As laid out in Chapter 5 of my book, the NFL has to juggle a lot of concerns and interests when making their schedule. At least nominally, NBC’s Sunday night package is the league’s premier package where the biggest games go out to a true national audience, but Fox and CBS’ late afternoon games actually attract larger audiences, Fox pays more than NBC for NFL rights, and both networks want quality matchups of their own, not just NBC’s leftovers. CBS also wants every opportunity to showcase the bigger-market, higher-profile NFC teams rather than the lower-profile AFC teams they’re usually stuck with. And then there’s ESPN, which pays nearly as much as CBS and NBC combined yet continues to get a mediocre package of games for its Monday night package, much to their everlasting consternation.
The league’s flexible scheduling rules reflect the need to appease all these masters. The constraints surrounding the ability of games to change networks within the season are fairly complicated, and they’ve become more complicated (or we’ve learned about more complications) as time has gone on, especially with the introduction of the “cross-flex” and early-season flexing in the 2011 contract. I’m going to focus primarily on rules affecting the Sunday night package, since that’s where things are simultaneously most complicated and most predictable, but I’ll also cover some rules governing the league’s other packages.
First, a recap of how the NFL schedule in general is structured. The first 16 weeks of the season are awarded to either Fox or CBS as a “doubleheader” week where most markets will get games at both 1 PM ET or 4:25 PM ET. The other network has a “singleheader” with one game that kicks off at 1 PM ET or 4:05 PM ET. Generally, if a market’s home team is playing on the doubleheader the singleheader will show a game in the opposite time slot, while if a team is playing at home on the singleheader network, the doubleheader network will be blacked out in that time slot and will only show a game in the opposite time slot. The blackout rules tended to be ignored in Los Angeles in 2017 (which also had multiple “double singleheaders” where both of the market’s teams were on the same network in different time slots), and in Week 17, both networks get doubleheaders in all markets with no blackouts. (For the 2019 season, the NFL is also experimenting with relaxing the blackout rules so that the doubleheader network can air a game at the same time the local team is playing at home, but only twice per team for the season.) The main 4:25 game on the doubleheader network is the showcase game for the entire afternoon, and usually the league tries to get it games at least close to on par with the Sunday night game, and avoids scheduling any games for the singleheader network at 4:05 that would leach too much of an audience away from the main 4:25 game, mostly West Coast games that can’t air earlier or some other sacrificial lamb for markets with teams playing in one of the doubleheader network’s early games. But the networks also want at least somewhat attractive games in their singleheader and early doubleheader time slots they can show to markets without a horse in any of their other games, and don’t want the league using the flex process to leave their cupboard bare in the early time slot.
Most games not chosen for primetime or another special national showcase like Thanksgiving will be regionalized as part of the Sunday afternoon packages. Typically, Fox will show games involving an NFC road team, while CBS will show games involving an AFC road team. Some games can be “crossflexed” from one network to the other either when the schedule comes out (as has been known to happen to CBS’ Thanksgiving game most years since crossflexing started) or as the season progresses. In-season crossflexes are usually either to bolster one or the other network’s lineups, providing a network with a decent game they wouldn’t otherwise have at 1 PM ET or a game worthy of the main national slot the doubleheader network doesn’t have on its own, or to ease logistical concerns, such as that resulting from a Sunday night game being flexed out (although this is less of a direct concern as a result of a rule noted below) or a game being chosen for 4:25, either to free up one network or the other from having to commit too many broadcasting teams or to ensure the doubleheader network has more games at 1 PM ET than at 4:25. No more than seven crossflexes can go each way over the course of the whole season (supposedly, although eight games were slated to go from CBS to Fox even before the 2018 season started), and the number of crossflexes going each way must be exactly equal.
