Perhaps so. We are so close to agreeing, and we do agree on the most salient point that the the metal at the surface of the die near the center devices flows inward to fill the voids in the die. Where we don't seem to agree is what happens closer to the edge of the die. My contention is that metal must flow outward to fill the void between planchet and collar and to fill in the rim. From your statements, you seem to believe that even in this region the metal flows inward, in spite of the lack of any external force to make that happen. I'd like to hear if you actually still believe this, or if you understand it could not be so.
Seriously, didn’t I present empirical evidence based on actual measurements with a micrometer just a couple of posts ago? If we had measurements of a planchet before and after upsetting, we could mathematically define how much metal was displaced. Then if we had a cross-section of a struck coin we could draw flow arrows showing how far, in which directions, the metal had to move. But even with only partial data we know a planchet that has a smaller pre-strike diameter than the final coin has to have some outward flow. A planchet that is thinner at the rim than the struck coin has to have some upward flow. The question I think we should investigate next is how to formalize this knowledge. What is the size of a planchet before and after a rim is formed by the upsetting mill? The metal moved at this stage is reshaped into the raised ridge around the planchet rim. If this had more volume than the rim needs, after it flows up and out against the collar, the excess flows inward towards the center. If the reservoir of metal in the upset rim is not sufficient to form the final edge and rim, then metal has to flow outward from the fields to supplement it. While this may seem like quibbling, I for one would like to see a cross-section of a planchet before and after upsetting so I could compute the volume of metal held in the pushed-up rim. Then I would like a cross-section of the final coin so we could see the volumes to fill. After that we could draw a diagram (once and for all) showing how far the metal moves, where from, and which direction it goes.
Unfortunately I don't have any blanks or planchets. I'd certainly be willing to do the cross-sections (they are one of my FA specialties) if someone wants to sacrifice one of each. That said, it's clear that the volume of planchet metal within some tbd distance of the edge of the finished coin is insufficient to complete the strike, so the metal near the edge will have a net outward movement. I suppose the quantitative analysis would tell us the "tbd", and thus the effective null-movement point (what I called "ambigous" earlier).