Writers of Pro Football Prospectus 2008

Most Recent FO Features

JonesJon15.jpg

» SDA: Early Playoff Elimination Round

TCU-West Virginia and Auburn-Ole Miss might as well be early playoff elimination rounds, with the losers likely knocked out of playoff contention.

04 Apr 2014

The NFL's Four-Minute Offense Is the Prevent Offense

by Scott Kacsmar

It has become routine in the NFL to see at least 60 times per season where a team wins the game after trailing in the fourth quarter. Most of the memorable comebacks were the product of a successful two-minute drill by the offense. Johnny Unitas is known for inventing the two-minute drill with the Colts in the 1950s, but it has grown to be a crucial part of the game that's practiced by every team each week. Some teams even use the no-huddle, hurry-up offense to start games now.

Anyone can spot a good two-minute drill when they see it, because it is very different from the rest of the game. The offense is not substituting or huddling, plays are called quickly at the line, there's a heavy emphasis on passing, receivers want to get out of bounds to conserve time, spikes and laterals become options and even some of the rules change in regards to fumble recoveries (credit the "Holy Roller") and pre-snap penalties. The offense gets to utilize four-down football and must understand when to use timeouts. With this being such a dramatic change from normal game flow, practice is paramount.

So what are teams practicing when it comes to the less popular four-minute offense we hear about? Former Browns GM Michael Lombardi said "It probably is the most vital element of any successful team in the league." My research on the last three seasons shows that the offense runs the clock out in only 27.9 percent of these four-minute situations. While teams still won 91.5 percent of these games, the four-minute offense seems to be an insignificant factor in that success. If mastered, that win percentage should be 100 percent. You cannot lose the game if you never give up possession of the ball with the lead.

In theory, this should be a much bigger deal than it is. All those great comebacks from the two-minute drill never have a chance to happen if the leading team's offense takes care of its own business. Yet through nearly 100 years of professional football, no one has made a Unitas-sized impact to show everyone how the four-minute offense should work. Bill Walsh said you have to practice the four-minute offense, but what should it look like? Historically, it has often looked like this: a team nursing a one-score lead conservatively runs the ball three times in a row against a defense expecting the run, the punting unit comes on, and the offense casually jogs off with hope the defense will not allow a successful two-minute drill.

That's great for my weekly column, because it just would not work without the dramatic finishes. But for anyone with an interest in strategy, this is completely unacceptable.

For all intents and purposes, the four-minute offense is football's prevent offense.

Is the Situation Aptly Named?

It is unclear who coined the term four-minute offense. Why not the five-minute offense or three-minute offense? What's a reasonable amount of time for an offense to successfully bleed the clock and maintain possession through the end? We can determine if four is suitable or not.

For years the average NFL drive has lasted approximately 155 seconds. That number is surely impacted by drives at the end of a half that are limited by the clock, so asking a team in clock-killing mode to keep possession for three minutes should be more than reasonable. Factor in the huge advantage that comes with having a first down after the two-minute warning (i.e. kneeldowns), and most of the time the team rarely has to burn a full four minutes.

In fact, of the 204 drives I studied over the past three seasons, the average start time was 2:37, or two seconds longer than the average drive duration in 2013.

The average drive also runs about 5.7 plays and gains 1.7 first downs. Since the four-minute offense can win the game and comes in a high-leverage situation, success should be a little harder than the average drive. Asking for six plays and two first downs should be reasonable. That is what makes this part of the game so strange; suddenly, it's actually better to gain 10 yards on three plays than to gain 10 yards on one play. Offensive efficiency in the four-minute offense is better described as clock efficiency. You want to gain yards, but you also want to take your time doing it, and scoring is no longer a primary goal. That break from normal NFL philosophies may be another reason why no one has created a good strategy yet.

What everyone in the NFL has to know is that if the offense gains a first down after the two-minute warning and the opponent is out of timeouts, the game is over. Kneeldowns are the only strategy at that point. Rarely does this get botched, except by Joe Pisarcik and the 1978 Giants, victims of the Miracle at the Meadowlands. That's the best lesson for what a team should not do at this point of the game. Risking a fumble by running a real play is not worth it.

Last season, Mike Smith's Falcons, usually on top of situational football, ran Jacquizz Rodgers three times for a meaningless 12 yards after the Rams used their final timeout with 1:44 left. That should have been three knees by Matt Ryan. It was a rare miscalculation.

Three knees will also kill off a team with one timeout left if the game clock is at roughly the 1:25 mark. With the defense down to one timeout, the offense can essentially burn 80 seconds just by running down the play clock twice. The time spent doing the kneeldowns and resetting the play clock should consume at least five seconds for multiple kneeldowns. Last year, Rex Ryan's Jets had 1:21 left against the Saints (one timeout) and ran Bilal Powell twice before taking the knee. That was pointless, as three knees easily would have taken care of that win.

In the last four minutes, if a team converts on third down twice, it is impossible for the defense to get the ball back (even with the maximum four clock stoppages) unless the offense commits one of the cardinal sins of the four-minute offense: an incomplete pass, a penalty, or going out of bounds to stop the clock. Also, defensive penalties will stop the clock, which can be harmful to the offense. It happened last season when Baltimore played in Cleveland. A Baltimore horse-collar penalty was tacked onto a gain already good enough for a Browns first down, but the clock stopped, essentially buying the Ravens an extra timeout. If a team can afford to give up the field position, then getting a penalty (think unnecessary roughness) on a play that already gained the first down may be a cheap strategy to use, but that's some heady, on-the-spot decision making for players to consider.

