I could understand if they’d prepared a package beforehand announcing Moore’s likely victory and then someone in the control room had hit the wrong button and inadvertently put it on the air. (Although why would they need to prepare a victory segment in advance?) But this clip doesn’t sound pre-packaged. The announcer cites some data about voters’ disapproval of Doug Jones’s platform based on, er, OAN’s “unofficial polling” that shouldn’t have been available until election night.

But it gets stranger. A barebones “Roy Moore wins” story also went up on the OAN website. It reads like a skeleton version of an article that would have been fleshed out with quotes, data, etc, on election night and then published. It’s easy to believe that the story was accidentally published too soon — or it would be easy to believe, if not for the clip above.

So what happened? Why would OAN gamble its credibility on a “Dewey Defeats Truman” announcement when it could have sat back and just tracked the returns as they came in? What makes it extra strange is that data nerds who do this for a living were oohing and ahhing over Jones’s numbers from the beginning after polls closed in Alabama. He was hitting the benchmarks he needed to hit in heavily Democratic areas and Moore was missing the benchmarks he needed in Republican ones. The New York Times’s election probability “needle,” which tracks who’s likely to win via algorithms pegged to the minute-by-minute returns, had Jones as the likely winner pretty much from the jump. Within the first hour of returns, stat-heads of various stripes were saying the race would be very close based on Jones’s numbers. All of that pointed towards exercising caution in calling the race. Yet OAN pounced. Why?

All I can figure is that they wanted to make a splash in right-wing media by sounding the trumpets for Moore long before the pros at Fox News did. The announcer specifically mentions that OAN’s president, Robert Herring, extends his congratulations to Moore on a hard-fought victory. Maybe Herring got on the phone to the control room, seeing Fox exercising caution and sensing a momentary vacuum in right-wing TV’s “tell me only the news that’ll confirm my biases” niche, and gambled on the fact that Moore would indeed win — which was the likely outcome. Then, after he did, OAN would have crowed that they called the race long before Fox did and therefore they should be red-state America’s go-to network on election nights. It was a big bet. They lost.

In lieu of an exit question, here’s a fascinating stat from lefty Matt Bruenig. There’s been hype all week about black turnout in Alabama on Tuesday night, for understandable reasons. It was an off-year special election, when turnout is typically low. If black voters had passed on the election, Jones would have been blown out. They were an absolutely necessary condition for his victory — but not a sufficient one. When Bruenig looked at the data, he found that blacks did in fact match the levels of turnout they reached for Obama’s two presidential elections, an impressive feat. But Democrats had been crushed in Alabama in 2008 and 2012 notwithstanding that turnout:

Jones couldn’t have done it without black voters but it was a massive shift among whites that actually won the election for him. Some who normally vote Republican stayed home, presumably repelled by Moore, while many others actually crossed over and voted for Jones. If that’s anything like a trend in next fall’s midterms, the GOP will be destroyed.