Should NFL Players Actually Want an 18-Game Season? Surprisingly, Maybe.
by Zach Binney
NFL owners have made their interest in extending the regular season to 18 games clear. The chatter has ratcheted up as we approach the expiration of the current collective bargaining agreement (CBA) governing NFL labor relations in 2021 – the point at which owners and players could negotiate this type of change.
NFL players and their union, the NFL Players’ Association (NFLPA), are generally opposed. They worry 18 games will result in more injuries and worse health for players (true), shortening their careers and degrading their earnings (less clear).
As a negotiating stance this makes perfect sense. But the 18-game season could actually put more money in players’ pockets without having as large an effect on their health and careers as most people assume.
Injury Consequences of an 18-Game Season
First let’s stipulate that all else equal, playing more regular-season games will cause more injuries, more hurt players, more wear-and-tear on bodies, and potentially more lifelong health problems such as arthritis and neurological diseases from two more weeks of hits every year. Nobody is arguing otherwise.
But there is also a concern that players would suffer financially due to injury-shortened careers. I’m not so sure about that.
Let’s start by looking at how (reported) injuries occur over the course of a season, per the NFL’s injury report.
Contrary to popular belief, injuries are not more common later in the year as wear-and-tear builds on players’ bodies. In reality, injuries occur at about the same clip throughout the regular season – 140-150 new issues per week, with about 15-20 of those being long-term issues sending players to the injured reserve list (IR).
The most dangerous time of the year is actually before the season starts: every year more than 300 injuries occur during training camp and preseason and linger into the regular season; 130 of them send a player to IR. (There are many additional injuries that occur and resolve during the preseason that we never hear about.) Other studies have shown that injury rates are worst at the beginning of training camp when players are first coming back and ramping up their training after the offseason. Survivor bias is also at play here – injury-prone players get weeded out early, and the players left by the regular season are on average less fragile.
The take-home message here is we should expect about 300 extra injuries per year from an 18-game season, with 40 of these severe enough to send a player to IR. But if a player can get two more game checks without aging another year and going through the gauntlet of another high-risk ramp-up period in training camp, that’s not a bad deal.
Financial Consequences of the Extra Injuries
Not every injury – even if they’re the same type – is equally costly. Timing is everything.
Players don’t just get paid per game. They get stuttered compensation through signing bonuses, roster bonuses, workout bonuses, performance incentives, and a variety of other creative, mostly team-friendly structures. Sometimes they get paid the same when they're hurt, sometimes less. Portions of some contracts can be guaranteed for injury, but that’s far from universal. More and more contracts also involve mechanisms like injury splits and per-game bonuses that result in players getting less money if they get hurt.
All this means that getting hurt late in the year is usually better for players financially. Most of the extra 300 injuries would affect at most 1-2 game checks, versus those suffered in training camp that can affect a full season’s earnings and possibly cost a player a “credited season” towards retirement benefits (more on that below).
That said, some of these extra 300 injuries will be major issues that could either bleed over into a future season, affect future contracts, or even derail a player’s career. How big a problem would this be?
Financially-Catastrophic Injuries from an 18-Game Season
We could start by looking at how often players suffering serious (IR) injuries late in the year start the next season injured or no longer play. We can then compare this with retirement proportions for age- and position-matched player-seasons without a serious late-season injury:
Of 510 IR trips beginning in weeks 13-15, 24.9 percent are followed by player “retirement” – that is, the player’s last regular-season game was the year of the injury.
In any given year, though, 14.4 percent of age- and position-matched players who play in at least six games but do not suffer a late-season IR injury also retire. The difference – 10.5 percent – is an estimate of how many of the IR players retired due to their injury.
With 40 extra serious injuries per year, about four (10.5 percent) might precipitate a player’s retirement. (We ignored less serious injuries as these are less likely to be the cause of a player’s retirement. Indeed, retirement proportions are similar between players suffering minor versus no late-season injuries.)
Further complicating things is that for many players, the bulk of their money will come on a single post-rookie deal not negotiated until their fourth to sixth season. In addition players need three “credited seasons” – basically, one in which the player is on the active roster for three or more regular-season games – to access many retiree benefits. Thus anything that either a) prevents a player from reaching his second contract (or shrinks its size) or b) prevents a player from reaching three credited seasons has an outsized impact on his finances.
How many retirements following late-season injuries happen at the most financially-catastrophic times?
About 13 percent will happen during a player’s second year – just before he gets access to many retirement benefits. About 15 percent happen during a player’s fourth season – when many are negotiating their post-rookie deal. Of the four injury-induced retirements annually, about 1-2 each year could be financially catastrophic.
Injuries can also have bad financial effects short of forcing retirement, though. Some portion of the 30 non-retirement serious injuries could affect a player’s future contracts – for example, a player coming off a late-season ACL tear is likely to command less in negotiations than a “fully healthy” player due to a perceived increased risk of future injury or performance declines. About 50 percent of non-retirement late-season IR trips happen between a player’s third and sixth season (data not shown). A rough upper-end estimate, then, is 30 non-retirement late-season IR trips x 50 percent = 15 serious injuries annually affecting a player’s second contract.
Would an 18-Game Season Shorten Careers?