Ultimately, a team is looking at one to three first downs to ice the game in the four-minute offense. Out of 204 drives since 2011, only 12 gained three first downs, and none gained more than that. Of those 12 teams, only one failed to run out the entire clock or score a touchdown. Last season, the Chiefs ran eight plays and punted the ball back to Dallas with just 16 seconds left. Jamaal Charles was pushed out of bounds (a big no-no) on a run that gained a first down. The Chiefs also converted on third-and-10 thanks to a defensive pass interference penalty on Morris Claiborne. That stopped the clock with 2:28 remaining and Dallas out of timeouts. Had the Chiefs converted on their own without the penalty, the game would have ended right there.

Sure, Dallas only had 16 seconds to work with, but all a team needs is one play to have the opportunity to score. Would you rather make the plays to win the game, or hope your opponent fails? The numbers certainly suggest most of the league has no problem with the latter.

The Four-Minute Offense Study (2011-2013)

So we should be comfortable with the idea of a four-minute offense, but how do we study it? For the last three seasons, I have used these requirements for the drive to count:

  • The offense is leading by 1-8 points (one-score game).
  • The offensive drive began with 4:00 or less remaining in the fourth quarter.
  • Drives that were only kneeldowns or three kneeldowns followed by a punt are excluded.
  • All kneeldowns are excluded from all drive stats.
  • If a drive produced zero first downs and included even one kneeldown, it is excluded.
  • Playoff games are included.

Perhaps the only controversy here is the time limit. Why not include drives that start at 4:05 or 4:13 too? Trust me; there were so many drives that started just outside of four minutes in 2013 that I nearly changed my rule out of frustration. But I stuck to four minutes so that I could make comparisons to previous research I had done for 2011 and 2012, and I believe this produces a fairer sample than branching out to something like 4:30.

The data's sample size would surely increase by going beyond four minutes, but we would lose some of the integrity of what the four-minute offense is supposed to accomplish. I also do not think including partial drives where there was a first down in the final four minutes would help. A drive could start at 4:04, the offense gets penalized for holding and suddenly it's first-and-20 with 3:57 left. The drive and its data have been compromised.

The elements present at the start of the drive are critical to studying it. Bad field position might make it dangerous to throw a pass with the quarterback standing in his own end zone. That's going to impact the play calling. For a study of this nature, only full drives should be analyzed.

There were at least 75 four-minute offense drives in 2011 and 2012, but this past season produced only 50. Here's a look at the drive breakdown:

NFL's Four-Minute Offense Summary
Year 2011 2012 2013
Total drives 79 Pct. 75 Pct. 50 Pct.
Punts 32 40.5% 39 52.0% 25 50.0%
Ran out clock 26 32.9% 18 24.0% 13 26.0%
Touchdowns 9 11.4% 4 5.3% 2 4.0%
Field goals 8 10.1% 8 10.7% 7 14.0%
Missed field goals 0 0.0% 1 1.3% 0 0.0%
Lost fumbles 2 2.5% 2 2.7% 1 2.0%
Interceptions 0 0.0% 1 1.3% 0 0.0%
Downs 1 1.3% 0 0.0% 2 4.0%
Intentional safety 1 1.3% 2 2.7% 0 0.0%
Three-and-outs 24 30.4% 26 34.7% 17 34.0%

Roughly half of the drives end with a punt, a third are three-and-outs and the goal of running out the clock gets accomplished 27.9 percent of the time. In 2013, there were just as many teams who ran out the clock (13) as there were teams who ran the ball three times and punted. These are not impressive results for the offense. Of the 23 field goals since 2011, 16 of those drives failed to even gain a first down. They were all set up by great field position. Only three of those 16 teams attempted a pass.

Touchdowns are rare with two scored in 2013. Often times, it's better to not score in the four-minute offense. I ranted about Marshawn Lynch not needing to score his touchdown against the Saints in the playoffs, but LeGarrette Blount's Week 17 run against Buffalo was even more unnecessary. With the Bills out of timeouts and the play starting at 2:40, Blount could have gone down at any point after 10 yards to end the game, but his 35-yard touchdown run produced a 34-20 final and added to his career day, even if it meant nothing to the Patriots.

Next are the four-minute offense drive averages for the last three years. The "End time" is the average time left in the game after the drive ended while the "End time*" is the same, but excludes drives where the team never gave the ball back. So when teams do get the ball back, they typically have at least 80 seconds to work with, which can be an eternity in today's fast-paced NFL.

Drive Averages
Year 2011 2012 2013
Start time 2:32 2:43 2:35
Lead (PTS) 5.24 5.16 4.64
LOS 39.3 35.7 39.8
First downs 0.9 0.8 0.8
Plays ran 3.6 3.9 3.7
Yards 19.8 15.9 14.5
End time 0:58 1:02 1:01
End time* 1:28 1:21 1:22

Teams leading by 4-8 points were more likely to run out the clock (30.6 percent) than teams leading by 1-3 points (21.7 percent), but I do not think we have enough data in this sample to draw any sweeping conclusions on that front. I did find it interesting that 21.7 percent of the teams leading by 1-3 points settled for a field-goal attempt as opposed to 7.6 percent of teams leading by 4-8 points. Kicking the field goal to take the dreaded 4-6 point late lead is nowhere near as advantageous as the field goal is for the team already up by a touchdown.

