I voted in California, too, using the old punch card system. How exactly is that not a paper trail?
As far as running the numbers, how do you propose to model the change in delegates based on an imaginery number of votes that were either transferred from Sanders to Clinton, or the number of “suppressed” votes that would have gone to Sanders as opposed to Clinton? If you want to insist that a large number of independents were not allowed in the Democratic primaries, that’s valid, but it doesn’t constitute voter suppression. Those rules were in place since 2008, at least. So Sanders supporters had ample time to inform themselves and act accordingly. Where I voted, I heard a number of independents state that they wished to vote in the Democratic primary. They were issued provisional ballots. In the article I linked to, a Sanders campaign spokesman did not dispute the final result going to Clinton; he was unhappy that the initial results changed the news narrative. But the reality is she won enough delegates before the polls closed in California.
I don’t understand how Clinton winning by a million votes in the scenario I proposed could completely change the entire nomination process. If it did, and Sanders won more total delegates while receiving one million less votes, we would be up in arms about a non-representative voting system.
Clinton won the most important states, and the exit polls aligned with those results. If you are searching for a magic result where Clinton would not have enough votes to win the nomination outright and then hope for a divided convention, that may be possible. But Sanders winning the nomination outright sounds like pure fantasy.
Please feel free to show mathematic models that could calculate these impossible to determine variables.