I started the process, then dumped the experiment. From what I could see, the edges of the sprue were rounded down, but the high point was basically unaffected. Yeah, you couldn't see the "sprue" with a quick glance, but you still had a high spot. That made it all that much harder to see so you could center it at the top while loading, though it certainly miked high enough to be an accuracy issue.
Maybe I didn't put enough "miles" on them. In any case, they went back into the melting pot without firing a single one.