It's fun to talk about the prospect of Mitt Romney running for president again, for a variety of reasons. We will grant our colleague Aaron Blake's assertion that Romney isn't actually popular enough to win in 2016, at least at this point. But the idea of Romney making a third consecutive run piqued our interest. How's that worked out in the past?

And the answer is: Pretty well, actually. Of the presidential candidates who have run three times and actually earned at least 20 percent of the vote in a general election, four of the five of them won two of the three races. One, poor old William Jennings Bryan, lost all three.

But notice that Mitt Romney doesn't actually meet that standard. Since "running for president" is a fairly vague quality which encompasses a lot of the nation's political fringe, we had to set a tighter guideline. After all: Did Thaddeus McCotter run for president in 2012? Does that count? So we applied the standard above, which includes being on the general election ballot and doing decently. And we went back all the way, too. George Washington went two-for-two.

That means that Romney would be running a second time, under our metric, which works just fine as a comparison anyway. So we'll start with the good news. Candidates who ran for president twice and only twice had the highest success rate, winning 63.5 percent of the time. Next was those who ran four times, who won, on average 62.5 percent of the time. Those who fared worst were people who ran once -- the category into which Romney currently falls.

Let's get the pretty pictures out of the way. Meet us down the page a bit.

OK. So the first thing we will add is that we're talking about some small sample sizes here. There have been two candidates that ran four times under our definition: FDR, who won all four races, and John Adams, who won once. So the 62.5 percent number doesn't tell us a whole lot. And, likewise, of course candidates who only ran once lost more often. If they'd won, they'd have usually run again!

Now let's game out Romney. Of those candidates who've run twice -- which would include Romney if he ran in 2016 and got the nomination -- five lost both times. Nine won once (say that ten times fast) and 12 won both races. Romney would want to be in that middle category. But there's bad news, as shown in the third graph above. Of those nine, eight won the first of the two races, not the second. The one candidate who lost his first bid and won his second was William Henry Harrison, who died of pneumonia a month after taking office. Not a great role model.

But we come back to the point above. If candidates lost a race and then won a race, they usually run for a third time. (Unless they die of pneumonia. Wear a coat to your inauguration!) So what about those candidates? The very good news for Romney is that if he were to win in 2016, no candidate that has lost a race and then won a race and then run for a third time has lost. He's basically guaranteed to be a two-term president -- if three previous instances can form a guarantee.

If Romney runs and loses in 2016, it's still not over. He could run again in 2020, and, if he's anything like John Adams, win. Or he could be like William Jennings Bryan, and lose a third time. (No candidate has ever run four times and lost four times.)

One last bit of bad news, though. On average, candidates' share of the popular vote drops between their first and second runs -- and drops substantially further between the second and third. So depending on how you look at Romney's possible 2016 bid, it's not likely he'd do better than he did in 2012, although the average drop is only 0.79 percentage points between campaigns one and two.

All of which is to say: We hope Mitt Romney runs in 2016. We need more data in our sample.