So thick is the desperation on the right now that some Rubio fans on social media are nervously arguing that this is an encouraging poll for him, sort of. The last poll of Florida, taken last month by CBS, had Trump up 41/18. Rubio’s gained 10 points since then. The lead’s been cut by seven points! Only 16 more to go!

…in Rubio’s home state. With just 19 days left. And oh, by the way, voting in Florida has already begun. Mar-a-Lago’s going to be rocking on the 15th:

Three striking results in this poll. The obvious one: Rubio plus Cruz doesn’t equal Trump. Assign Marco every last one of Ted’s votes and Rubio still doesn’t take first place. He’d need all of Cruz plus nearly all of Kasich to pass Trump, and that’s with Jeb Bush’s voters having already been redistributed. In fact, Trump is so close to 50 percent now that if Cruz dropped out before the 15th, it’s possible he’d pick up enough of Cruz’s voters to get there. Rubio and Cruz continue to cling in their interviews to the idea that Trump has a ceiling in the low 30s and that the sole remaining candidate against him will inevitably inherit the great bulk of the party. It’s nonsense on stilts. Trump’s ceiling *was* in the low 30s before any votes had been cast, when it was still an open question whether he could win any primaries. The more he succeeds, the more some undecideds who are well disposed to him see him as being worth a vote. There’s a word for that, in fact. It’s called “momentum.”

Two: We’re used to seeing this by now in polls but it’s no less remarkable for the repetition. Trump is not only beating Rubio and Cruz among white, born-again evangelicals, he’s crushing them. He’s more than doubled Cruz’s take here. Everyone else could drop out today and Cruz still couldn’t beat Trump head-to-head nationally losing his core voters that way. Meanwhile, the now-familiar educational split shows up here vividly with Rubio edging past Trump by six points among college grads and being roundly destroyed by Trump among those without college degrees. (“I love the poorly educated!”) Rubio would need to overwhelm Trump among more educated voters to erase that advantage but Trump is holding his own there, which is yet another reason why it’s silly to think he has a fixed ceiling. He’s the one guy in the race who’s showed appeal across demographics. As Peter Beinart says, “Trump’s not doomed in a two-way race because he’s not a niche candidate.” What’s Rubio supposed to say or do in the next three weeks to either shrink Trump’s numbers among less educated voters or greatly boost his own among college grads?

Third: The silver-bullet explanation for why Trump can and must fail to win the nomination is that too many Republicans say they’ll never vote for him. Ideological conservatives will happily talk your ear off about their personal lists of 57 reasons why they could never support the guy. Some polls have showed as many as 39 percent of Republicans claiming that Trump is in never-ever territory for them. That would sink any other candidate, but Trump’s had a preternatural knack for converting people who like him into supporters at the polls. If 60 percent of the party is open to voting for him, that’s plausibly more than enough for him to get to 51 percent. The thing is, it may not even be true anymore that Trump is the candidate whom Republicans are most likely to say they’ll never support. Quinnipiac asked about that in their Florida poll today. Results:

More Republicans say they’d never vote for Ted Cruz than say they’d never vote for Trump. And more evangelicals say they’d never vote for Cruz than say they’d never vote for Trump. I guess Trump and Rubio screaming “liar!” at him for the past month worked, huh?

Even I can’t handle this much pessimism so here’s another new poll result, this time from WaPo, about Rubio’s comparative advantage among Latino voters who plan to vote in a Republican primary:

All things considered, not a bad result for Trump, though. He’s in second place despite the “deport ’em all” rhetoric that was supposedly going to alienate every last Latino American. And if you’re looking towards the general election, arguably you should prefer a candidate who’s a bit weaker with Latinos but much stronger with the white working-class to one who’s a bit stronger with Latinos and weaker with the white working-class. In theory, Romney’s 2012 take of 27 percent of the Latino vote against Obama is basically rock bottom; any Republican, including Trump, might do as badly but it’s unlikely they’ll do worse. Meanwhile, it’s unlikely that any Republican will win so much more of the Latino vote in swing states that Romney lost to flip them from blue to red without improving among other demographics too. If Trump pulls 27 percent of Latinos but adds many millions more white working-class voters — the “missing” voters in 2012, per Sean Trende — then he should outperform Romney, although perhaps not by enough to actually win.

I’ll leave you with this Namath-esque quote from a Rubio advisor:

Media needs to chill. The FL Q poll #'s are way wrong. We are going to win Florida. Period. Take it to the bank.

— Todd Harris (@dtoddharris) February 25, 2016

Update: Much better for Rubio fans, but a comeback here still likely depends on Cruz getting out and Trump’s Super Tuesday success somehow not building his momentum in Florida:

FL's premier polling shop @VoiceofFLBiz:
Trump 34
Rubio 27
Cruz 17
Kasich 5
Carson 5 pic.twitter.com/JSjTTJiRda

— Marc Caputo (@MarcACaputo) February 25, 2016