of my childhood math phobia. But I can be logical.) That link led me to a related article on the spring scale, and this "a-ha" phrase: "the scale markings on the spring balance are equally spaced". In other words, if it takes exactly one unit of (additional) force to move the needle from 4-5, it also takes exactly one unit of (additional) force to move the needle the same distance from 9-10.
This rule suggests a binding design such that the distance the spring must move from "at rest" to "release" is always the same, regardless of the DIN setting. Is that how things actually are? If it is, then the release force over and above a given DIN setting should also always be the same. Let's say that force is quantified as "fred." Then the release force is always DIN + fred.
So, if the distance the spring must be moved before releasing is constant, and the force to move it that distance is constant, the only thing that varies is the preload (DIN), or how much force we need to put into the system before we move the spring that last increment. If all this is correct, it SOUNDS like the nature of the travel between the at-rest position and the release position should not be different at DIN 8 out of 15 than it would be at DIN 8 out of 10.
Therefore I suspect that people must be saying that as you get closer to maxing out the total travel of the spring, the "spring scale" principle outlined above must break down, and the additional force required to move the spring a given distance must ramp up or become more progressive. Correct? If that does happen, are binding manufacturers compensating for this by designing the bindings in some way such that the spring needs to be moved a shorter distance to release under higher DIN settings? If so, that would explain the argument alexzn and others are making, which seems to assert less elasticity in this scenario.