Relax, relax. It’s a joke. The elections don’t start getting canceled for real until 2020.

His point here is similar to Frank Luntz’s point last night about how Rubio and John Kasich might have fared against Clinton. Namely, how can it be that a candidate as bad as her might possibly win? She was relegated today to being the warm-up act at her own rally, for cripes sake. She’s 12 days away from becoming the most powerful person in the world and she’s still in the “Puppet Show and Hillary Clinton” stage of her political appeal. I … guess we know the answer to the question of how she can conceivably be winning, although some of the replies to the Luntz post last night surprised me in arguing that Rubio and Kasich would surely be trailing too. How could either of them win, commenters asked, when they got crushed in the primaries by Trump? Trump humiliated Rubio in his own home state of Florida, winning by nearly 20 points. If a candidate that strong is trailing Clinton (narrowly) in Florida now, obviously Rubio would be getting flattened by Clinton there, right?

Well … no. One doesn’t follow from the other. To see why, spend 30 seconds scrolling through this nifty graphic. There are around 30 million voters in a Republican primary; there are another 100 million non-Republican voters to be wooed in the general election. Is it possible that a candidate who had less success with the first group would have more with the second, for a net gain in votes overall? Sure, why not, if their politics lined up better with the general electorate? To believe that Rubio would be less competitive than Trump in Florida, for instance, you need to believe that some meaningful number of Trump fans in Florida would have sat out next month’s election if Rubio had won the nomination fair and square. Some would have, especially given Rubio’s immigration record, but the same logic Trumpers have spent the last six months using — “what about the Supreme Court?” — would have convinced most of them to grit their teeth and vote Rubio anyway. So Rubio would have started with most of the same Republican base that’s showing up for Trump in Florida now, and from there he would have branched out into the general electorate to try to woo swing voters. Trump’s numbers are abysmal among millennials and will certainly be terrible next month among Latinos (although maybe not more terrible than Romney’s in 2012). How many voters in those groups would have switched from Clinton to Rubio is unknown, but given how low Trump has set the bar, there’d be some. Rubio wouldn’t have had the same baggage Trump does among women either, which would have reduced Clinton’s advantage further. He wouldn’t have been nearly as strong as Trump among white working-class voters, I’m sure, but he also certainly would have been stronger among the white college grads who have been abandoning Trump. Remember that Bloomberg poll from a few days ago showing Trump ahead by two in Florida? Trump’s favorable rating there was 46/51 — which is surprisingly solid by Trump standards. Rubio’s rating was 50/44.

You can run that same thought experiment with Kasich, of course, who actually did win his home state of Ohio. Trying to compare the relative strength of candidates among a primary electorate versus a general electorate is like trying to compare how they’d do running in two entirely different states, one very red and one purple-bluish. How would, say, Ted Cruz do against Clinton in socially conservative Utah? (Here’s a clue.) And how would he do in, say, Pennsylvania? Given how well Clinton has consolidated Democratic support there and how Cruz lacks the same appeal as Trump to PA’s white rural voters, he’d probably be trailing badly. We understand this point, that different candidates play better or worse with different electorates, intuitively in most cases, but suggest that a guy who lost the primary might have been stronger in the general and some people go to pieces over it. Just remember this: All of the models this year that rely on “fundamentals,” like how long one party has controlled the White House, how the economy’s doing, and so on, favored a Republican victory this fall if the party nominated a “generic Republican.” If they’d nominated a bland, replacement-level politician — John Kasich is a perfect example — the GOP probably would have won. A good politician might have won easily, a bad politician might have won very narrowly or, ahem, lost, but a generic Republican probably would have won based on the “fundamentals.” (And that assumes a generic Democrat on the other side. Hillary Clinton may be below replacement-level herself.) The party didn’t want a generic Republican this year, though, because it tried that in 2012 and it didn’t work out, even though the fundamentals pointed to a different outcome at the time. But that doesn’t mean a generic R wouldn’t have been a better choice for the general election. It just means Republican voters didn’t care enough about electability to let that reality trump Trump.