The New Overtime Format

The NFL finally admitted its overtime format was broken when it revised the rules to allow a team a possession to match or beat a first-drive field goal. Although I have been shouting from the rooftops that OT was broken, I was not strongly in favor of any particular fix. I thought that adjustments such as returning the kickoff line to the 35 from the 30 would do a lot to reduce or eliminate the advantage the coin-flip winner had, but if the NFL wanted to keep it's sudden death format, there weren't many other good options.

The solution the NFL settled on strikes me as too complicated, and it only partially addresses half of the problem. I'm reminded of my favorite non-sequitur (all too common in politics and public policy):

We have a really big problem. We must do something.
This is something, so we must do this.

Well, the NFL has certainly done something.

There were really two issues with the previous system. One, it favored the coin-flip winner dramatically. And two, it was increasingly likely the coin-flip loser would never have a chance to play offense, while the coin-flip winner would never need to play defense. The recent fix only partially addresses the second problem because a first-possession TD would still end the game before the other team gets the ball, and it may not affect the first problem much if at all.

As I laid out in my previous analysis of the new format, the coin-flip winner will still enjoy a substantial advantage. In any situation where the teams either both fail to score or both score FGs on their first possessions, the rules revert to the old sudden death format in which the coin-flip winner has a 3:2 advantage. The only combination of events that reduces the advantage is when the coin-flip loser scores a touchdown following a first-possession FG by the coin-flip winner.

In the event tree below, the math works out to where the coin-flip winner would have a 56/44 advantage compared to a 60/40 advantage in recent years. That would be an improvement, but the truth is we can't really predict what the new advantage will look like. The probabilities I used are educated estimates, but they depend on how coaches react under the new rules.


Known for their irrational risk aversion, NFL coaches are now faced with a myriad of risk-reward decisions. It's clear that a coach needing a FG to stay alive will use all four downs to keep his drive alive, but there are many more situations that aren't as obvious.

On the first possession in OT, what will a coach do if faced with a 3rd and 7 on the opponent's 29? Will he run it up the gut for 4 yds to set up a FG, which may or may not hold up to win the game? Or will he be thinking touchdown and call for a pass, going for a fist down but risking a sack or turnover?

What about the coach of the team with the second possession? Knowing he needs a FG to stay alive, what would he do in the same situation? His decision is very different because even a successful FG gives the ball to the opponent under the old sudden death rules, rules that heavily favor whoever has the ball. Maybe he should even go for it on 4th and 7.

Coaches don't like the new rules because they are unequipped for these kinds of decisions. There is no existing conventional wisdom for their decision-making intuition to fall back on. I'll probably have a field day second-guessing them next season, but I'm not sure I like the complexity.

I think the entire OT discussion is missing one very important underlying problem. The NFL has certain requirements about its OT format it wants to preserve. It wants a quickly-resolved, exciting, and fair game. It wants the format to resemble 'true' football that incorporates all three phases of offense, defense, and special teams. These requirements dictate what kind of OT rules are necessary, and that's the way it should work. But there is one unspoken requirement the NFL seems to have that makes no sense: Overtime must begin with a coin flip.

Wait... Why? Why must there be a coin flip? Why should something wholly arbitrary and entirely random dictate all the rules that follow? Why should the flip of a coin bestow an advantage of any size to either team? Remove the coin flip from the process, and things become much clearer.

The unfortunate reality is that no system will be perfectly fair. Some team will likely have an advantage over the other. In fact, granting one team a significant advantage over another is a celebrated part of the NFL. For most playoff-bound teams, the main question in the latter part of the season is who earns home field advantage in the playoffs. The home team wins 57% of the time--in all games, not just the tiny few that go into overtime.

The question is, how do we determine which team gets that advantage? In baseball, the team that bats second has an advantage in the 9th inning (and in extra innings) because it knows what is needed to win. If the first team doesn't score, the second team knows it can adjust its strategy to play 'small ball,' increasing its chances of scoring just one run at the expense of possibly scoring more. If the first team scores two runs, the second team knows it needs a rally and would adjust its strategy accordingly. For baseball fans, this edge is accepted as a natural part of home field advantage.

The NFL could eliminate the OT coin flip and assign first possession to the home team. Or, it could simply use the result of the coin flip to begin regulation. Why does there need to be a second coin flip at all?

In either case, knowing ahead of time who would start OT with the ball would indirectly mitigate the advantage of first possession. If I'm coaching a team that just scored a TD to tie the game late in the 4th period, and I know the other team would have the first possession in OT, I'd want to go for the two-point conversion. I'd probably rather take my chances on the 2-yard line than have to kick off to start OT, under either the old or new format. The same principle applies to a team that can decide between attempting a game-tying FG or going for a TD to take the lead.

