Writers of Pro Football Prospectus 2008

Most Recent FO Features

SmithAle05-QB.jpg

» 2017 ALEX: Midseason Report

The latest ALEX update looks at the recent draft class that is struggling, the unusual Chicago strategy, and what's gotten into Alex Smith? We also looked at Tyrod Taylor's declining ALEX, but rising conversion rate that Buffalo just sent to the bench.

20 Sep 2016

Clutch Encounters: Week 2

by Scott Kacsmar

After a historic slate of season openers, Week 2 was also very competitive, as 11 of the games featured a comeback opportunity. This time, however, the defenses were more up to the task, as fourth-quarter lead changes dropped from 13 in Week 1 to just three this week. Strong defense was a central theme and the main explanation for how teams with Trevor Siemian, Case Keenum, and Sam Bradford at quarterback respectively beat and outdueled teams led by Andrew Luck, Russell Wilson, and Aaron Rodgers.

This emphasis on defense made it all the more surprising that Buffalo fired offensive coordinator Greg Roman on Friday, a day after a 37-31 loss to the Jets. Sure, under Roman, the Bills had gone 2-7 at game-winning drive opportunities, only having scored twice on 19 drives. That failure to close would upset any franchise, but enough to make a power move 18 games into the current regime? This is especially troubling when the problems are clearly more on defense after the No. 2 ranked unit of 2014 finished 24th last year under Rex Ryan. Notice how the Jets improved after replacing Ryan with Todd Bowles, and the Saints even looked better on defense this week now that ex-coordinator Rob Ryan has gone to Buffalo with his brother.

If this early one-eighth of the 2016 season has reminded us of anything, it is that coaching matters (Bill Belichick without his superstars), health matters (Russell Wilson's ankle), and a great defense can still control a game (the defending champion Denver Broncos). It is an embarrassing situation for Buffalo, heading towards a 17th-consecutive season without the playoffs, to scapegoat its offensive coordinator after two games, while ignoring the other problems going on with the team.

Game of the Week

Tennessee Titans 16 at Detroit Lions 15

Type: 4QC/GWD
Largest Fourth-Quarter Deficit: 12 (15-3)
Head Coach: Mike Mularkey (5-19 at 4QC and 5-20 overall 4QC/GWD record)
Quarterback: Marcus Mariota (3-6 at 4QC and 3-6 overall 4QC/GWD record)

Depending on the perspective, the Detroit Lions have almost been either the league's most impressive 2-0 team, or its most disappointing 0-2 team. On the one side, Detroit has done a fine job in overcoming Calvin Johnson's retirement, and has mostly controlled play through its first two games. On the other hand, the Lions blew an 18-point lead in Indianapolis and now a 12-point lead to the Titans as the defense has been shredded for 43.8 yards per drive. The actual path taken has led Detroit to a 1-1 question mark, but this could be a loss that comes back to haunt the Lions, who committed 17 accepted penalties at home.

Meanwhile, how about the uncharacteristic 12-point fourth-quarter comeback for the Titans? We have not seen that from Tennessee since Vince Young led a 21-point comeback over the Giants in 2006. This is only the fifth comeback win for coach Mike Mularkey, who still has the worst record in the league in those opportunities at 5-20. Of course, it helps when the first play of the fourth quarter was a 30-yard touchdown pass from Marcus Mariota to Delanie Walker. The Lions have surrendered a lot of big plays to tight ends to start this season.

During the offseason, we highlighted an obscure record for Matthew Stafford. In 2015, he became the first quarterback to ever complete 60 percent of his passes in each game of a 16-game season. Stafford's streak reached 18 games in Week 1, but ended four games short of the record streak (Tony Romo, 22 games) with a 22-of-40 day (55.0 percent) against the Titans. In particular, Golden Tate had arguably the roughest game of his career, catching 2-of-9 targets for 13 yards. Stafford just missed a wide-open Tate deep down the field on a big third-and-4 with 7:15 left.

Tennessee took over and started its long march from its own 17. DeMarco Murray's sixth touch on the drive set up a third-and-7 at the Detroit 12 at the two-minute warning. The Titans tried to free Tajae Sharpe up on a pick play, but it was not well done and almost guaranteed a fourth-down situation. Time ticked away as neither team had chosen to use any timeouts. This was odd, especially for the Titans, who would have been in dire straits had the next play failed. On fourth-and-4, Mariota threw a pass up for Andre Johnson that the veteran somehow hauled in for the go-ahead touchdown with 1:13 left.

This play was more of a parting gift from Johnson rather than signaling the arrival of Mariota. What a great catch. Unfortunately, the Titans failed on the crucial two-point conversion after Mariota's designed rollout led him right into a sack.

