I’ve made the point in multiple contexts that percentages, such as those used to describe either antibiotic resistance frequencies or electoral outcomes, often don’t make much sense without the underlying numbers–what I’ve called the ‘denominator problem.’ Finally, a political pundit figures this out, and realizes its implications (boldface mine):

The results in both of the recent special elections were surprising to many journalists who thought that Democrats didn’t have a chance in what had historically been reliably Republican districts. But the conventional wisdom is founded on faulty analysis and an incorrect understanding of percentages.

The essential mathematical concept that a shockingly large number of people in politics fail to understand is the difference between percentages and raw numbers. Reporters see that Tom Price, a Republican, received 60 percent of the vote in 2016 in the Georgia’s Sixth Congressional District and quickly conclude that the district is conservative. Percentages, however, are only of limited analytical utility (for example, if a stock price increases by 10 percent, that means a whole lot more to somebody who has a billion dollars of that stock—a $100 million increase in wealth—than it does to somebody who only has $100 and gets a bump of just $10).

What the percentages masked in Georgia is that while the Democrat only received 38 percent of the vote in that district in 2016, that 38 percent equals 125,000 people. If Jon Ossoff had gotten 97,000 votes in the first round, we would now be calling him congressman (and we may yet have that pleasure if his campaign mobilizes the core Democrats in the district in June). As it was, Ossoff received 92,000 votes and nearly pulled off the outright win.

This situation of high Democratic turnout making seats competitive enough to flip will replicate itself across the country heading into the 2018 midterm elections. If Republican turnout does fall significantly—as it has in the special elections and as it did during the last Republican presidential administration—then Democrats have a golden opportunity. Presuming a Republican decline of 36 percent—as occurred in 2006 during Bush’s presidency—then Democrats only need to get, on average, 84 percent of those who came out in 2016 to vote again in 2018.

Some asshole with a blog might have made this point once or twice. That said, I’m not as optimistic: I think there will have to be a few ‘conversions’ as Democrats have done a piss-poor job of getting their voters out in off-year elections. But if they can, high Democratic turnout–and low Republican turnout–could put Democrats back in charge of the House. Now if there were only policies that could rally Democrats… thinking… thinking…