The NFL has a page laying out how flexible scheduling for Sunday night works, but it’s only been tinkered with since 2007 (and as a result, at one point still mentions that late doubleheader games start at 4:15 PM ET despite elsewhere giving the correct time of 4:25), leaves out several important factors governing flexible scheduling, oversimplifies what it does include, and in some places is just plain wrong (namely, it implies that only games in the Pacific and Mountain time zones are scheduled for 4:05 or 4:25, but while that may have been the case the first few years of the flexible scheduling regime, more recently the schedule has included games further east targeted for the feature slot, and in some cases for the late singleheader slot, which on occasion can even move back to 1 PM ET). I’m going to attempt to expand on and clarify the NFL’s page, presenting the pertinent facts about flex scheduling in an accessible manner.
There are really three different flex scheduling periods with different rules for each: the early flex period, which generally runs from Week 5 to 10; the later flex period, the period where flex scheduling was in place before 2014, which generally runs from Week 11 to 16; and Week 17. (In years where Christmas falls on Sunday or Monday and bumps either the afternoon schedule or the Sunday night game to Saturday, no flex takes place that week since games can’t be flexed from one day to another on two weeks’ notice, and to compensate the late flex period begins Week 10.) The NFL and NBC have been known to imply that there are only minor differences in how flex scheduling works in the early and late flex period, in some cases publishing schedules with the same flex-scheduling bullet for everything in Week 5 or later, but in fact the early flex period is really an emergency valve, intended to avoid situations like what happened in 2011, when a game between the Saints and Colts, with the latter winless with Peyton Manning injured, saw the Saints blow the Colts off the field and the ratings go into the tank, with the embarrassment to the league not helped by the fact the game was scheduled against the World Series for only the second year and lost to baseball in the ratings. (The Colts’ other SNF game that season, the intended Brady-vs-Manning showdown against the Patriots, ended up getting flexed out several weeks before the deadline to make a decision without a replacement being named at the time, the only time that has ever happened.)
Not only can the early flex only be triggered twice a season as the NFL’s page mentions, but CBS and Fox can protect one game each once the league starts thinking about triggering it, the same amount of time before the game as the league has to make its decision. That means the league generally only has the option of picking the third-most attractive game, second-most if it happens to be on the same network as the most attractive. That, coupled with its tendency to stick with the tentative if flexing out would result in only a marginal improvement, means it would pretty much have to be a Peyton Manning-esque disaster looming for the league to use an early flex. CBS and Fox also have the option to protect games in the late flex period, but all protections in those weeks must come in after Week 5 (Week 4 in years where the late flex period begins Week 10 as above) and of the six weeks in that period they may protect only five games each, leaving one week unprotected. Despite protection’s importance in determining what games may be flexed to Sunday night, not only does the league not acknowledge it on its flexible scheduling page, the list of protected games is not public knowledge – it leaked out for most of the first several years of the flex scheduling era, but I don’t believe there’s been a leak since 2014, the first year of the crossflex era.
Another constraint on the NFL is appearance limits. No team may appear more than six times across all of the league’s primetime packages on NBC, ESPN, and NFL Network, and only three teams can appear more than five. Thursday night games simulcast on Fox and NFLN count towards the total, but not Thanksgiving or other fully-national games on Fox and CBS that are part of their Sunday afternoon packages. (I would expect London games airing primarily on NFL Network to count towards the total, but in 2018 pretty much everyone that wasn’t me seemed to assume the Eagles were not maxed out on primetime appearances despite having a London game on NFL Network that would have maxed them out, and two games were flexed to Sunday night that didn’t seem to make sense for Fox not to protect them unless Fox felt the need to protect Eagles games.) In addition, no team may appear more than four times on NBC, including the Kickoff and Thanksgiving Night games. There are also rules regarding the balance of games taken from Fox and CBS to populate Sunday nights over a period of several seasons, and while I don’t know exactly what those rules are and in any case there aren’t any hard-and-fast limits on the scale of a single season, there have been times when the need to address balance issues has forced the league to flex in a specific network’s game late in the season.