Detroit only trailed by a point and still had all three timeouts remaining. Sound familiar? This was the same situation the Lions faced in Indianapolis a week ago, except now the Lions had 36 extra seconds and the support of the crowd. Stafford usually does well in these moments, but Anquan Boldin did him no favor with Detroit's seventh dropped pass of the game. Stafford overcame a Brian Orakpo sack by finding Eric Ebron for 29 yards on third-and-19, but the Lions never gained another yard. On third down, Stafford did not recognize Perrish Cox leaving Boldin to jump the pass for Ebron, and the corner's interception clinched the Tennessee win.

Under Jim Caldwell, the Lions have lost an NFL-high five games with a lead of at least 12 points since 2014 (including playoffs).

Clutch Encounters of the Winning Kind

Dallas Cowboys 27 at Washington Redskins 23

Type: 4QC/GWD
Largest Fourth-Quarter Deficit: 3 (23-20)
Head Coach: Jason Garrett (20-30 at 4QC and 25-32 overall 4QC/GWD record)
Quarterback: Dak Prescott (1-1 at 4QC and 1-1 overall 4QC/GWD record)

With so many big personalities in Dallas, the coach is not always in the spotlight. Since officially taking over as the head coach of the Dallas Cowboys in 2011, Jason Garrett is 42-42 (including 1-1 in the playoffs). Garrett (2011-13) and Jeff Fisher (1996-98) are the only coaches in NFL history to finish .500 in three consecutive seasons with the same team. Whether it is with Tony Romo or one of the backups, Garrett consistently has Dallas in close games. Sixty-two percent of his games have featured Dallas with a 4QC/GWD opportunity, the highest rate among active coaches with at least six seasons of tenure.

Sure enough, both games this season have been close against two NFC East rivals, and the Cowboys have started 1-1. Washington has fallen to 0-2 and the blame has largely been placed on quarterback Kirk Cousins. Is that fair? In this game, he certainly botched a critical red zone sequence when Washington led 23-20 in the fourth quarter. After two unsuccessful fade passes, Cousins never saw safety Barry Church in the end zone and threw a bad interception with 10:35 left. While that cost his team potential points, Cousins was not on the field when rookie Dak Prescott calmly guided his offense 80 yards for the game-winning touchdown. Two penalties on Washington defenders saved Dallas from difficult third-down situations, and Cole Beasley made a key conversion on third-and-11 in the red zone. Alfred Morris capped off the drive with what had to be a satisfying 4-yard, walk-in touchdown against his former team with 4:45 left. Dallas led 27-23.

Cousins has had some issues in the past with interceptions in these moments, but here he dinked and dunked his way towards midfield. The decision to run the ball on third-and-1 was sound, but the failure to convert cost the Redskins more than 30 seconds as they chose not to hurry the fourth-and-1 play before the two-minute warning. In need of a touchdown, that was too much wasted time. On fourth-and-1, Cousins picked the right receiver (Pierre Garcon), but linebacker Justin Durant read the play well and knocked the pass down. Washington still had all three timeouts, but it was encouraging to see Garrett let Prescott attempt to throw on third down to seal the win. Prescott did not take a chance against the rush and slid down for an official sack.

Washington's task was difficult: drive 90 yards in 90 seconds without a timeout. Two big mistakes by the offensive line crushed Washington's effort. The first was a sack allowed to Tyrone Crawford, but the worst was center Kory Lichtensteiger flinching on the snap when Cousins wanted to spike the ball at the Dallas 36. Not only was that a 5-yard penalty for a false start, but a 10-second runoff left the Redskins with one Hail Mary shot at the end zone. Cousins failed to even keep the pass in bounds for any realistic shot of a game-winning catch.

Not only is Washington 0-2, but both losses were at home. Only the 1987 Colts and 2003 Eagles have made the playoffs after starting 0-2 with both losses at home.

New Orleans Saints 13 at New York Giants 16

Type: GWD
Head Coach: Ben McAdoo (1-0 at 4QC and 2-0 overall 4QC/GWD record)
Quarterback: Eli Manning (28-46 at 4QC and 35-48 overall 4QC/GWD record)

Forget last year's 52-49 shootout; we know the Saints are rarely that explosive away from the Superdome. But in the last four meetings between the Saints and Giants, the home team scored at least 48 points each time. So how does this end up as a 16-13 game with a blocked field goal returned for a touchdown helping out the Giants? New York's 9-point offensive outing was not for a lack of trying. The offense amassed 417 yards, but lost three fumbles, missed a field goal, and failed on a fourth down at the 3-yard line. With the Saints getting impact plays on defense, it was time for the offense to answer.