One thing I will speak conclusively on: whether the lead is a field goal (84.8 percent) or touchdown (86.1 percent), NFL offenses grossly favor running the ball in the four-minute offense.

The splits for play-call ratios by down for 2011 and 2012 led to one of the most consistent tables I have ever prepared. I have added 2013 and also made two adjustments for every season. Scrambles by quarterbacks on designed passes are now considered passes and plays negated by penalty are included as well.

Four-Minute Offense: Called-Play Ratio
2011 2012 2013
Down Run Pass Run% Down Run Pass Run% Down Run Pass Run%
1st 112 6 94.9% 1st 114 5 95.8% 1st 71 6 92.2%
2nd 83 12 87.4% 2nd 89 15 85.6% 2nd 51 13 79.7%
3rd 49 22 69.0% 3rd 50 29 63.3% 3rd 32 19 62.7%
4th 2 0 100.0% 4th 1 0 100.0% 4th 1 0 100.0%
TOT 246 40 86.0% TOT 254 49 83.8% TOT 155 38 80.3%

Why even bother defending the pass on first down when teams run the ball 94.6 percent of the time? That's 297 runs to 17 passes since 2011 -- the very definition of predictable. There's a little more passing on second down, but it's not until third down when you can actually doubt a team's willingness to run. Even then, that can become obvious based on distance and time remaining.

The sad part is this seems to be a league-wide epidemic where even the teams with the best quarterbacks conform to the run-heavy, conservative approach.

The most egregious example in 2012 was Peyton Manning's Broncos. When two first downs would have knocked Baltimore out of the playoffs, the ball went to Ronnie Hillman on the ground on five consecutive plays. Thanks to that, Jacoby Jones and Rahim Moore will always have a place in NFL lore. The Broncos did take that situation to heart this offseason and were the most successful four-minute offense in 2013. They had a league-high three drives to run out the clock (twice against San Diego). In the AFC Divisional round, the Broncos exorcised the playoff demons with a four-minute drive that included four passes -- something only done one other time in my database of 204 drives. The Broncos looked like they were quickly headed for a three-and-out drive, but the Chargers left Julius Thomas wide open on third-and-17. It was an aggressive, successful drive to close out a playoff game, which is exactly what you should get from a record-setting offense.

New Orleans also has one of the best quarterbacks in NFL history, but the Saints epically blew the four-minute offense in 2013. It was so bad they picked up two of the league's four losses thanks to game-winning drives in the final minute by the Patriots and Panthers. While Rob Ryan's defense gets the majority of the blame, the problem started with Sean Payton's loaded offense failing to get any first downs. I have detailed the unbelievable clock mismanagement the Saints suffered in Foxboro before, but here's a reminder. With two plays needing to gain a total of eight yards to win the game, the Saints handed off to Pierre Thomas for a yard and called a naked bootleg by Brees on third-and-7. Cue Tom Brady and Kenbrell Thompkins for the win.

Going into Super Bowl XLVI, former Giants offensive coordinator Kevin Gilbride said "you need to attack, not step back" in the four-minute offense. I only have three drives in the sample from the Giants over three years. They ran the ball nine times and had zero passes. At least they produced a touchdown and ran out the clock once, but that's not the mark of an attacking offense.

The Colts were very good in the four-minute offense in 2012 behind Andrew Luck with a league-high four drives that ran out the clock. Last season, they only had one opportunity. In Houston, Trent Richardson carried the ball twice, so needless to say the drive lost yardage and was a three-and-out. Houston ended up missing a 55-yard game-tying field goal with no time left. A lack of aggression nearly forced overtime.

In 2013, the Patriots were the only team willing to go for it on fourth down, and this time it was not a fourth-and-2 in their own end. A botched snap against Atlanta failed and the Patriots hung on for the win thanks to red-zone defense. Atlanta faced the only other fourth-down play, but it was a very strange call by the Saints to let Brees lose 11 yards on a give-up play, losing field position in exchange for time. Again, the Saints were not very sharp in this department last year.

Dallas had one of the season's biggest errors in Detroit when Tyron Smith was penalized for holding after a third-down run. That stopped the clock at 1:07 and Dallas kicked a field goal to take a 30-24 lead. But that's too much time for Matthew Stafford to lead one of the game-winning drives of the year, finishing with a quarterback sneak score with 12 seconds left. Dallas even had two cracks at the four-minute offense in that loss, but only gained one first down (via penalty).

In trying to find a 2013 team which took an aggressive approach, Joe Philbin's Dolphins are about as close as it gets. Against San Diego, Ryan Tannehill started the drive with back-to-back pass plays on first down. That's two of the league's six passes on first down in 2013. He also had two passes on second down, dropping back five times total on the drive. A second-down pass backfired against Buffalo when Tannehill was sacked and fumbled, setting the Bills up for a game-winning field goal. Miami only had one such drive in 2012, but that also featured a rare second-down pass.

Before giving Miami too much credit, let's wrap up this 2013 review with one of the most glaring examples of why the four-minute offense needs better results. Against Pittsburgh, the Dolphins led 31-28 and had a first-and-goal at the 9 with 2:23 left. Since they could not get a first down without a touchdown (at least not without a penalty), Miami's best strategy was to go for the kill and get the touchdown to take a 38-28 lead. Instead it was the most conservative approach possible: three straight runs, allowing the Steelers to use the two-minute warning and final timeout to get the ball back with 63 seconds and 80 yards needed to win the game.