In a perfect world, flex scheduling would serve the purpose it was promoted as serving when it started out back in 2006-07 and is still reflected on the league’s flex-scheduling page, to “ensure quality matchups on Sunday night in those weeks and g[i]ve surprise teams a chance to play their way onto primetime”. To some extent, this has happened (the game that eventually replaced that Pats-Colts game in 2011 was one involving the Detroit Lions only a few years removed from going 0-16), and generally when the league announces a flex they will try to bring in a game involving two good teams and usually the best game on paper that wasn’t protected and doesn’t run into other constraints (and almost always picks a game involving better teams than the game being flexed out), but ultimately the league is still in the business of delivering ratings to NBC, and certain teams (and players) are more likely to be flexed in and less likely to be flexed out than others. Notably, the Dallas Cowboys have never been flexed out of Sunday night, despite repeatedly being in games that absolutely would be flexed out if one of the teams involved wasn’t America’s Team.
On paper, because the league wants to get NBC the best possible games CBS and Fox’s influence on the process should be limited to the protections, but as above they pay the league plenty of money too, and they aren’t above politicking to keep high-value games even if they missed their chance to ensure they kept them by protecting them (as was notoriously the case with Patriots-Broncos in 2011 when Tim Tebow was having his magical run with the Broncos). There’s no evidence of any restrictions on having both halves of a divisional matchup on NBC (and in fact one year, one half of Cowboys-Giants was scheduled for NBC and the other half was the Week 17 division title game, resulting the next year in one half being played before the flex period and the other half on Thanksgiving, both on Fox), and CBS and Fox have regularly protected games where the other half was scheduled on NBC, but a questionable flexing decision in 2013 suggests the league does try to avoid putting both games between division rivals on NBC, or at least to avoid flexing both in. On the flip side, the league can put pressure on CBS and Fox as well; notably, in 2013 CBS protected a game between the Chiefs and Broncos, two teams with two of the best records in football, but then “voluntarily” relinquished it when the time came to make a pick, because they were the singleheader network that week and so, with the game being in the Mountain time zone, the game would otherwise be pinned to 4:05 PM ET with limited distribution.
As above, the league tends to stick with the tentative if the only alternatives available are only marginally better (or, on occasion, involve teams that would send ratings into the tank regardless of their record), which serves the purpose of minimizing disruption to fans, teams, and stadium staff. The league is supposed to announce whether or not a game will be flexed out and for what twelve days in advance (so the Tuesday after the games two weeks before), usually publishing a press release with any schedule changes on Monday, but has been known to bend that by a day or so on rare occasions if the politicking is particularly intense (as with Patriots-Broncos 2011 again). It was reported in 2017 that from that point forward flexes would be one-to-one – regardless of what network a game being flexed out would normally go to, it would always go to the network that had their game flexed in – but while the league sometimes followed this rule in 2018 (in one case even when it would have made much more sense for the flexed-out game to go to its normal network) in other cases they sent the flexed-out game to the network it would normally be on, so as always, the league basically can do whatever it wants. (I would assume this counts as a crossflex if the game is going to a network it wouldn’t normally go to, but haven’t looked into how this affected the balance of crossflexes in 2018 – 2017 saw no flexes). When this rule is followed, it eases the logistics for both networks, who don’t need to commit any more or less broadcast teams than they otherwise would.
When analyzing the schedule for the chances of a flex, I always start out looking at the tentative game and what the records of the teams in it are, and how valuable a game it looks like it might be as a result. Once we pass the week where protections are due I try to guess what those protections are, which can be difficult because some of the protection decisions in years where they’ve been leaked have been rather surprising and/or puzzling, and it’s hard to know how to balance raw quality vs. name and ratings value; I often list several options for games that might have been protected, and hedge my bets by listing multiple weeks a network might have left unprotected. I typically stick with those predictions unless and until any hard information about protections comes out that contradicts them. Then I look for all the unprotected Sunday games involving only teams at or above .500; I might look at teams a game below that mark if the tentative looks particularly putrid and the pickings of over-.500 games are particularly slim. This does mean I’m somewhat buying into the league’s original “play your way into primetime” rhetoric and there have been occasions where the league will blindside me by flexing in a game that wasn’t even on my radar, but those occasions tend to be inexplicable even from a name value perspective, and I do factor considerations beyond raw record into my analysis and predictions (like never picking the Cowboys to be flexed out).
