Fact-Checking Roger Goodell's Claim that Thursday Night Football Is "Better"

"Almost by every barometer," NFL commissioner Roger Goodell said last month on Colin Cowherd's radio show, "the quality of the games is better on Thursday night." The commissioner doubled down on Thursday Night Football in his annual state of the league address Wednesday, saying, "We've seen high-quality football on Thursday night." But popular opinion, and every barometer VICE Sports could find, disagrees.

One can almost picture Joe Pesci reacting to Goodell's assertion that Thursday games are better: Better? Better how? Better like a clown? Did he mean "better" than all the other regular-season games, or "better" than TNF games used to be?

Either way, a Twitter search for "Goodell TNF" turns up a host of references to Sean Spicer, "alternative facts," and the Why You Always Lying Guy.

Lets give the commissioner the benefit of a doubt for a moment: What barometers could we use to determine whether the TNF games are "better"? Let's start with the elements of a good game: a compelling matchup, a competitive contest, a close finish.

(Note: Across all of the following numbers and charts, I've excluded the extra Monday Night Football game in Week 1—which this season might have been the worst game in modern history—and the Thanksgiving games, which aren't truly part of the Thursday Night Football schedule.)

First, to compare matchups we can look at two numbers coming out of Las Vegas: the point spread, per Pro Football Reference, and the over/under. Both serve as useful barometers for people's expectations for an upcoming game—a low point spread means the game is expected to be close, and a high over/under means the game's expected to be high-scoring.

The average point spread of 2016 Thursday night games (4.47) was slightly lower than the Sunday or Monday slates. In theory, that's good; it means the wise guys expected close TNF games this year. But the average over/under is two full points lower than the combined average of the other three slates—Vegas reduced the spreads partially because they didn't think teams would score much. And those games, generally speaking, are not considered as exciting.

Maybe Goodell meant the other sense of "better," as in this year's Thursday night games were better than last year's Thursday games. Nope, or at least not according to Vegas:

When talking anticipated matchups, of course, whether the teams are actually any good is a huge factor—and per Pro Football Reference, the 2016 TNF slate featured fewer winning teams than any other showcase series. Just like last year, this year's Thursday slates featured just 14 team-games played by squads that went on to have a winning season.

It's no wonder so many TNF games turned into snoozefests.

Even throwing out Thanksgiving's late Pittsburgh/Indianapolis blowout, four Thursday night games had a margin of 14 points or more at the start of the fourth quarter, more than either SNF (three) or MNF (two). But in a year when the single-season fourth-quarter comeback record was broken, how early was it truly safe to turn off the TV?

On Thursdays this year, it was pretty early—at least relative to every other night. I used Pro Football Reference's Win Probability models to find out the point where each of the games hit 99 percent WP for the winner and stayed there. Though many snoozers were effectively over well before then, once the trailing team drops below a 1-in-100 chance of making a comeback and shows no sign of trying, you're well and truly wasting your time:

The chart above shows the average portion of unwatchable seconds across the whole season, with each pie representing 60 minutes (3,600 seconds) of action. Across 2016, the last interesting second was wrung out of the average Thursday night game with 7:31 left to go in the fourth quarter. That's far worse than the Monday or Sunday slate of prime-time action, and more than twice as bad as last year's TNF series!

And this is a charitable definition of "watchable." The Arizona Cardinals' Thursday night beatdown of the San Francisco 49ers technically hit 100 percent win probability with 5:56 left to go, but a meaningless Blaine Gabbert touchdown pass briefly cut that down to 96.2 percent before topping out for good with 1:56 left to go.

As you might expect, this is all reflected in the TV ratings, the ultimate measure by which the cable-money-driven NFL must judge itself. The two Thursday Night Football packages, per Ad Age, finished fourth and fifth among the six national broadcast slates the NFL offers. (The NFL also livestreamed Thursday night games on Twitter for the first time this season; per Variety, it drew an average audience of 243,000 viewers, compared to 15.4 million TV viewers.)

So no, Commissioner Goodell, Thursday Night Football games aren't high-quality, and they aren't getting better. They weren't better this year than they were last year, and they weren't better than either of the other two prime-time slates. For every really good game (like the season-opening Super Bowl rematch, which dramatically finished 21-20), there were a handful of stinkers like Texans-Patriots (0-27) and Rams-Seahawks (3-24).

We have run out of barometers by which to judge Thursday games; all of them have pointed toward a cloudy future. If Goodell has any others by which the TNF forecast is sunnier, we hope he'll be kind enough to share at the next annual press conference.

Want to read more stories like this from VICE Sports? Subscribe to our daily newsletter.