This morning I tweeted that, for all the hype about Marcomentum, no poll of New Hampshire has showed him within single digits of Trump yet. The closest was last night’s dubious CNN poll, which had Rubio within 11 but also had a margin of error of nearly seven(!) percent.

Enter the Boston Globe/Suffolk survey, which has Rubio within 10 via a respectable sample of 500 likely Republican voters. That makes three polls in a row where he’s at 17-19 percent and has second place to himself. Seems safe to say at this point that the Rubio surge is real, with a potentially gamechanging debate less than 36 hours away.

New York businessman Donald Trump has a nearly 10 point lead over the field — but the gap between him and the rest of the field has shrunk since his disappointing second-place finish in the Iowa caucuses. Trump received 29 percent in the survey, while US Senator Marco Rubio of Florida took second place with 19 percent.

Ohio Governor John Kasich placed third in Suffolk University/Boston Globe poll with 13 percent, former Florida governor Jeb Bush had 10 percent , US Senator Ted Cruz had 7 percent, New Jersey Governor Chris Christie fell to 5 percent, and Ben Carson had 4 percent.

“What a difference a caucus makes,” said David Paleologos, the director of the Political Research Center at Suffolk University. “By exceeding expectations in Iowa, Marco Rubio is converting likability to electability even more so than Ted Cruz, who, like many conservative Iowa winners of the past like Mike Huckabee and Rick Santorum, can’t seem to convert an Iowa win into a major showing in New Hampshire.”

Fully a third of likely voters say they might change their minds, so the race could go anywhere by Tuesday. Also worth noting: For all the abuse thrown at the CNN poll last night, the data there lines up (with one important exception) with the two newest polls of New Hampshire today. CNN had Trump at 29, Rubio at 18, Kasich at 12, and Bush at 10. The Globe is within a point of those numbers for each of them and WSJ/NBC is within two points for Trump, Rubio, and Bush:

nb

All three polls are within a point of each other on Chris Christie’s support too: Two of them have him at four percent and one has him at five, which puts him a fairly distant fourth among the four-man race on the center-right. I wonder, seeing these numbers now, if he wishes he had a do-over on his “destroy Rubio” strategy this week. Even if he hasn’t driven his own numbers down by going sharply negative, he’s got no realistic chance to win and he’s alienating the guy from his lane of the party who’s most likely to be in a position next year to give him a federal job. Oh well. As for the big discrepancy between the polls, that has to do with Cruz: CNN had him at 13 percent, WSJ/NBC had him at 15 percent, and the Globe has him at … seven. I’m going to assume that’s an outlier, unless Trump’s accusations of cheating in Iowa have gained so much surprising traction within the past 24 hours that they’ve ruined Cruz in New Hampshire. If he pulls another weirdly low number in another poll this weekend, Team Cruz will have a real damage control problem. And Trump, who’s now taken to hugging Cruz again, may go right back to screeching about fraud.

One more bit of data for you, this time for a new YouGov poll. That’s an … interesting trend on the right:

yg

A 30-point decline for Trump in less than a week. He’ll claw some of that back and be the favorite again if he wins on Tuesday (although maybe only a slight favorite if Rubio finishes a strong second), but the days of Trump seeming invincible to a majority of the party are gone. If you subscribe to the “Wizard of Oz” theory of Trump’s success, that his mystique will fade once voters realize he’s not all-powerful, there’s your best evidence yet that he’s already on his way to also-ran status.

One ominous note in closing for Rubio fans: You’re still not seeing Kasich’s or Bush’s supporters peeling off today. Rubio doesn’t need to climb in a poll for that poll to bring good news; a survey that showed Bush or Kasich fans shifting to “undecided” would be a big deal for him, since he’s the likeliest beneficiary of their last-second doubts. Both candidates have been steady this week, though, meaning that Rubio may have to pick off Trump supporters (or Cruz supporters) if he’s going to win in an upset. That’s harder to do than consolidating the center-right vote, at least in theory. Exit question via Noah Rothman: If you believe CNN, Romney’s holding off on endorsing Rubio because he hasn’t seen any data suggesting that Rubio stands a real shot of winning New Hampshire. Being within 10 points of Trump and trending upward when there are still 33 percent willing to change their minds seems like a real shot, no? Plus, what would reliable data look like in this case? New Hampshire polling is often garbage. If Rubio’s within 10 here, he may well already be tied with Trump in reality. Take a chance.