That analysis doesn’t start in earnest until a week before any flexes are due, when I look at what the potential record comparisons (and in some cases, narratives) might be after the coming week’s games and what chance any of the leading alternatives have of overcoming the tentative game bias or each other, and if one game’s position is secure enough (usually the tentative being strong enough to overcome any challenges) I’ll make a prediction right then and there. If the tentative game looks particularly attractive, and there are no unprotected alternatives even nearly as good, I may even make a prediction two weeks in advance. Otherwise I’ll make a “last-minute remarks” post on Monday reassessing the situation after Thursday and Sunday’s games, whether or not the Monday night game might have any impact on the calculus, and making a final prediction then, before looking to see if the league has already announced any changes or lack of change.
As a whole, Week 17 has increasingly become an entirely different beast from the rest of the regular season over the course of the flex-scheduling era as the league has sought to maximize fairness in determining the last playoff spots, between the move to all division games, the way it’s scheduled, and the numerous rules mentioned above that get thrown out or completely changed. The entire schedule gets set only six days in advance, often being announced on NBC itself during the Week 16 Sunday night game, and the overarching concern is making sure teams in the late doubleheader slot have something to play for no matter what the results are in the early slot. As mentioned, both CBS and Fox get coast-to-coast doubleheaders with no blackouts.
NBC gets the only primetime game, but as they’re airing “Game 256” and so their game needs to have the same implications no matter the result of all the other games, their options are fairly constrained. As such, until 2018 all of NBC’s games since the advent of the all-division-games era had been effective division title games (all of them in the NFC and most of them involving the East or North divisions, as it happens), even if the loser might still make the playoffs as a wild card. 2018 changed that with NBC airing a game between the Colts and Titans with the winner set to receive the last AFC wild card spot. In theory, if the standings set up just right NBC could also air a game where one team is guaranteed to win the division or claim a wild card berth with a win but would fall behind the winner of another game with a loss (so either a game ahead of two teams playing each other they’d lose a tiebreaker to, tied with two teams they’d lose a tiebreaker to, or a half-game ahead), but the NFL would probably prefer to avoid that situation where the other team has nothing to play for, and probably wouldn’t even consider it if the other team did have something to play for but only if the game is played earlier in the day. (See here for the league’s tiebreaking procedures.) To maximize flexibility, there are no protections this week, and starting in 2016 Week 17 has been exempt from team appearance limits, so the league doesn’t need to worry about maxing any teams out they might want in Week 17 when making flexes earlier in the season.
In 2017, when no game meeting any of these criteria came up, the NFL didn’t schedule a Sunday night game at all, and with teams in both Western divisions vying for wild card spots, scheduled substantially more games in the late slot than in the early one, something they wouldn’t do for any other week, so committed are they to ensuring as many teams have something to play for as possible. Honestly, it’s a bit of a miracle they were able to avoid that fate each of the previous seven or so years, and I wouldn’t be surprised if the league makes radical changes to how Week 17 is scheduled during the next contract negotiations.
Whoof. That was a lot of words for something that, given all the restrictions and the league’s tendency to stick with tentative games if there isn’t an overwhelming case to flex it out, really only comes into play once or twice a year outside Week 17, and sometimes no games are flexed out at all (though that’s only happened in 2014 and 2017). But now you have some idea of the thought process that goes into scheduling decisions, why your game’s time might have changed on two weeks notice, why a certain game was or wasn’t flexed out, why NBC is still showing the putrid Cowboys (or alternately, the world-beating Cowboys playing some jobber team) this week, and why the Week 17 schedule is so funky.