Now that might be a dire situation for the Steelers, but we have seen teams pull this off before. Miami gave them the chance. What sounds even grimmer is having to go 74 yards with three seconds left. Well, after a couple of laterals, Antonio Brown was inches away from doing just that. Brown barely stepped out of bounds before scoring what would have been one of the most improbable game-winning touchdowns in NFL history. You could put it up there with Stanford-Cal and the River City Relay.

As long as the offense is on the field in a one-score game, they always have a chance. The way to prevent allowing these chances is the four-minute offense, so stop making it the most predictable and risk-averse part of the game.

Conclusion: What's a Good Strategy?

The four-minute offense will never have a strategy as clear-cut as the pass-heavy two-minute drill. It's a whole different mindset of offensive football, but it has to get smarter than 95-percent runs on first down. At least give the illusion of a pass on first down. That might improve the running success. Sometimes, a simple smoke screen to a receiver in single coverage should be an easy completion that gains yards and keeps the clock moving.

Teams cannot fear the interception. Not even 3.0 percent of passes in the NFL are intercepted these days. There's practically no desperation at play with the four-minute offense, so risky passes are rarely ever called, let alone thrown (unless Ryan Fitzpatrick is your quarterback).

Don't get me wrong. The four-minute offense should continue to lean on the running game, but the split cannot continue to be over 80 percent. Mixing in some safe passes on first and second down will bring that number down and should increase the rate of teams successfully running out the clock.

Nothing will stop the few teams built on running the ball and playing defense from sticking with the conservative approach in the four-minute offense. Frankly, as we have seen, no matter what an offense does in the last four minutes, they still are likely to win the game. The record is 176-16-1 (.915) since 2011. But I'm still focused on the 16 losses and the numerous close calls that did not have to be had the offense kept the defense on the bench.

The four-minute offense didn't help the Jets beat Tampa Bay in Week 1 last year. It failed and Tampa Bay took the lead with 34 seconds left. The only reason the Jets won was a fortunate game-winning drive thanks to Lavonte David pushing Geno Smith out of bounds.

The Ravens beat the Bengals in overtime in Week 10, but that was after A.J. Green caught a tipped Hail Mary to tie the game. Why did that happen? Baltimore's rich quarterback handed the ball off three times in a row and the offense punted to set that drive up.

There's always going to be that dilemma of calling a pass when one more first down wins the game versus handing off to burn another 40 seconds. That element of time is becoming less of a burden with offenses able to move the ball quickly, often needing less than a minute these days. Since 2010, there have been 27 one-minute drills to win a game. There were 26 in the entire decade of the 1990s and 36 in 2000-09. The game is slowly changing, but the four-minute offense remains untouched and undeveloped.

Ever get the feeling more games are lost in the final minutes than are won thanks to these "prevent" offenses and defenses? Somehow, much of the league has adopted late-game strategies that do not allow the best players to do what they do best. If you have a good quarterback, let him continue to throw and win the game. If you are having success with blitzing, keep bringing pressure.

If something worked for 56 minutes, then why not stick with it for the final four? When the Saints really need a first down in the first half, they don't usually depend on Khiry Robinson, right? So why are they using him and a fullback to invite a loaded box with the game on the line? That's not who the Saints are.

The only thing worse than the prevent defense is the prevent offense.

Posted by: Scott Kacsmar on 04 Apr 2014

40 comments, Last at 09 Apr 2014, 10:46pm by Karl Cuba

Comments

1
by MJK :: Fri, 04/04/2014 - 2:03pm

Good article, on a very interesting issue. I think that one thing that will come into play is that savvy defensive coordinators realize that, in a 4-minute situation, giving up only a minimal (<3 yards) gain or incurring a loss of yardage is even more important than normal, and giving up a 10 yard play on 1st down is actually preferable to giving up, say, a 5-8 yard play. So they will sell out and risk single coverage or a cover-0, and load the box. In other words, offensive coordinators should expect success rates on "high percentage" plays to be lower than typical. You see this on 4th down conversions all the time--but it's even more true in a 4 minute scenario.

Kind of like bringing your outfielders into the infield in the bottom of the 9th in a tie game with less than two outs and a runner on 3rd...

4
by steveNC :: Fri, 04/04/2014 - 2:49pm

Interesting article, enough to inspire a registration. I wonder how many CBS/FOX announcers realize this fact, that giving up a first down or TD on first down is often not as bad as giving up a third down conversion in a 4-minute situation. They certainly don't seem to mention that element of defensive strategy that often.

5
by nat :: Fri, 04/04/2014 - 5:39pm

...giving up a 10 yard play on 1st down is actually preferable to giving up, say, a 5-8 yard play

Isn't second and short always better for the offense than first and ten, and by enough to make up for a few yards of field position?

You're not wrong that the calculation would be a little different in the four minute situation. But maybe not by much.

The first down gives the offense a complete new set of downs to use up clock, with a normal chance of moving the chains again. The other only gives them two downs to burn clock, but a better chance of moving the chains. Giving up five yards may still be a better deal for the defense if all that matters is the average time until they get a stop.

2
by nikshaw :: Fri, 04/04/2014 - 2:34pm

Nik Shaw

I always thought that if anyone invented the two-minute offense, it was Bobby Layne, long before Unitas entered the NFL.

