> the rest of the population would probably have voted along much the same proportions as those who did

According to Pew, the non-voter block is much less white, less educated, and poorer than the voting population. You can probably make several cases for which way that block breaks, but I would certainly not say they would vote proportionally the same as voters.

Whatever you assume about how they would have voted, they didn't “literally” vote.

And the assumption that people who choose to vote have an equal distribution of political views to those who choose not to vote is ludicrous; people who, if they voted, would only vote for someone who isn't a major party candidate are going to be vastly overrepresented among nonvoters simply because they are more likely to feel participation is futile; it would be less inaccurate than assuming equal distribution—but still inaccurate—to assume every eligible nonvoter is equivalent to a vote against at least all of the major party candidates.

I don't think this is true. It's been shown over and over again that the more people who vote, the more likely Republicans are to lose the election. This is why Republicans have gone all in on their disenfranchisement efforts.

voter turnout * share = .667 * .469 = 0.312823 or did you count in people not eligible to vote? like kids and foreigner?