“For as bad as all this s—t is,” said Jeff Roe, the Republican strategist who steered Ted Cruz’s 2016 presidential campaign, “you can certainly see how he could win.”

It’s a big “if” resting on a tower of uncertainties — explosive Republican turnout on Election Day, including among new, same-day registrants; tepid enthusiasm for Biden, especially among young people of color; disruptions at polling places that depress Democratic turnout; legal victories related to ballot counting; and a full-on embrace of Trump by the few voters who remain undecided.

Trump will need most or all of those things to happen in multiple states to engineer a victory, according to interviews with more than a dozen Democratic and Republican strategists and campaign officials. But as they run their final simulations of the Electoral College map, they see a credible — if distant — opportunity for a surprise.

Overall, forecasters put Trump’s odds at about one in 10.

Last week, a Democratic strategist who works with major party donors drew up a series of potential maps that included one Trump victory with 279 Electoral College votes, one tie and one scenario in which it’s “259-259 going into 3 a.m. and Pennsylvania is the last state that hasn’t been called.”

If that happens, all of Trump’s preparations for post-Election Day legal challenges will loom large. “If we don’t win on election night, they will fight this, and they will come away with a victory,” the strategist said.

Trump’s advisers and allies say they are more optimistic about his prospects than they have been in months. Polls tightened slightly in some states following the second debate, and they think the latest strong economic numbers could help his cause with late-deciding voters.

Trump is sprinting across the battleground map, with 14 rallies in seven states over the weekend and on Monday. And he hasn't ruled out some last-minute rallies on Tuesday.

The best-case scenario for Trump — the one most of his supporters pin their hopes to — is that the polls are off. State-level polling was off four years ago, and there have been some eye-popping discrepancies in battleground states this year, though Biden has consistently led in most of them.

An ABC News/Washington Post Poll released Wednesday showed Biden leading by 17 points in Wisconsin. The same day, a Marquette Law School poll had Biden up 5 points in the state.

If public polling averages in battleground states are off by even 2 percentage points, Trump would currently be within 1 percentage point of Biden in Florida and North Carolina, and just over 1 point behind in Arizona. Ohio and Georgia wouldn’t be as worrisome for Republicans as they are now. And the margin in Pennsylvania, the linchpin state for Trump, would be cut to a margin of just over 3 points, based on FiveThirtyEight’s polling average.

“He’s going to have to pull two inside straights, where he pulled one last time,” said Mike Madrid, a Republican strategist and one of the founders of The Lincoln Project, a Republican group opposing Trump.

However, Madrid said, “If he does have a dramatic overperformance [with his base demographic of non-college educated whites] in places that we’re talking about, it’s going to set most of those states in motion.”

He said, “I could easily see them winning North Carolina and Florida.”

Trump’s team is focused on states the president won four years ago, though his aides know he likely won’t sweep Pennsylvania, Michigan and Wisconsin, the three Rust Belt states he flipped in 2016.

Outside of Florida — a must-win for Trump — Pennsylvania is the most important state that Trump and Biden are fighting over. Trump could win Pennsylvania and still lose the White House. But if he doesn’t carry Pennsylvania, where he is polling behind by about 5 percentage points, he has almost no chance. That explains why Trump has been focusing relentlessly on the state and headlined four rallies in the state on Saturday.

In his last-ditch effort in Pennsylvania, Trump is decrying protests following a fatal police shooting in Philadelphia — a play for suburban voters that’s fallen flat elsewhere — while hammering Biden for his comments about hydraulic fracturing and transitioning away from oil dependency.

“How does Trump win Pennsylvania? Not easily, I’ll start it there,” said Christopher Borick, director of the Muhlenberg College Institute of Public Opinion in Allentown, Pa.

The effect of fracking on the outcome in the state “is fairly baked in already,” Borick said. However, if Trump can improve his standing with seniors and drive “historically strong turnout numbers” in rural areas, he said, it’s still possible Trump could win.

If he can engineer an upset, Trump will likely rely instead on variations in turnout and the mechanics of the election, including post-election court challenges and Trump’s extensive field operations. Democrats in Arizona remain concerned about the possibility — however unlikely — that enough Mormons could break late to Trump to put him over the top in the state.

Democrats in Minnesota on Thursday scrambled to warn voters not to mail ballots anymore, after a panel of federal appellate judges suggested late-arriving ballots could be in jeopardy of being invalidated. The Trump campaign’s efforts to limit voting in other battleground states could still depress Democratic turnout or the number of ballots that are ultimately counted.

Trump himself has predicted that the election ultimately “will end up in the Supreme Court.”

“His path is a narrow Electoral College victory, it’s getting unprecedented turnout on Election Day, it’s having everything go right on Election Day,” said Tom Bonier, CEO of the Democratic data firm TargetSmart. Then, he added, Trump “has to win in court … The one thing that is just repeating in my mind now is just the lack of any sort of margin for error for him.”

Of the 17 states Trump had initially targeted, he has largely abandoned three states that he lost in 2016 but hoped to pick up — Colorado, Virginia and New Mexico. Trump is concentrating on just three states he didn’t win in 2016 — Minnesota, Nevada and New Hampshire — to help make up for states he might lose. For example, he could lose Wisconsin but carry Minnesota or both Nevada and New Hampshire and still win the same number of electoral votes.

“It’s about trying to hold what you got," said Scott Jennings, who worked for George W. Bush and is close to the Trump White House, "and knowing you can lose one or two states.”

That’s a luxury that Trump does not seem to have attained.

Trump campaign officials acknowledge he's the underdog heading into Tuesday. But they insist they're more confident in his chances than they have been in months. Trump has moved past his Covid-19 diagnosis and his much-maligned first debate performance, and last week's big GDP numbers focused attention on the economy — the president's best issue — at an opportune time, they say.

"While we take nothing for granted," said Steve Cortes, the Trump campaign’s senior adviser for strategy, "we are increasingly confident.”

In Florida, the president's top campaign lieutenants insist Biden's lead is narrower than Hillary Clinton's was at this point in 2016. Polls compiled by RealClearPolitics in the last presidential cycle showed polls with the Democratic nominee tied with Trump, trailing him by 3 percentage points and leading him by as much as 4 percentage points in the final week before the Nov. 8 election.

"Even though the national numbers are much worse than where we were against Hillary at this point, the battleground numbers are virtually the same," a Trump campaign adviser said.

Gabby Orr contributed to this report.