3
by duh :: Fri, 04/04/2014 - 2:46pm

Seems like on first and second down the ideal play call might be a designed roll-out with the QB told if the pass isn't there to just take what is there and stay in bounds.
either that or maybe go run some plays out of the single wing ;-)

6
by Steve in WI :: Fri, 04/04/2014 - 7:38pm

Great topic for an article. This has really started to frustrate me in the past few seasons...teams are becoming increasingly adept at driving down the field in two minutes (or less) and yet coaches are content to run the ball up the middle three times and punt with a one-possession lead.

I also agree that if safe passes are called, the risk of an interception drops to be next to nothing (and let's not forget, there's always the risk of a fumble with a running play).

7
by Q :: Sat, 04/05/2014 - 6:31am

When I read about how even teams with the Best QBs go ultra conservative, I was stunned not to see a giant pic of McCarthy.

I am very surprised the article did not bring up the notion of blame. Since most coaches are risk averse and conventional wisdom is to run, run, run/pass, Head Coaches do not want to do things that could lead to them being blamed.

As the article indicated, the blame for the loss to the Pats generally went to Ryan and the Defense, not Sean Payton for his playcalling on Offense. If Payton calls 3 straight passes and NO goes 3 and Out, Sean's playcalling is the lead on Sportscenter and Skip Bayless is spending 20 minutes on it the next day.

With a 95% run rate on 1st Down, I want to see a 1st Down Flea Flicker

8
by MC2 :: Sat, 04/05/2014 - 11:18am

Good article, and I generally agree. One thing that bugs me, though, is this part:

Often times, it's better to not score in the four-minute offense. I ranted about Marshawn Lynch not needing to score his touchdown against the Saints in the playoffs, but LeGarrette Blount's Week 17 run against Buffalo was even more unnecessary.

I actually hear quite a bit about this (usually from announcers who think they're being really clever), and technically the logic is correct, but it all strikes me as a classic case of much ado about (basically) nothing. After all, scoring these "mistake" TDs probably reduces your chances of winning from 100% to 99.9%, so what's the big deal? There are dozens of mistakes in every game that are much more costly (in terms of win probability), and yet many of them go almost completely unnoticed.

9
by dmstorm22 :: Sat, 04/05/2014 - 2:15pm

The playoff game was probably as close to the worst case scenario without actually losing. The Saints scored a quick TD to make it 23-15, recovered the onside kick, and then moved it to the 40 before Colston's hilariously-forward lateral.

Lost in that was Earl Thomas injuring himself on the TD drive. He was able to play the rest of the season, but why subject your players to extra snaps.

Lynch's case is tough because a normal 10-yard run might not have taken off enough time to get to the 2:00 warning, but some fo the other instances Scott mentioned were far more obvious.

14
by MC2 :: Sun, 04/06/2014 - 4:46am

Even there, though, they weren't all that close to losing. The Saints still would have had to score the TD, and then convert the 2-point attempt, and even that would have just sent the game to overtime. So, even after all that, at the time just before Colston's miscue, Seattle was still a significant favorite to win the game.

10
by Jazzaloha :: Sat, 04/05/2014 - 3:12pm

@MC2

The probability may not change, but I can think of two nerve-wracking situations--that could have been avoided had the offense not scored. In the Lynch example you reference, the game would have been over had Lynch not scored. Instead, the game came down to the last play, with the Saints having a chance to score. The other example that comes to mind is the second Super Bowl between the Giants and Pats. I think Bradshaw scores a TD (instead of stopping at the one; this would have required the Giants to kick FG, however), leading to a Hail Mary that Gronk missed on. Rooting for both the Seahawks and Giants in those games, the ending was nerve-wracking, and in the case of the Seahawks, that could have been avoided.

11
by Jazzaloha :: Sat, 04/05/2014 - 3:21pm

Some random comments:

1. The author says, "Frankly, as we have seen, no matter what an offense does in the last four minutes, they still are likely to win the game. The record is 176-16-1 (.915) since 2011. But I'm still focused on the 16 losses and the numerous close calls that did not have to be had the offense kept the defense on the bench."

To me, this suggests the current approach is sound--and that worrying about the 16 losses seems unreasonable, "greedy." If your chance of winning is 90% with a one score lead, then I would think the wisest strategy would avoid silly mistakes; and trying to move from 90 to 100% seems unwise if it means making a costly mistake (that could come by passing more on first down).

2. Having said that, one of the passes that I see and like are the play action passes from heavy run formations. I'm thinking of the play action passes that go to a FB, TE or even the RB. The routes are usually short, so they seem like high percentage plays.

3. I do think that executing a good four minute offense is important. Actually, I would expand that to include an offense that can effectively run time off the clock when a team has the lead in the second half. Looking at this issue is interesting when you think about the hurry-up, aggressive offenses. Can they excel in four-minute situations? I'm skeptical because the philosophy behind these offenses seem diametrically opposed to the requirements in a four minute offense. This is one of the reasons I'm not too keen about these offenses.

13
by Jerry :: Sun, 04/06/2014 - 3:14am

Your first point looks right to me. Opting to take the time off the clock seems to work.

Generally, I wonder exactly how valuable each 40 seconds is. I don't know if there are enough games in the database to compare a trailing team starting its last drive at 1:30 vs. 1:00 (all other things being equal), but that would be interesting to find out.

