[G]eneral-election trial-heat polls offer one obvious way to project the odds of winning with this or that nominee. But early in the nominating contest, they sometimes measure name identification as well as popularity. And as Democrats learned to their eternal sorrow in 2016, polls can get it all wrong, too, even late in the contest.

So barring definitive evidence that a particular candidate is decisively stronger as a Trump opponent than others, what do Democrats (and for that matter media observers) mean by “electability” in the first place? What prejudices do they bring to that discussion?

I would discern five basic ideas about the abilities that equip a candidate to do well in 2020:

This is the oldest idea, and probably the one most often embraced by media folks. Its basis in social science is the median voter theorem — the basic idea of which is that the candidate that puts themselves in the center can win the most votes:

“The median voter theorem as developed by Anthony Downs in his 1957 book, “An Economic Theory of Democracy,” is an attempt to explain why politicians on both ends of the spectrum tend to gravitate towards the philosophical center. Downs, as well as economist Duncan Black, who proposed the theory in 1948, argue that politicians take political positions are far as possible near the center in order to appeal to as many potential voters as possible. Under certain constraints/assumptions, Black says, the median voter ‘wins,’ and the outcome ends up as a Nash equilibrium.”

For obvious reasons, the median voter theorem is unpopular among ideologues in both parties who view “the center” as a place where the unprincipled and the timid gravitate. There are also some problems associated with defining “the center” in the first place, as Perry Bacon Jr. has noted:

“Would it help the Democrats in 2020 if they had a ‘centrist’ at the top of the ticket? All else being equal, it’s probably safe to conclude that candidates more removed from the mainstream of American political thought will do worse at the ballot box. There is some evidence, for example, that Republican presidential nominee Barry Goldwater (1964) and Democratic nominee George McGovern (1972) lost by larger margins than other factors would have predicted in their elections because of the ideological extremism of their voting records.

“But ideology is somewhat complicated to measure, particularly for people who haven’t served in legislative bodies (like Montana Gov. Steve Bullock, a potential Democratic candidate in 2020) or in any political office at all (like Trump.) Trump’s Muslim plan was perhaps the most radical idea proposed by any recent presidential candidate, but voters had trouble pinning the candidate down on a left-right spectrum before the election. Trump, according to the Pew Research Center, won the plurality of 2016 voters who described their views as ‘mixed’ and basically was even with Clinton among self-described independents.”

All sorts of candidates, moreover, can make a plausible claim to represent views within the political “mainstream.” As my colleague Eric Levitz has pointed out, the democratic socialist Bernie Sanders has been careful to identify himself with initiatives that are broadly popular:

“Bernie Sanders’s signature policies — Medicare for All, tuition-free college, a $15 minimum wage, a giant tax hike on the wealthy, and a $1 trillion infrastructure stimulus — all boast majority support in most surveys, and overwhelming bipartisan support in a few.”

Still, the strong tendency in many circles to view “the center” as ideologically moderate is an explicit talking point in the campaigns of Amy Klobuchar and John Hickenlooper, who are reasonably close to the Clinton-Obama “centrist” heritage that won four presidential elections for Democrats since 1992. And it’s certain to be a fundamental argument for Joe Biden’s candidacy as well.

2. Win crucial swing voters.

Another prevalent way to judge candidate “electability” is to frame 2020 as essentially a battle over a particular set of crucial swing voters to whom particular candidates do or don’t appeal. Without question, the most popular contestants for key swing voters next year are the Rust Belt white working-class voters — many of whom voted for Barack Obama in 2008 and/or 2012 — who helped Trump win Michigan, Ohio, Pennsylvania, and Wisconsin, and thus, the presidency in 2016.

Midwestern natives Amy Klobuchar, Pete Buttigieg, and Tim Ryan will talk early and often about their geographical and psychological solidarity with these voters. And Joe Biden’s alleged popularity in this sector will be a big deal for him, too; it’s no accident he is launching his candidacy with multiple events in his native Pennsylvania.