Anyway, there are millions of NFL fans, and there are just as many ideas for fixing OT. (And I think half of them are collected in comments at this site.) I think the 28-4 vote by the teams to change OT was really more of an admission the old system was broken rather than an endorsement of the new format.

  • Spread The Love
  • Digg This Post
  • Tweet This Post
  • Stumble This Post
  • Submit This Post To Delicious
  • Submit This Post To Reddit
  • Submit This Post To Mixx

33 Responses to “The New Overtime Format”

  1. Anonymous says:

    I think there should be a coin toss, but the winner should start with the ball at their own 20 (no kick-off). Then the first to score 4 points in OT wins. This would obviously extend the length of OT (relative to past seasons) and it may cause an uneven amount of possessions, but it fairer and simpler than both the rule from last year and the newly accepted one.

  2. Anonymous says:

    but it fairer and simpler than both the rule from last year and the newly accepted one

    should be

    but it's fairer and simpler than both the past rule and the new one.

  3. Anonymous says:

    Brian,
    For OT, how much of a difference would it make in the expected win/loss probabilities if the kick off point was advanced 5 yeard to the 35 yard line?

    If thats not enough to drop from 60-40 towards 50-50, then how about kicking from the 40?

    Or what if the winner of the coin flip can decide to get the ball at the 20.

    - Jets Fan

  4. Brian Burke says:

    Jets Fan- This should answer your question. Unfortunately, once you get past the 35, you're just guaranteeing a touchback.

  5. Anonymous says:

    Yay a post. I thought you died. It's been like almost a whole week

  6. Brian Burke says:

    It's my off-season too. :) Besides, there's the NCAA Tournament going on.

    Actually, you can expect about 1 post/wk in the off-season, maybe more often as the draft approaches. I'm working on other projects at the moment, including the player stat pages. Right now, I'm working on advanced stats for defensive players. Here's a sample, just so you don't think I've abandoned my post. I'll explain what +WPA and +EPA mean in a forthcoming article.

  7. William T says:

    My NFL overtime suggestion:

    Play sudden death, but no placeholding field goals (drop kicks only).

  8. Anonymous says:

    I liked the auction idea. Each team submits a bid for a starting field position. The team that picks further back gets to play sudden death from there.

  9. Kevin says:

    You may be right that the home team in baseball has a statistical advantage from the 9th inning on, but it's worth noting that the visiting team has certain strategic advantages as well. For example, if they are tied facing a man on third with zero or one outs, they can bring the outfield in, bring in a pitcher who has a high strikeout rate, and/or change their lineup to improve their defense knowing they have to get outs. These are adjustments that the home team may not feel confident making in similar situations.

    In other words, the home team knows exactly what's needed to win, but the visiting team knows exactly what's needed to avoid losing.

  10. Rickter says:

    I've been saying for a few years now, the coin toss should be thrown out, and the OT should begin like it was the start of a new quarter. Teams switch sides, and the team that has the ball retains possession. If one team scores on the last play of regulation, they kick off to start the overtime so the opponent has an opportunity to respond to the score, which would almost guarantee most coaches would go for two since the 60% chance of converting is better that the 40% chance of winning the game after kicking off to start OT.

  11. billsfan says:

    I wonder what the win-probability trees look like for some of the other "reasonable" proposals to fix OT, or if it is even possible to do sudden death OT in a WP-neutral manner? It seems like playing a full 5th quarter (with a new kickoff) is the only "fair" option.

  12. Dr Obvious says:

    I love the idea of an auction, but I don't think we'll ever see anything like that.

  13. brett says:

    I'd be okay with the new playoff overtime format if the 2nd team got a chance to respond to an opening drive touchdown. Using the numbers from your event tree, if the 2nd team matches the touchdown 75% of the time, it would bring the overtime odds down to 50-50. 75% is high, but not that high since they can use all 4 downs. Using a more conservative estimate of 40% would give the coin-toss winner about a 53% chance of winning.

  14. brett says:

    Also Brian, if you are looking for topics to post about during the off-season, I would love to hear your thoughts on 4th down decisions in the new overtime format based on WP in different down/distance/field position situations. My gut tells me there are very few situations in which kicking a FG is the best decision.

  15. Marver says:

    I don't understand why the game can't just continue at the end of regulation in a tie game. I always thought it was a sham that teams would lay down at the end of regulation to force overtime, and deciding by coin flip certainly makes less sense than letting the team with current possession keep possession.

    The second guessing from the media following the first overtime playoff game that doesn't end on a first possession touchdown will be overwhelmingly irritating.

