A leftover from last night. In case you thought there’s even a whisper of a chance that the GOP will win this November, disabuse yourself of that now. It’s not just #NeverTrump that threatens to sink the party. It’s #NeverCruz too.

When asked what they would do if Cruz were the GOP nominee in November, only 65 percent of Wisconsin Republicans said they’d vote for him. The remainder instead would vote for a third-party candidate (18 percent), vote for Democratic front-runner Hillary Clinton (7 percent) or not vote at all (6 percent).

The numbers got slightly worse for the Republican Party when voters were asked to consider Trump as the GOP nominee. Just 61 percent said they’d vote for the brash businessman, with the rest defecting to a third party (16 percent) or to Clinton (10 percent)—or simply staying home (9 percent).

Question: Are the #NeverCruz people really #OnlyTrump people? I think #NeverTrumpers would be willing to vote for damn near any third-party nominee with conservative cred — Tom Coburn, Rick Perry, you name it. With #NeverCruzers I’m not sure. I think many of them are imagining the third-party candidate in this scenario to be Trump himself. Right? If Cruz wins on the convention floor, Trumpers will be expecting Trump to declare his independent candidacy, sore-loser laws be damned, and to battle on to November. I don’t know if he’ll do that. He’d have zero chance to win and he’s not a guy who seems eager to spend down his nest egg on building a formidable campaign even in the current favorable circumstances (if he was, he’d probably have the nomination locked up by now). Why would he spend to build one for a lost cause? I think he’d sit it out, encourage his fans to write him in — which many would — and spend the fall sniping at Cruz in media appearances. He’d still do major damage to Cruz as a write-in, but not to the tune of 18 percent. I think 5-7 percent is realistic, with plenty of other Trumpers reluctantly resigned to voting Cruz once it’s gut-check time in late October. Although don’t misunderstand: I think Cruz is a longshot this fall, albeit not as much as a longshot as Trump. Even with an excellent organization, his odds of beating Hillary after a bitter floor fight and splintered party can’t be more than 40 percent.

If you missed the rest of the exit polls last night, skim through them now and see what jumps out. Right away, you’ll find the numbers among men and women were nearly identical for all three candidates with Cruz winning 48/35 in both groups. That explains why the now infamous ARG poll of Wisconsin was so wildly wrong. ARG correctly had Cruz winning big among women but they put Trump ahead among men … 57/23. They missed by 47 points. For all the hype, including from me, about a looming gender gap in Wisconsin, there was no gap at all. Cruz crushed it among both sexes. Meanwhile, ideologically, in the past Cruz wins have been defined by running up the score with “very conservative” voters. He did that again last night, 65/28, but he also topped Trump among “somewhat conservatives,” 47/36, and finished just 11 points behind among moderates, a group among which Trump usually kills him. (Relatedly, Cruz tied Trump at 40 percent among independents, an amazing neutralization of the threat posed by Wisconsin’s open primary.) If you’re invested in the idea that last night was a turning point, writes Nate Cohn, that’s your best evidence. It may be that centrists, who haven’t been warm to Cruz thus far, are finally cutting Kasich loose and starting to line up behind Cruz as the last Not Trump standing. If Wisconsin is the start of a national trend in that regard, with Trump no longer able to count on his core demographics to deliver for him, he’s in trouble.

It appears that many moderate voters, who have long been the biggest obstacle to Mr. Cruz, finally broke his way. According to exit polls, Mr. Cruz won 29 percent of them — far higher than the 12 percent he won in Michigan and 15 percent in Illinois. Mr. Kasich’s share of the vote among both self-described “moderate” and “somewhat conservative” voters dropped…

But perhaps the best reason to think it still might be part of a broader phenomenon for Mr. Cruz is that he has outperformed expectations at every point since Super Tuesday. He nearly doubled his support in the contests immediately after Super Tuesday, as Mr. Rubio faltered. He posted strong showings on March 15 — like clearing 40 percent of the vote in Missouri and 30 percent in Illinois — that were largely overlooked because it was not enough for him to win…

If Mr. Cruz’s step forward is representative of the rest of his race, Mr. Trump will no longer be on track to amass a majority of delegates. Here’s one way to think about it: If Mr. Cruz outperforms our model by the same amount that he did Tuesday, Mr. Trump will go from a favorite to an underdog in California, Indiana, Maryland and Montana. Pennsylvania would be competitive. Mr. Trump would not even be near the 1,237 delegates needed to win the nomination.

There’s no great mystery to what’s happening here. Rubio fans began drifting towards Cruz weeks ago, even before Rubio dropped out, as they concluded that Cruz was a better long-term bet to stop Trump than Marco was. Once Rubio made it official and quit, that drift accelerated. (In fact, per Philip Klein, Cruz has won more delegates since Rubio left the race than Trump has.) Now you’re seeing something similar happen with some of Kasich’s supporters: As he continues to lose, and lose badly, in one state after another, parts of his base are reluctantly accepting that he’s a lost cause and swinging around to Cruz as their #NeverTrump choice. It may be that Cruz is going to run into diminishing returns with the most loyal Kasich voters, who are likely firmly centrist and have resisted all appeals to switch to Cruz so far. But then, it might not matter. Cruz’s goal at this point isn’t to get to 1,237 before the convention or even to win states per se. His goal is to hold Trump as far below 1,237 as he can. If he can split the remaining delegates on the table in the primaries with Trump 50/50, he’ll accomplish that easily. And who knows? Maybe some of those stubborn Kasich dead-enders will be newly alienated by whatever boorish thing Trump says this week or next week or the week after that. Trump will continue to produce defectors for Cruz through his own behavior. The only question is how many.

In fact, here’s a stat that sums all of this up nicely:

Trump’s audience isn't growing, @alexcast says. Average vote share after 30 states is 35.6% —almost same as 35.3% in his 1st victory, in NH

— Jennifer Jacobs (@JenniferJJacobs) April 6, 2016

Nate Silver fleshed that out a bit:

Trump came close to his results in neighboring states tonight. Cruz had a breakthrough, though. pic.twitter.com/Fr9OFsNZ0y

— Nate Silver (@NateSilver538) April 6, 2016

Trump has his loyal third of the party (okay, slightly more than third) but it’s not growing. It’s Cruz’s numbers that are growing as the race becomes the “Trump vs. Not Trump” contest he’s craved all along. Trump has to reverse that in the few remaining midwestern states to come, like Indiana, to get to 1,237 before Cleveland. He’s going to crush Cruz in New York in two weeks, but that landslide has already been priced into his chances at winning a majority of delegates before the convention. He needs some mojo in the toss-up states, not the sure things, to help himself. Where does he get it?