Yesterday Axios published a memo from Doug Sosnik, the political director of the winning Clinton ’96 campaign, laying out Biden’s likeliest path to the presidency. Gotta be Pennsylvania + Michigan + Wisconsin, right? Dems will try to rebuild the blue wall in the Rust Belt and having a nominee from Scranton is going to help them do it.

Not so, said Sosnik. It’s true that Pennsylvania and Michigan are very much in play this year but Trump’s job approval in Wisconsin has been consistently higher than it is in most of the country. That state may be a heavy lift for Biden. And if Trump holds it, it could be decisive. Winning PA and MI gets the Democrat to 268 electoral votes. He needs one more state (or, I guess, he needs to pick off the two congressional districts in Maine and Nebraska that award one electoral vote each to the winner in their district). Which state can he get?

Arizona, says Sosnik. They elected Kyrsten Sinema two years ago and Mark Kelly looks increasingly likely to bump off Martha McSally this fall, raising the unthinkable prospect of two Democrats representing Goldwater’s home state in the Senate. Dems obviously can win statewide there. Pennsylvania + Michigan + Arizona is Biden’s road to the White House.

If you believe this new poll from OH Predictive Insights, he leads by seven points in AZ right now.

Biden’s ahead among all income groups, and among the “double haters” in the electorate (the people who view both nominees unfavorably) he’s up 63/6(!). Trump won that group from Hillary Clinton four years ago, which may have been decisive. Biden’s numbers are less noteworthy in the graph above than Trump’s are, though: With the exception of December 2019, when Republicans may have rallied to the president briefly amid the Democrats’ impeachment effort, it’s been a year since he topped 44 percent in one of OHPI’s Arizona polls. With no serious third-party candidate in the mix this year, that number has to change or he’s a sitting duck in November.

The new numbers for McSally in her Senate race against Mark Kelly are even worse. Gruesome, really:

While the April poll of 600 likely voters favored Kelly 51% to McSally’s 42%, in May it’s now 51%-38%

In May 2019, this same tracking poll showed Kelly up over McSally, 46%-41%, among likely voters in Maricopa County.

In May 2020, Kelly has climbed to 54% in Maricopa County while McSally has dropped to 36%.

That’s a stunner when you consider that Maricopa County in recent years always has gone for Republicans (well, except for now-ex-state Superintendent Diane Douglas and McSally).

At the rate McSally’s going, the GOP isn’t even going to spend money on her this fall. Better to write the seat off and give the cash allotted to her to someone with a stronger chance of holding off a tough opponent, like Steve Daines in Montana.

Should we believe these results, though? Two of the smarter election-watchers whom I follow on Twitter, Nate Cohn and Harry Enten, groaned this morning that OHPI (and other state pollsters) aren’t weighting by education when tabulating their results. That’s a big deal because it risks undercounting Trump’s base of white voters without a college degree; what OHPI apparently did was weight its results according to the 2016 Arizona exit poll … which is itself unweighted by education, notes Cohn. Bottom line: He estimates that polls that aren’t weighting properly education-wise are overstating Biden’s lead by three to four points. So what OHPI is claiming as a seven-point lead for the Democrat may be more like three points, which would be more in line with what polling in Arizona showed in March.

So the state’s still competitive. That’s the good news for Trump. The bad news is that the polling in Biden’s other two decisive states, Michigan and Pennsylvania, has him holding healthy leads in the range of five to six points. Team Trump has been pessimistic about winning Michigan again since last year and reportedly has grown more discouraged lately, and of course Biden is a Pennsylvania native. If those two states slip away from Trump, the whole election could come down to Arizona.

Although, realistically, if PA and MI are out of reach before next fall, that bodes ill enough for Republicans that Biden will probably have multiple pick-up opportunities across other battlegrounds — Florida, North Carolina, Wisconsin. Winning any one of them would hand him the White House. But as unlikely as it may seem, Arizona’s his best bet right now.