As established above we are not likely to see an outbreak of career-altering injuries in games 17 and 18 that cause player career lengths to crater. However, it’s possible that two extra games of general wear-and-tear – especially for high-demand positions like RB – could shorten some players’ careers absent a specific reported injury.
The NFLPA recently claimed that an 18-game season would reduce the average career length 15 percent from 3.3 years to 2.8 years, a major concern as it would drop the average below the three credited seasons needed for many retirement benefits. There are several problems with this.
First, among players who retired from 2012-16 the average NFL career lasted about 5.3 seasons, with a median of four seasons (the number for credited seasons is slightly lower due to injuries causing players not to be on the roster for at least three games). This is updated data but consistent with an earlier FO investigation. This is longer than either the NFLPA or the NFL claimed in the above article. The NFLPA in particular has understated the length of careers as a negotiating tactic. The numbers they quote (around 3.5 years) represent the average experience of all active NFL players. That’s like claiming U.S. life expectancy – which is about 77 years – is in fact 38 years, the average age of Americans alive right now.
Second, we can estimate the effect of an 18-game season on career lengths by recognizing that we already have a group of players who play longer seasons: guys on repetitive playoff teams. A proper comparison is tricky because we have to adjust for skill: players on playoff teams are on average better and expected to have longer careers. Here is a first attempt at this sort of analysis, which roughly adjusts for skill by looking only at players with a First-Team All Pro appearance:
Among All-Pros there is no association between the number of games per year they play and their career length. Splitting by position reveals similar results but suffers from small sample sizes.
An important next step would be to look at a broader population of players and adjust for skill using a comprehensive rate stat not tied to a player’s career length (e.g. EPA, WPA, or DVOA), but that’s beyond the scope of this post. Whatever we do, though, probably won’t show the average career dipping by 15 percent, or to 2.8 years.
The Details Matter
There are several ideas for how to implement an 18-game season. Some of these are better than others:
1. Add two games to the end of the regular season, no other changes: This is the “default” plan we’ve assumed in all our analyses above. The major downside is it adds two more weeks of work to a player’s year, reducing their offseason rest and recovery time by the same amount. This could have some negative health and career effects.
2. Add two games to the end of the regular season, cut two preseason games (and two weeks of training camp/preseason time): This would have the benefit of keeping players’ offseason rest and recovery time the same. The major potential downside is you shorten the ramp-up period before the regular season, causing players to be brought back up to in-season shape more quickly. A shorter, more intense training period could cause more injuries, but how this would weigh against two fewer weeks of practices and preseason games in which to get hurt is unclear.
You could also cut two preseason games while keeping the training camp and preseason period the same length – that’s a fine plan.
3. Add two games to the end of the regular season, add a second bye week: Could a second bye week mitigate some of the damage of an 18-game season? It would give players some extra badly-needed recovery time, but would it reduce injuries?
Based on data from the past five seasons, injuries are just as common the week after a bye as two weeks before. (We can’t use data from the week before a bye because injuries that occur just before the bye but resolve in the week and a half before the next injury report don’t get reported.)
It's doubtful a second bye would reduce serious injuries, and it would cut the players’ offseasons by yet another week. But if players want the extra in-season rest they should ask for it.
4. Add two games to the end of the regular season, but only allow players to play up to 16 games: No.
Even as a big fan of chaos…no. The main benefit would be that players play the same game load they always did, which will prevent some injuries in some players. But everyone will still be subject to two more weeks of practice injuries while getting much smaller salary increases because of simultaneous roster expansions. Speaking of which...
5. Roster Expansion: The value proposition for players of two extra games is they get more money for the added risk of injury and wear-and-tear. If the extra games are coupled with expanding regular-season rosters, the benefit is reduced. Furthermore, their extra risk probably wouldn't change much: bigger rosters could cut in-game exposure in certain circumstances (such as bigger DL rotations), but coaches are going to use their best players as much as possible. Less benefit, similar risk, bad deal.
Two more games is worse for players’ bodies, but we are not likely to see an outbreak of financially-catastrophic injuries or cratering of career lengths. There would be about 300 more injuries each year (40 of these serious enough to land a player on IR), but we estimate just four additional retirements each year from these injuries, with 1-2 coming just before a player secures retiree benefits or a second contract. There could also be up to 15 injuries per year that affect a player’s earning potential for a second contract without forcing retirement. We did not find an association between regular-season + playoff games played per year and career length for All-Pros, though it’s possible other players could suffer shortened careers from general wear-and-tear not associated with a specific injury.
Balanced against this is the fact that, assuming per-game pay doesn't drop, hundreds of players would benefit financially from the opportunity to accrue two more game checks before aging another year and rolling the dice to make it through another dangerous training camp ramp-up period. Assuming a 10-15 percent increase in the players’ total revenue pool it’s quite possible the average player would benefit financially – with the caveat that roster expansions would dilute that benefit.
To counter any career-shortening concerns the NFLPA could consider negotiating for shorter rookie contracts, restrictions on the franchise tag, or tweaks to how players accrue credited seasons for retirement (for example, granting players a bonus credited season if they suffer a severe injury in the last two weeks of the season).
Regardless, the NFL and/or NFLPA would be wise to engage with analysts to help them think through and model these issues, with the goal of a winning outcome for both sides.