Thanks for the link. I read through the articles and I did like the PHR one. I agree with you in a sense - that the increase or decrease in overlap area does not have a linear relationship with the increase or decrease in degrees. The area of overlap can grow exponentially as the lobes get closer together (depending on the exact cam profile, of course).
But I do think that on a purely factual basis, decreasing LSA by 2 degrees will increase overlap by 2 degrees. And increasing LSA by 2 degrees will decrease overlap by 2 degrees.
However, to tie back in to the point you made, the 2 degrees of increased overlap are moving in the direction where the cam lobe is getting taller. And the 2 degrees of decreased overlap are moving in the direction where the cam lobe is regressing to the base circle.
So it's always 2 degrees, but these 2 degrees may be "bigger" than those 2 degrees!
That's how I understand it at present, anyway.