A result from this weekend’s YouGov poll that caught my eye as a possible explanation for Trump’s surprising “regret” comments this past week on the stump. Kellyanne Conway, his new campaign manager, is a pollster by trade. If YouGov’s seeing this sort of number in a must-win battleground, chances are Conway’s seen something similar in her own surveys. That explains the sudden, uncharacteristic Trump lunge towards remorse.

YouGov targeted people who said they’re not voting for Trump or Hillary and offered them a list of things he or she might conceivably do to get them to change their minds. The top score for Hillary among Hillary opponents came when they were asked whether they’d reconsider if she convinced them she was telling the truth about her emails and server. Fifteen percent said yes, but that number is largely worthless because there’s little Clinton can do at this point to convince anyone. You either believe her or you don’t. She’s not going to say anything between now and November to change any minds.

Trump, however, does have it in his power to change minds, at least in theory. Here’s what happens when you ask Trump opponents whether they’d reconsider if he apologizes to people he’s insulted. Even if only half of these people changed their minds, that’s a big deal in a tight, valuable state like Ohio:

ygapol

Note that a majority of anti-Trump Republicans are open to reconsidering, which confirms my suspicion that relatively few GOPers are hardcore #NeverTrumpers. Some of the red votes he’s lost were, and maybe still are, gettable. There are other change-your-mind issues on which Trump scores relatively high as well. Ask people if they’d reconsider if he convinces them he’s prepared to be commander-in-chief and 19 percent say yes, including 52 percent of anti-Trump Republicans. (Fully 70 percent of Ohio voters overall say Trump would be a risky choice as president.) Ask them if they’d reconsider if he stops saying controversial things and 14 percent say yes, with 51 percent of anti-Trump Republicans agreeing. Obviously there’s some overlap among people who answered yes to each question but the point is that there are voters still out there whom Trump can convert and, importantly, the bar he needs to clear to convert them is comparatively low. Stop acting like a jackhole, show a little phony contrition to sympathetic critics like the Khans, and study hard enough for the debates and interviews so that you sound like you know what you’re talking about rather than like a guy on a barstool. Overtaking Clinton may be difficult given her organizational advantages but it should not be difficult for him to make the race closer. Which is yet another reason why the whining about #NeverTrump voters is so irritating.

YouGov, by the way, has Clinton up six points in the four-way race in Ohio. A new Monmouth poll of Ohio this afternoon has her up four, thanks to the by now familiar phenomenon of Clinton winning a bigger share of the vote among white women at Trump’s expense than Obama did at Romney’s. Those two margins, four points and six points, are right in line with the RCP average of Ohio, which is also four points. Both Nate Silver and Michael Barone have noticed something quirky over the past week: Hillary’s lead in national polls is shrinking but her lead in the swing states, which is the more important metric, remains solid. How can that be? If the race is tightening nationally, why isn’t Ohio? Barone’s theory:

The evidence is somewhat scanty, but it points to Clinton running better in target states than she is nationally. That’s out of line with experience, in which target states taken together vote pretty much the same as the nation as a whole: In 2012 the top ten target states, which cast 29 percent of the nation’s votes, voted 51-48 for Barack Obama over Mitt Romney, very close to the 51-47 margin nationally and in non-target states.

Possible explanation: Clinton’s campaign has been running significant amounts of TV advertising in most target states (though not Colorado or Virginia, where she has been polling well ahead) and Trump has not. Interesting question: If Trump starts running equivalent amounts of TV advertising, can he close the apparent gap between national and target state preferences? If not, it’s hard to see how he can get to 270 electoral votes.

Hillary’s campaign has already spent $70 million on ads and just reserved an additional $80 million worth of airtime for the fall. It may be that her swing-state polling would be tracking her national polling if Trump was contesting the airwaves in those states, but with Hillary running almost unopposed there until recently, swing voters have stuck with her. Hopefully for Trump’s sake it’s not too late to undo that. The RNC will have to place a bet soon on whether it is or isn’t possible by deciding whether to keep its funds earmarked for the presidential race flowing to Trump or to redirect them to Senate races. They might not have any choice but to stick with Trump, notes the Times: Because so many deep-pocketed Republican donors aren’t contributing this year, the RNC is more dependent on Trump’s small donors. If they cut him off, they’ll need those big donors to step up immediately and make up the shortfall for when Trump’s base turns their back on Reince and crew.

By the way, there’s a new poll of Pennsylvania from a pollster I’m unfamiliar with that shows Trump up five points over Clinton, which, if true, would represent a 15-point turnaround in the state in a matter of a week or so. Some Trump fans are touting it, but I’d wait for a major pollster to confirm that finding before concluding that the race has shifted. Exit quotation from Monmouth: “A solid 57% majority say they would have backed [John] Kasich for president over 33% who would support Clinton if this had been the match-up.”