15
by The Powers That Be :: Sun, 04/06/2014 - 4:40pm

For what it's worth, Football Reference's drive finder says there have been 712 drives since 1999 that started with 2:00 or less left in the game and the offense trailing by 1-8 points. Here's the breakdown of those drives:

Time left(s), Drives, TD, "Viable" FG attempts, Success Rate
111-120 92 18 16 37.0%
101-110 86 17 13 34.9%
91-100 56 7 10 30.4%
81-90 29 4 2 20.7%
71-80 34 5 4 26.5%
61-70 71 8 10 25.4%
51-60 53 3 7 18.9%
41-50 51 2 5 13.7%
31-40 41 0 3 7.3%
21-30 65 2 9 16.9%
11-20 75 3 5 10.7%
1-10 59 0 1 1.7%

I arbitrarily excluded FG attempts greater than around 54 or 55 yards, on the argument that those weren't really "successful" offensive drives. That eliminated a whole bunch of missed FGs and exactly one made FG. (To be precise, I'm still counting the drives, just not counting them as successes.) Similarly, missed viable FGs are counted as successes, since the offense "did its job".

The one "successful" FG drive with 10 or fewer seconds left was in a HOU-TEN game in 2005. Houston returned the kickoff to Tennessee's 13 yard line...and Kris Brown missed the FG.

The data looks surprisingly linear, with every extra 10 seconds worth about 3 percentage points in success rate.

18
by Jerry :: Mon, 04/07/2014 - 3:41am

Thanks. If I have a chance, I'll try to put together the matrix with starting field position to see how strong the effect is.

19
by BJR :: Mon, 04/07/2014 - 9:33am

Well, the wisest strategy in the 4 minute offence (or in any game situation) is the one that improves your chance of winning the game the most. If avoiding mistakes is deemed to be paramount, then the best strategy for some teams might be to simply kneel 3 times and punt, even if it does leave adequate time on the clock for another drive. In all honestly, I would sometimes prefer it if coaches did this rather than tacitly admitting they are avoiding errors at all costs by calling predictable running plays. After all, is running into the teeth of the defence any less likely to produce a turnover than passing?

21
by tuluse :: Mon, 04/07/2014 - 10:16am

Running up the middle is significantly more likely to keep the clock running than passing.

23
by BJR :: Mon, 04/07/2014 - 1:33pm

Of course. But the point of this article is that teams should be passing more in the 4 minute offence, and the comment above suggested that one of the reasons they are not passing more is the risk of a costly error, i.e. a turnover. I don't believe that passing is very much more likely to lead to turnovers (within reason) than running, given what the 4-minute defence is likely to be expecting.

12
by Karl Cuba :: Sat, 04/05/2014 - 3:25pm

What would be very useful in trying to assess the ideal run/pass ratio is the expected completion percentage against defenses stacking the run to stop the four minute offense.

If defenses are behaving as you suggest then you would expect the completion percentage, or at least the DVOA, to be higher than usual.

16
by Sifter :: Sun, 04/06/2014 - 6:32pm

I would have thought the D's strategy would be very similar to a goal-line D ie. run blitzing, not worried about zones 10+ yards behind you. Therefore offenses should be running their most successful goal line plays. Rivers did a quick study on that last year (http://www.footballoutsiders.com/ramblings/2013/three-cone-drill-goal-li...). From that, teams are generally splitting run/pass 50-50 in all goal-to-go situations, and 60-40 inside the 3.
If I were running a 4 minute offense, I'd start in the heavy package for 1 play, then substituting to spread on 2nd down, and passing/running based on who the D subs in. Screen pass preferably to keep that clock running.

27
by Karl Cuba :: Mon, 04/07/2014 - 9:23pm

Wouldn't the obvious difference between the goaline D and the four minute defense be that when you run the package designed to operate on your one yard line there isn't plenty of space for receivers behind you?

28
by Jerry :: Tue, 04/08/2014 - 3:41am

The most obvious difference is offensive. On the goal line, the offense is trying to score. In the four-minute drill, they're trying to kill clock, so the incompletion is a much worse result.

17
by Dennis :: Sun, 04/06/2014 - 9:34pm

This reminds me of the old Madden trick (92-93 era, I have no idea if it still works) of smashing your defender into the O-line to get the offsides penalty and stop the clock. Even if it gave the offense the first down, it still gave you a chance to stop them isntead of letting them run out the clock.

22
by Jimmy :: Mon, 04/07/2014 - 11:37am

It still worked as of about 2001/2. I know this because I managed to win an unlikely game by somehow getting an eight yard loss on the first down. Which meant that even with three five yard penalties they still ended up on fourth down with the clock stopped. Yes it was dumb. Yes I am still proud of it anyway.

20
by BJR :: Mon, 04/07/2014 - 10:01am

Some of the language in the article is a little confused. The author talks about the "cardinal sins" of the 4 minute offence (incomplete passes, going out of bounds, etc.), before generally berating coaches for playing too predictably/conservatively. But it's very difficult to generate first downs without risking these cardinal sins. Which is it to be?

24
by Scott Kacsmar :: Mon, 04/07/2014 - 5:05pm

The league-wide completion percentage has been over 60% since 2007 (61.2% in 2013). Factor in scrambles and sacks (both will keep the clock moving) and that number goes up. Factor out a lot of the tougher deep throws that a QB just won't be looking to throw in these situations and I think there's a great chance of keeping the clock moving with a called pass in the 4MO.

