On Sunday, Bernie Sanders's campaign put out a press release titled, "Electability Matters."

"A new NBC News/Wall Street Journal/Marist poll," it read, "found that Sanders does better than Clinton against the leading Republican candidates by an average of 6 points in Iowa and a stunning 21 points in New Hampshire. Specifically, the poll put Sanders 13 points ahead of Republican frontrunner Donald Trump here in Iowa."

A fine point was put on it: "Bernie's substantial advantage over Republicans in the general election versus Secretary Clinton is another important reason that Democratic primary voters should choose him as our nominee," the campaign's Jeff Weaver wrote.

It was a continuation of an argument that Sanders has been making a lot of late. If Democrats are worried about keeping the White House, Bernie Sanders is the better choice for the general election.

The only problem is that the argument isn't a great one.

We can start with those numbers from Iowa and New Hampshire. When it comes to state match-ups, the extent to which the Democrat wins New Hampshire -- a state that's gone for the Democrats in five of the last six elections and which is worth less than 1 percent of the country's electoral votes -- doesn't amount to a whole lot. Iowa is slightly different, with a few more electoral votes and somewhat closer recent elections (although still a tilt toward the Dems). Sanders does do better against the leading Republicans in those states, but Clinton still beats Trump in both.

Iowa and New Hampshire both have a distinction which has been to Sanders's benefit in the primaries there, too: They're very white. Clinton has much more support among non-white voters than Sanders, which is why she's up big in South Carolina. Once we start talking about states like Missouri or Georgia, we can expect the calculus to shift.

There aren't head-to-head match-ups contrasting Sanders and Clinton in every state just yet. So we can approximate that by looking at how they fare in national head-to-head polling.

The most recent poll that asked about Clinton and Sanders versus Ted Cruz, Marco Rubio and Trump was one by Quinnipiac University, released late last month. Clinton beat Trump and Rubio and tied Cruz. Sanders beat Trump but lost to Rubio and Cruz. The spreads:

For the most part, there wasn't much demographic difference between the candidates. Clinton mostly did better with women and worse with men than Sanders.

Both did better among voters with college degrees, though Clinton did slightly less badly among those without.

However, the story here isn't that Clinton is the most electable. It's that it's a wash. Or, more accurately, that general election polling is not predictive this far out, as we keep trying to explain.

Why not? For one thing, Quinnipiac also found that Clinton and Trump were the only candidates recognized by the vast majority of Americans. Cruz, Rubio and Sanders all had more than a quarter of people who didn't know enough about them to have an opinion.

This is why we have elections, so people can get to know candidates and decide who they want to vote for. If I asked you who you supported for the presidency, Tony Randomguy or Jane Anonymous, who would you pick? The odds are good that if I said Democrat Tony Randomguy and Republican Jane Anonymous, you'd have at least some guideline of how you wanted to play it. Most people have ended up voting their party in recent elections. But then all we're tracking is partisanship.

When we last talked about the faultiness of general election polling, we showed this graph, which contrasts a Gallup national poll from January 2008 with the actual margin of the general election. In January 2008, Barack Obama trailed John McCain by two points, while Clinton trailed him by one. Clinton was therefore "more electable," in the parlance.

But Obama ended up winning by seven points. Meaning that the January poll 1) wasn't predictive and 2) didn't mean that Obama wasn't electable, clearly.

Voters care about electability. What Sanders is trying to do here, one can assume, is not really to argue that Democrats will lose with Hillary Clinton -- an argument that may have some resonance. It's to argue that Democrats can win with him, too. By pointing out that he runs ahead of Clinton in some scenarios, it tells people who know about Clinton that he, too, is a real candidate who could be viable in November.

It's now 2016, at long last, but the election is still a long time away. At the beginning of 2008, the Dow Jones was at 13,200. By Election Day, it was at about 9,300 and dropping. Lots can happen that isn't related to Bernie Sanders or Hillary Clinton. Trying to argue who would win the general election right now is like asking who would win between Tony Randomguy or Jane Anonymous. For which there's a poll below! See if you can predict the winner when polls close at 8 p.m. Friday.

I suspect you can't.

Vote in our hypothetical election!

This is a non-scientific user poll. Results are not statistically valid and cannot be assumed to reflect the views of Washington Post users as a group or the general population.