You lost me. You are throwing a lot of data out without the analytics. Your also saying the distributor never moves and you set initial by welding inside. Assuming everything else in the distributor is perfect for your new set up of higher initial and less mech adv. Which it is not. I also think some of your assumptions about vacuum are wrong. Was this just a linear computer equation you used, or a generated graph from actual experimentation?

I just measured several Prestolite dual point cams. Everything from the Hemi Cap 15deg distributor adv (0 initial), to a tach drive race cam, and several intermediate ones. The inside of the slot is at the exact same point at about 7/16" from collar. The difference is the length moving outward. I do not have as many Chrysler cams loose, but did check two 8s versus a 11 and again the inside distance from collar was the same. In this case 3/8".

Having studied the differences in shaft tip clocking, and cam stop flat location, it was all about housing orientation to ensure vacuum cannister, oiling port access was not obstructed by manifold, fuel lines, etc... Watching a distributor being fitted on a hemi with cross ram manifold with vacuum advance (street motor). You had limited amount of rotation for setting initial timing. In fact one Chrysler type would not work. When I inspected why, the tip of shaft was clocked slightly different around 10 deg (Chinese knock off). Would not have mattered in a 440 or stock setup. This prevented getting intitial high enough because vacuum cannister hit manifold.

As said in beginning you can do this either way, but that does not mean your tuned right. Weld up to limit mech adv. Set initial by housing rotation, set advance curve by spring choice and tension (adjustable).

Sure, the weights cut a curve, but the rotation is small compared to 360 degree circle. So it can be treated as linear. Well within a tolerance of the mechanical play, bearing clearance, spring rate etc.....