  16. Anonymous says:

    For people who dismiss moving the kickoff back to the 35: this does not have to mean more touchbacks. Simply make kicking the ball out on the fly a penalty, and taking a touchback on a kick that lands in play a penalty.

    The ball goes to the 10 or 30 yard yard on a touchback, depending on who gets penalized. The result? Almost all kicks get returned AND field position after the kickoff is pushed back by 5 yards.

    This could be used only for OT, for all kickoffs that start a half, or for all kickoffs - although the last two would reduce scoring.

  17. Anonymous says:

    I would favor giving the home team the advantage (i.e. the equivalent of having won the OT coin flip) for all games except the Super Bowl, when the team with the better record in the regular season (subject to tiebreakers if they had the same record) gets the advantage.

  18. Mark says:

    What happens in double or triple OT?

  19. Nate says:

    I can't understand why Rickter and Marver's suggestion hasn't gotten any traction. It's fair, it's simple, and it's exciting. What's the downside?

  20. Dave says:

    Re: Rickter, Marver and Nate

    The downside is obvious: in a tie game, there would be no pressure on the team with the ball to score before the end of regulation. Remember Adam Vinatieri's dramatic 48 yard field goal to beat the Rams in Super Bowl XXXVI as time expired? If the game continued into the "fifth quarter" as it were, the Patriots could have taken their sweet ass time, 4 yards and 40 seconds at a time. They would have kicked a chip shot field goal five minutes into overtime and it would have been boring and terrible.

    If there's one thing I've discovered in discussing different overtime scenarios it's this: NOBODY CAN AGREE ON ANY SOLUTION. Every suggestion has a downside- either it's less dramatic, or it's not fair enough, or it takes too long, or it leads to ties. No matter WHAT the league decided on, most people would find something to complain about.

    Instead of criticizing the league on this, let's give them some credit for doing something right for once. I actually find this rule to be a good combination of fairness (although it's not perfectly fair), brevity (although it's not perfectly brief), and drama. Nobody can name a proposal that beats this one on all three of those criteria.

  21. Samuel says:

    Dave-
    good point. The pressure of the clock at the end of the game is very dramatic. You could imagine teams intentionally running out the 4th quarter clock, so that the score would take place in overtime rather than regulation, just to make sure that it would be sudden death, with no need for a last second kickoff (e.g., music city miracle).

  22. Anonymous says:

    I don't think the two games or so affected by coin toss in a regular season really matter but certainly for post season...

    Take out the toss (other than Superbowl)

    Home team has option of first or second drive

    Team One starts at 15. If Defense score, game over. FGA only allowed on 4th Down. PAT only; no TPC

    Whatever the result (other than Defense score)..

    Team Two starts at 15. Matching FGA again only allowed on 4th Down. If TD, only PAT allowed

    If FG or Off TD are matched - with or without the PAT, then Team One resume at their 15

    Play is now continuous until a score but a FGA is still only allowed on 4th Down

    Hugh

  23. Josh Robben says:

    Seems like the auction for field position would preserve the suspense of who gets the ball first in OT, mitigate the first-possesion advantage, and add a wildly facinating new element to OT games. Seems like a perfect and simple adjustment to the new format!

  24. Anonymous says:

    I like what the NFL did, because as you pointed out before, it really help to level the number of games essentially decided by a coin toss while managing to keep many of aspects of regulation play involved. Although your suggestion would help, i dont think coaches knowing they won't get the ball first in OT would make up for a 10% change that this new rule is estimated to adjust for. Plus it's not quite the same as baseball, because in baseball every inning you're playing til 4th down. The bases are wiped clean and there is no "protecting field position by punting on 4th down" instead of being gusty and going for it. But in sudden death football, you have the option to be aggressive and go for it on 4th down or protecting field possession.

    With that being said, most ppl don't realize football OT is more similar to baseball than to hockey, soccer, or wrestling (real wrestling). Ppl may argue that in football the defense can score, unlike baseball. BUT what they dont realize is on any DRIVE, there is less than a 2% chance the defense will score a TD or safety(turnovers may set things up, but that's a change of possession with the offense then taking the field). In other words, from the 158 OT games in the last decade it would be estimated there were THREE OT games won by the defense scoring off the offense's first possession drive in the last decade, so you NEED some sort of guaranteed possession system in football OT. Baseball is obv 0% and it's hard to calculate soccer, hockey, and wrestling because their play is continuous but i'm guessing it's a lot higher than 2%. But regardless football isn't that far off from baseball. and THAT would be an interesting stat breakdown to see shown.

  25. Nate says:

    Dave,

    Sure, in some situations the game would be less exciting, like a long last-second field goal with the game tied. In other situations it would be more exciting -- for example, if a team scored a touchdown to tie the game near the end of regulation, they'd now go for a two-point conversion. Overall, it seems like a wash.