Every skill guy has to know he can't go out of bounds in this situation no matter what type of play it is.

25
by tuluse :: Mon, 04/07/2014 - 6:22pm

Marion Barber. Chicago Bears.

30
by BJR :: Tue, 04/08/2014 - 12:12pm

This sounds appealing. But equally, without the defense having to factor in the threat of a deep throw or a sideline throw, shorter passes are going to be more difficult to execute and less effective.

I agree that 95% run on first down in the 4MO sounds sub-optimal and leaves scope for a little more diversity. Still, if we are determined to keep the clock moving at all costs (and the defense knows this) the options are severely limited.

35
by Scott Kacsmar :: Tue, 04/08/2014 - 2:52pm

"shorter passes are going to be more difficult to execute and less effective."

I doubt that, honestly. The same way play-action works even when your running game's awful and you've given the defense no reason to expect a run, the same way we see open sideline routes in the two-minute drill even when it's expected, defenses play to their instincts and aren't as smart as we like to think.

So I think an offense can run normal pass plays in the 4MO with 4-5 targets filling the deep, intermediate and short areas with a checkdown option, and defenses would react just the same (assume shotgun) as they would under usual circumstances. Where you can probably really trick them is to go big and do a play-action pass to a FB or something like that. Or a TE screen to your blocking TE who rarely catches the ball.

So either something creative or something that's more bread-and-butter to what a team usually does to try to have success. Enough with the most conservative running plays in the book on every first and second down.

The clock doesn't have to run after every play for the 4MO to be successful. Like the Saints example I used in the comments, when you know one first down wins the game, I'd sacrifice one incompletion if it improves my chances for a conversion for that clinching first down.

26
by Jazzaloha :: Mon, 04/07/2014 - 7:11pm

@Scott

But here's my problem: if teams with a one score lead with four minute left have about a 90% chance of winning regardless of what they do--then I would think they would have to avoid doing things that would decrease that probability versus trying to find ways to increase the probability.

QBs may be far better at passing the ball now and if they're making short passes, the odds of turning the ball over decrease even more. But why risk lowering your 90% chance of winning? I'm not sure if passing would lower the odds--and if did by how much. But I do think passing, no matter what the type of pass, is riskier than running the ball.

34
by Scott Kacsmar :: Tue, 04/08/2014 - 2:44pm

The 90 percent is a bit misleading, and there's a lot of correlation issues there. The driving force behind that winning percentage is that it's just really hard for a team to get the ball back in the last four minutes and drive for a game-tying or go-ahead score and go on to win the game. The leading team's offense is having very little impact on that record (especially in terms of successful offensive plays) when in theory they should have the most control over it.

I didn't include a list of the 50 drives from 2013 for obvious reasons, but there are some cheap data points that enter into this every year. While I set a limit of no more than 4 minutes at the drive's start time, you could argue I should close the range on the other end and have at least 2:01 left (keeping the two-minute warning in play).

Example: Broncos at Chiefs. Broncos up 35-28 with 1:46 left. Chiefs have one timeout left. The most realistic hope the Chiefs have left is forcing a three-and-out so the Broncos punt and KC gets it back with 10-15 seconds left. Yeah, the Chiefs have almost no shot at scoring a TD that quickly, but again, with the Antonio Brown play in mind, a chance is still a chance even if it's a bad one. But in that situation, Peyton could have taken a dive three times and it's still a 99% win for the Broncos. That's why I eliminate drives with all knees and I'd probably eliminate this drive if I wasn't trying to build up the sample size. What actually happened was Montee Ball got the carry on first down and broke one for 28 yards. Chiefs used that last timeout and PM took three knees to end it. Arizona did the exact same thing to Seattle in December. Those were the two 1-play drives in the 4MO last year.

On drives like that where the opponent has one timeout and it's past the 2:00 warning, teams would be crazy to not just run the ball. Like I said in the piece, if it's around 1:25, they would be crazy not to do three knees.

For 2013, 16 of our 50 drives started at 1:59 or later. Those drives featured 44 runs and 2 passes. The passes happened because the opponent had 2 or 3 timeouts left. The opponent still got the ball back 10/16 times, but with an average 0f 0:45 left. Still, Detroit beat Dallas and the Panthers beat the Saints.

This will matter more to certain teams every year. The Saints were the main team in 2013. The four-minute offense is really the difference between being a 6th seed and a 2 seed for them. They wouldn't have had to go to Seattle on a rainy Saturday in the NFC-D. Maybe they get lucky and a Carolina or SF knocks off Seattle that day instead and the Saints host the NFC-C and get to the Super Bowl. Those final minutes in Carolina (or even NE) were a season-changer.

After the Panthers used their second timeout with 1:52 left, the Saints faced a 2nd-and-9. Any first down will essentially end the game. A non-penalty first down that stays in bounds is guaranteed to end the game. This is the situation I'm really talking about. Give Brees two plays to gain 9 yards with passes and end the game. Instead they take the safe approach of running twice and Carolina gets it back with 0:55 left. The old thought would be 2nd down has to be a run to make them use their 3rd timeout, but that's exactly where I think the strategy can change. Take two shots at getting the first-down pass instead of putting it all on the third-and-long pass. And the Saints wouldn't even do that once it came up.

Losing the game in the last minute is so devastating that I'm surprised coaches are this content with letting the opponent have that opportunity. Kyle Orton quickly throwing a season-ending INT to give the Eagles the NFC East in Week 17 does not make the Eagles' preceding three-and-out drive look any better. Some other day that 2-point lead will not hold and Orton or some other QB will make the plays necessary to set up the game-winning FG and knock you out of the playoffs.