The belief that particular candidates will hurt Democrats among swing voters can become a factor as well. There’s a widespread if often quiet fear that Hillary Clinton’s gender and perceived cultural elitism killed her candidacy in the Rust Belt. This fear could hurt female candidates — particularly from the coasts — in 2020.

3. Run to the left

For many years there has been a persistent progressive dissent against the median voter theorem holding that on the contrary, Democrats need a more strongly ideological candidate to win an electorate left cold by centrists and the Washington bipartisan “Establishment” they embody. A subset of that dissent is what I’d call the “hidden majority hypothesis,” articulated here by Colin McAuliffe and Sean McElwee:

“[T]he path forward for Democrats needs to include mobilizing marginal voters, individuals who drift in and out of the electorate. These voters are overwhelmingly more supportive of progressive policies than individuals who consistently vote. According to Cooperative Congressional Election Studies (CCES) 2016 data, individuals who voted for Barack Obama but stayed home in 2016 preferred Democratic candidates in the House 83 percent to 14 percent (the rest preferred a third-party candidate). Ninety-one percent of those nonvoters support increasing the minimum wage to $12, 72 percent believe white people have advantages, 76 percent support a renewable fuel mandate, and 82 percent support an assault weapons ban.”

Another “run to the left” argument is that Democrats need an “insurgent” candidate to counter Trump’s anti-Establishment “drain the swamp” rhetoric while authentically representing the increasingly left-wing views of reliable Democratic voters.

The most powerful evidence for aiming the 2020 Democratic candidacy to the left, of course, is negative: Look what happened to Hillary Clinton. There is some evidence to support the emotionally powerful claim that Sanders would not have lost, sparing the country a Trump presidency.

4. Energize the base

A theory that sometimes overlaps with the “run to the left” prescription simply holds that 2020 will be a savage turnout battle and Democrats need a candidate whose identity will help mobilize the party base. This approach generally begins with an analysis of disappointing 2016 turnout by African-Americans and Latinos, and the importance of minority voters in states trending Democratic like Arizona, Georgia, and Texas.

This theory offers a particularly strong boost to Kamala Harris, an African/Asian-American woman. But Cory Booker relies on it as well, as does the sole Latino in the field, Julián Castro. It’s unlikely, but if Stacey Abrams were to jump into the race late, it would be significant that she’s devoted most of her career to registering and mobilizing young and minority voters. And speaking of young voters, candidates not named Joe Biden will likely boast of their appeal to Democratic-trending millennials and post-millennials.

5. Charm the electorate

There’s one theory of electability that denies ideology or geography or ethnic, gender, and racial identity offer the best Democratic path to victory. It’s all about a candidate’s charisma or “relatability” or likability, as Molly Ball recently suggested:

“As always, style and charisma are likely to matter to voters at least as much as policy papers and voting records. ‘These candidates are going to try very hard to distinguish themselves from each other, but their positions are pretty similar,’ says Democratic strategist Rodell Mollineau. ‘It’s going to be a lot more about the framing of your worldview than one specific vote.’”

Arguably Donald Trump won not because of policy positions or money or endorsements, but because he embodies his own sinister brand of charisma, based on an ability to understand and amplify the inchoate fury of white voters who feel threatened or left behind by technological or demographic or cultural change. Some Democrats think they need their own brand of rhetorical enchantment, and are attracted to candidates like Beto O’Rourke and Pete Buttigieg, who offer “narratives” instead of policy papers, and just seem refreshingly likable.

Any successful candidate is going to give off some of that sweet aroma of charm, but maintaining it regularly is tough.

So next time you hear someone boast of this or that candidate’s “electability,” it’s useful to discern what is meant, and which often-unstated premises are involved. It will make everything easier for Democrats if polls in January of 2020 show one candidate beating Trump by 20 points and the rest of them trailing the much-feared demagogue. They probably won’t be that lucky