    Thanks,
    - Nate

  26. Anonymous says:

    Nate--WRONG. Have you ever seen anybody hurry up to score before the end of the 3rd quarter?

    If you keep the ball into OT, nobody will try to hurry up with the game tied late in the 4th Q (and that's the games we're discussing--those who are tied at the end of the 4th Q which will go to overtime).
    Maybe Brian can help with this--Brian, when does the last score IN REGULATION take place (on average) for OT games? In the last 2 minutes? Last 5?

  27. Tarr says:

    Rickter, Marver, and Nate are right. Simply continuing the game *IS* the best, and obvious, soultion. It's perfectly fair, it's decisive (i.e. it won't result in more ties), it causes games to end as quickly or more quickly than they usually do, and it doesn't distort the game in any significant way.

    The common complaint about removing time pressure at the end of close games is a red herring. I've put this up before, but let's examine the various situations that teams can face when they have the ball in what is likely to be the last meaningful posession of a game:

    Down 8+ points: no change in strategy from the current OT system.

    Down 7 points: more likely to go for 2, because you know your opponent will have the ball in OT. Result: more strategy, more drama, fewer games going into OT.

    Down 4, 5, or 6 points: no change in strategy from the current OT system.

    Down 3: More likely to go play for the touchdown, in stead of getting conservative once you get into field goal range and taking your chances in OT. Result: more strategy, more drama, more time pressure, fewer games going into OT.

    Down 2 or 1: no change in strategy from the current OT system.

    Tied game: removes the time pressure. The game basically feels like an OT game right away. Result: a wash strategically, but less drama, and more likely to go into OT.

    Leading: no change in strategy from the current OT system.

    So, in total, you don't increase the incidence of OT, nor do make the game more simple strategically.

    If you regard ANY change in endgame strategy as inherently bad, then you can of course reject the above argument. But I think it's impossible to look at the above breakdown and argue rationally that this is somehow taking the drama out of endgame situations.

  28. Tim says:

    @Tarr,

    I gotta disagree. You're missing the most obvious side effect of having that form of OT. For ex) there is 2 minutes left on the clock and a team is driving and the score is tied. But with your system they have no reason to go to a hurry up offense or try to move the ball quickly. They can just take their sweet time (like its the beginning of the 2nd quarter) and go down and score sometime at the beginning of OT. That's where the lack of excitement will come from.

  29. Tarr says:

    Tim,

    I addressed that exact situation. Did you read my post or just skim the first and last paragraphs? I noted explicitly that that situation has less drama. I also noted that two other equally likely situations have more drama.

  30. Jason says:

    Just for fun...

    Franchise Possession Arrows:

    All teams start with a "possession arrow" equal to zero.

    If you win a coin toss for overtime, your p.a. is adjusted by -1.

    If you lose a coin toss, your p.a. is adjusted by +1.

    When teams meet that have different p.a. values, the higher team automatically wins the toss (and both scores are adjusted).

    [Optional: A team which loses a tiebreaker for a playoff spot (but not for playoff seeding) (and excluding head-to-head) gets a +1; the team winning such a tie-breaker gets a -1.]

    A team which has clinched a playoff spot, or been mathematically eliminated, may chose to defer.

    If both teams defer, a coin toss is used. However, a team which defers will never have their p.a. adjusted positively OR negatively.

    Playoffs count double, whether a p.a. difference is used or a coin flip is required (because of tied p.a.'s).

    ...

    Complicated, perhaps, but the details can be handled off-the-field and don't require any additional thinking for our mentally taxed head coaches.

  31. Anonymous says:

    I was wondering if any one can answer this question for me. On the diagram if a team scores a Field goal on the 1st possession why does the 2nd team have a higher probabilty of scoring both a Touch Down and a Field Goal then the 1st did on the 1st possession?

  32. Brian Burke says:

    Because the second team has nothing to lose by going for it on every 4th down. Either they go for it or they automatically lose. Punting buys them nothing, so they'll never punt, so they have a higher chance of scoring.

  33. Tim says:

    @Tarr, I'm more than a lil late responding to your response but i'll give it a shot regardless. Yes, I read it and you didn't address it. YOU skimmed over what would happen in OT when the score would be tied as having less drama because it'd be one of many situations. Problem is that's the ONLY situation that matters. How much harder will teams try to get TD's instead of field goals? I'm guessing going for it on 4th down (when they're in field goal range) has a worse success rate than if they kicked the field goal and let it go to OT. Teams have a worse chance of winning the game by getting a 2 pt conversion than if they tied it and let the the other team get the first OT possession. So the most important situation, you know the one you "skimmed over," that'll be the situation most affected...making the game worse.

Leave a Reply