Pop off pressure has nothing to do with the regulator diaphragm.
Still don't have a good explanation of how non-pressurized fuel line will make a ski run like "crap"
maybe you should go back the drawing board because pop off pressure is made by the fuel pressure against a closed needle and seat. Once the fuel pressure has piped the beadle open it makes its way into the regulator portion. Diaphragm acting as a bowl to keep fuel in and float so when fuel runs low bypasses the pop off.
You assumed that because you use a pop-off gauge to put pressure against the needle, that must be how the carb works. You are wrong.
There is a nipple on the diaphram that pushes on the end of the rocker arm. The diaphragm pushes on the rocker arm, which opens the needle and seat. That is why you can't bend the rocker arm more than 0.020". If you bend too much one way, the diaphram can't open the needle, and you starve for fuel. Bend too far the other way, and the diaphragm always pushes on the arm, causing your needle to leak. Having tuning issues? Don't forget to check your rocker arm...
The diaphragm regulates the fuel flow through the seat. However, since we have no way to measure the reaction of the diaphram during regular running condition, we sample test the "pop-off" backwards (basically) by pushing the needle open (rather than using the diaphram to test the spring. Also, this is why there are 2 parts to checking the popoff. You 1st check the popoff PSI the way we all know, but then you reassemble and let it sit at 10psi for 10 min, if it drops, you possibly have your diaphragm pushing on your rocker arm (bent rocker, damaged diaphragm, dented diaphragm cover).
Your pop-off pressure (in the way we test it) is mostly irrelevant, however, since we can't measure the working pressure of the diaphragm, we use the only measure that is available to us, that which we get from the pop-off test.
Edit: god, I can't believe I wasted all that time writing that when I could have just said "go read page 12". Don't be stubborn. Your foundation is wrong...