Think of the current 4MO as a half measure that's reliant on the defense. I'm looking for more of a full measure where the offense tries to end the game, and if they can't, we still have our last line of defense too.

38
by Jazzaloha :: Wed, 04/09/2014 - 2:47am

Wait, I'm genuinely confused. Are you saying that being a little more aggressive on offense (i.e., passing) is even more preferable when there's less time (2:00 minutes or less)?

In any event, my actual opinion is that a team should decide if they should be aggressive or not based on the specifics of the situation. How are both teams defenses and offenses playing that day; how many points are you winning by? I might even factor in weather conditions (e.g., are you throwing into the wind or not?).

I don't remember the Saint-Panther game, but giving Newton only 55 seconds left doesn't seem like a bad decision. If it were Brady or Aaron Rodgers, I might want be more aggressive in trying to get the first down.

Losing the game at the last minute--by giving the ball back to the opposing offense--can be devastating, but I'm not sure it's that much more devastating than if you made a crucial mistake on offense because you were trying to aggressive in trying to get a first down.

One final, somewhat tangential, note that you haven't touched on. We're talking mostly about passing on running situations, but there's another factor to consider--namely, the style and philosophy of the team is also important. If you're a physical, run-first team, this team might still be able to run successfully even when the defense knows a team will run the ball--because running in a physical style is their identity and bread-and-butter. Additionally, run-first teams usually invest in running the ball early, which can result in dividends during the last four minutes. To wit, the defense may know the run is coming, but they may just be worn out to stop the running. If what I'm saying is accurate, and that the four minute offense is truly important, then this indicates a huge advantage to run-first offenses and some significant drawbacks for pass-first offenses, especially the hurry-up offenses that attack all parts of the field.

29
by nat :: Tue, 04/08/2014 - 11:31am

Nice topic, and a thought-provoking article. I think we can forgive the more obvious hyperbole (e.g. "If mastered, that win percentage should be 100 percent.")

Judging from the "three-and-out" percentages and the dearth of turnovers, it looks like these offenses are being reasonably successful, if cautious, on their first set of downs. These are situations where the defense can afford to risk long gains in order to force turnovers or punts quickly. And the offense would be right to reduce their chance of moving the chains for a higher average time burned.

But I agree, it can be frustrating to see an offense blatantly signal the run with no real threat to pass.

31
by Vincent Verhei :: Tue, 04/08/2014 - 12:21pm

In the last three years, on passes with a one-score lead in the last four minutes of a game (whenever the drive began), teams have called 281 pass plays, with 165 completions, 91 incompletions (including 5 interceptions), and 25 sacks. Nearly one-in-three plays (32.4%) ended in an incompletion or interception. That's a lower rate than the overall league average (36.2%), but I'd still argue that's an awfully high risk of handing your opponents a free timeout.

32
by BJR :: Tue, 04/08/2014 - 12:49pm

Good numbers, thanks

33
by Karl Cuba :: Tue, 04/08/2014 - 2:18pm

That's interesting; that the difference between a normal D and the four minute D is only about four percent. I would have expected the difference to be greater.

36
by Scott Kacsmar :: Tue, 04/08/2014 - 3:03pm

But if you use the last 4 minutes, you're likely adding a lot of 3rd-and-long plays that are the ending to drives that started with 8, 9 or even 10+ minutes left in the game. That's a bit different from going out there in the last 4:00 and just looking to get 1-3 first downs.

If you use 2011-13, last 4:00, up 1-8 and look at 1st and 2nd down plays only (area that needs the most change), QBs are 73/111 (65.8%) with 8 sacks, 6 TD, 3 INT. One of the picks was a Brad Smith pass that needed aborted.

Either way, these are passing stats that are superior to league-average numbers.

I think those other drives that start with more time left are important too and that might be more of what Lombardi was talking about. If you're up 7 with 7:30 left, can you add another score and really make it hard for them to come back? But I also think that's a different mindset from the 4MO since your goal is to score there, and usually a team will stick with their base offense to accomplish that. It's only when the game gets down to the final minutes that we see this run-heavy approach swallow up the league.

40
by Karl Cuba :: Wed, 04/09/2014 - 10:46pm

Except for in San Francisco, where they run the fifteen minute offense.

37
by Scott Kacsmar :: Tue, 04/08/2014 - 4:40pm

Might go without saying, but I am in favor of being more conservative when leading by 3 or by 7-8 points. More aggressive when it's 1-2 or 4-6 points (danger zone).

39
by nat :: Wed, 04/09/2014 - 9:44am

This is an interesting first attempt at an analysis. It would be even better to look at the results of sets of downs rather than drives. In particular, look at sets of downs that start within a certain time range, say 4:30 to 2:30. How did each conclude: first down, score, loss on downs, turnover, punt, failed kick? How much time did each use up? How did field position change?

If you wanted, you could also limit this by field position to exclude plays near both goal lines, which would eliminate the three runs and an intentional safety sequences and the three sneaks and a field goal ones, too.

The best thing about analyzing sets of downs is that it puts the focus on incrementally improving the chance of winning, rather than treating running the clock down to zero or not as a binary result. That naturally leads to balancing the risks of turnovers and clock stoppages against the rewards of additional plays to run.