Bernie Sanders gestures as he greets voters at the Reading Terminal Market on primary election day in Philadelphia on April 26. (Dominick Reuter/Reuters)

Bernie Sanders entered the 2016 Democratic nominating contest without really expecting to win, as the New York Times reported last month. His goal was to draw attention to the role of money in politics and the need to address income inequality — a goal that he has, by now, surpassed by orders of magnitude. That focus, combined with Sanders's appeal as an outsider and broad antipathy (particularly among white men) to Hillary Clinton, fueled an unexpected surge of support for Sanders. It manifested first in crowd sizes, then in poll numbers — and, finally, in votes.

As Sanders's campaign evolved from non-starter to actual contender, so has Sanders's argument for his path to victory. As that path has looked less possible, those arguments have grown increasingly complex.

The big picture

Last fall, Bloomberg's John Heileman spoke with Tad Devine and Jeff Weaver, Sanders's chief strategist and his campaign manager, respectively. Devine and Weaver outlined a plan for beating Clinton in the nominating process that involved a number of components: Compete on money, undercut Clinton's support with Democrats, improve with black voters and convince Democrats that Sanders can win in November.

The campaign accomplished the first, cobbling together millions of small donations to beat Clinton consistently for the first three months of the year. They accomplished the latter, too, with Sanders regularly outpolling Clinton against presumptive Republican nominee Donald Trump. (At the time Devine and Weaver spoke to Bloomberg, a majority of Democrats thought Sanders definitely couldn't beat the unknown Republican nominee.)

On the other two metrics, Sanders's campaign was not successful. Clinton's favorability with Democrats has dipped slightly; it was 81 percent last May and 75 percent in March. In between, though — and right before voting began — it rose back up to 84 percent. Black voters, though, have consistently preferred Clinton to Sanders, according to exit polls. The average across all states gives Clinton a nearly 4-to-1 advantage. In only five states has Sanders hit the 30 percent of black support the campaign was aiming for in South Carolina (which wasn't one of the five states).

Early on, some Sanders supporters insisted that black voters would switch to Sanders once he proved he could compete by winning early contests — mirroring what happened to Barack Obama in 2008. That didn't happen.

The early sweep

Early victories became a key checkbox for Sanders's long-term efforts. "His strategy for capturing the Democratic presidential nomination was based on sweeping all three early-voting states," the Times' Patrick Healy and Yamiche Alcindor wrote, suggesting that Sanders figured that such an outcome might prompt Democratic supporters of Clinton to reassess his (and her) viability. (That this was unlikely to affect the results of the fourth contest — South Carolina — and other heavily black Southern states is left unmentioned.)

That sweep didn't happen. Sanders essentially tied Clinton in Iowa, crushed her in New Hampshire — but then lost in Nevada. Clinton, largely on the strength of the black vote, continued to win big victories through March 15, building up the lead that Sanders hasn't been able to touch.

It's worth noting that the period between New Hampshire and Nevada was also a period when the most frustration at Democratic party superdelegates was expressed. At that point, Sanders led with pledged delegates, but Clinton's advantage with the unpledged superdelegates gave her more votes at the convention. A MoveOn petition demanding that superdelegates withhold their decisions until the votes were in garnered tens of thousands of signatures,

With Clinton's wins, this quickly became unimportant.

Momentum

After mid-March, the territory shifted dramatically in Sanders's favor, with a number of caucuses in states with small black populations. The campaign has been talking about momentum for a year, focusing earlier on the scale of the events it was holding and, later, on his gains in the polls. After Clinton jumped out to her big pledged delegate lead, though, the campaign started talking about winning states as providing momentum. This, it clearly hoped, might provide an argument to superdelegates that the party as a whole had shifted to Sanders — that the will of the voters was no longer that Clinton be their pick.

Sanders confidently predicted that he'd win New York, the next state on the list. Instead, his purported momentum slammed into a wall there. A closed contest in Clinton's home state handed her a big victory.

The pledged delegate Hail Mary

The night of the New York primary, Weaver went on MSNBC to present the campaign's theory for how the election would end. It was a hodgepodge of arguments, centered on the idea that there were enough outstanding delegates that big wins from the Sanders campaign could still leave him with more pledged delegates than Clinton by the time the final contests occurred. The math then was prohibitive, but Weaver figured big wins out west and a close contest in, for example, Pennsylvania would put the campaign back on track to overtake Clinton.

But Weaver went a step further.

The superdelegate switcheroo

That interview was also one of the early instances of the Sanders campaign specifically suggesting that superdelegates could and should abandon Clinton to support Sanders, potentially setting aside both pledged delegate and popular vote leads by Clinton to offer Sanders the nomination at the convention. (In 2008, the superdelegates did the latter, ignoring Clinton's popular vote majority to vote for Obama.) After all, Sanders did better against Trump (who was close to locking up the nomination at that point) in national polling, and he did better with young people and independents.

What happened next is that Clinton continued a rout in the northeast, winning Pennsylvania by more than 10 points and simultaneously widening the pledged delegate gap. Any idea that superdelegates would have to pick between two essentially tied candidates moved even further away from reality.

Momentum, part two

This is where we are now. Sanders still, bafflingly, insists that he can take the lead with pledged delegates, which would almost certainly require winning California by at least 30 points.

His wins this month in Indiana and West Virginia ostensibly renewed his argument that he was the candidate with momentum headed toward the Philadelphia convention. "We've won two states in a row," a fundraising email the day before this week's contests in Kentucky and Oregon touted (ignoring a loss in the territory of Guam). "Imagine the boost in momentum that a pair of wins tomorrow could deliver for our political revolution."

He didn't get a pair of wins.

Ties count for momentum, too

If New York killed Sanders's first argument that he had the momentum, his narrow loss in Kentucky was probably more damaging. It came later in the process and came in a state that wasn't Clinton's home turf. A week earlier, he'd won adjacent West Virginia by a wide margin.

So, as our Dave Weigel wrote this morning, Sanders has reframed the outcome of the Kentucky primary, which he lost. "That’s 21 victories for us so far," Sanders wrote in an email, "plus three more virtual ties where the margin was less than one percent of the vote." Kentucky falls into the latter category.

It's up to California

At this point, Sanders has only one real argument left: A big win in California on June 7 will close both the pledged delegate and vote total gaps — and will reinforce Sanders's (highly debatable) argument that he's done better in later contests.

Update: In an interview with NPR on Thursday, Weaver made this latter point explicitly. "If we can substantially close the gap between Secretary Clinton and Sen. Sanders in terms of pledged delegates, he can go into the convention with a substantial momentum from having won the vast, vast majority of states at end of the process," he said.

That the vote in California occurs the same day as the New Jersey primary, where Clinton leads by a wide margin, doesn't matter -- nor does the fact that a week later, D.C. Democrats will almost certainly give Clinton another, final win. California, in essence, is Sanders' last chance to make his case.

Until a new strategy emerges.