As we highlighted last week, win or lose, Drew Brees often has an answer. This time Willie Snead split the safeties and Brees found him for a 17-yard touchdown to tie the game at 10. In response, New York failed on three plays at the 1-yard line before kicking a 19-yard field goal with 8:46 left. Was that decision to kick a response to the low score, or the fact that the Saints had just stopped the offense three times? In most cases, the Giants should have tried a fourth-down attempt to try getting a touchdown lead on Brees with half a quarter to go, but Ben McAdoo played it safe. The Saints eventually tied the game on rookie kicker Wil Lutz's 45-yard field goal with 2:54 left.

Eli Manning is getting comfortable with his latest trio of wide receivers as rookie Sterling Shepard had his first 100-yard receiving game. But much like on the game-winning drive in Dallas last week, old reliable Victor Cruz returned to the forefront. After Odell Beckham Jr. dropped a potential touchdown, Cruz made a tougher catch down the sideline for 34 yards that probably helped the Giants more than a Beckham touchdown would have. With the Saints out of timeouts, the Giants made the very rare, but smart decision to take three kneeldowns to set up a 23-yard field goal as the final play of regulation. Josh Brown made the kick, and the Giants are 2-0 after better clock management than they had last season.

Manning's 35th game-winning drive moves him into 10th place all time.

Atlanta Falcons 35 at Oakland Raiders 28

Type: GWD
Head Coach: Dan Quinn (4-5 at 4QC and 6-6 overall 4QC/GWD record)
Quarterback: Matt Ryan (24-32 at 4QC and 33-33 overall 4QC/GWD record)

Oakland has joined the 1967 Falcons as the only defenses in NFL history to have started a season by allowing back-to-back 500-yard games. The Saints and Falcons are no slouches, but the Raiders were hoping for more on defense after bringing in Bruce Irvin, Sean Smith, and Reggie Nelson. A week after allowing 6 receiving yards to the tight ends in New Orleans, the Raiders allowed 180 yards to Atlanta's ungodly trio of Jacob Tamme, Austin Hooper, and Levine Toilolo. The defense also caught a bad break when a negative ALEX pass by Matt Ryan on third-and-6 was tipped into the air by Tevin Coleman and caught by Justin Hardy for the game-winning touchdown.

This defense has already forced Jack Del Rio to roll the dice more on offense, like going for the game-tying touchdown on fourth-and-2 in a situation where many coaches would have kicked a field goal. While Michael Crabtree came through with another score, not all of these plays are going to work out for the offense. Del Rio may have gambled too much when facing a fourth-and-2, down 28-21, at his own 49 with 7:16 left. Amari Cooper had just seen a touchdown wiped out after an illegal touching of the ball after he was out of bounds. To go for it here was not nonsensical, but trying to line up in a heavy formation and running Jalen Richard up the gut was. Richard finished the game with seven carries for 17 yards.

Atlanta only had to work half the field and added another touchdown to take a 35-21 lead with 4:35 left. The Falcons could have done a much better job of covering the middle of the field and tackling, but Oakland picked up a quick touchdown to keep hope alive. However, Ryan's third-and-3 conversion to Mohamed Sanu all but put the game away. Oakland had just two seconds left when it got the ball back, which was only enough time for a failed series of laterals.

The Raiders may not see another strong offense until Carolina in Week 12, but without improvements on defense, more shootouts are likely for this team.

Baltimore Ravens 25 at Cleveland Browns 20

Type: 4QC/GWD
Largest Fourth-Quarter Deficit: 1 (20-19)
Head Coach: John Harbaugh (18-37 at 4QC and 25-40 overall 4QC/GWD record)
Quarterback: Joe Flacco (18-34 at 4QC and 25-37 overall 4QC/GWD record)

This goes down as one of the stranger 20-point comebacks in NFL history. Cleveland is just the eighth team to score at least 20 points in the first quarter before getting shut out in the final three quarters. The last instance was when the 2011 Bills opened up a 21-0 lead on the Patriots before losing 49-21. Baltimore calmed down after the poor start and Josh McCown, filling in for the injured Robert Griffin III, cooled down immensely after a hot start that saw him find rookie receiver Corey Coleman for two touchdowns.

But the young Coleman also lost his cool in the fourth quarter, getting caught in a post-play skirmish with Jimmy Smith that led to a 15-yard penalty. Those were big yards as kicker Patrick Murray missed a 52-yard field goal two plays later. Justin Tucker, one of the best kickers in the game, came through with a 49-yard field goal that put the Ravens ahead for the first and last time at 22-20.

For McCown, this game unfortunately had the familiar ending of pain and misery. He injured his shoulder during the game, but continued to play on even though the injury is serious enough that he could miss several weeks moving forward. The refusal to come out may have been a mistake as Cleveland did not score on any of its final eight possessions. Trailing 25-20 with 2:53 left, McCown had one last shot at a 75-yard touchdown drive to win the game. Too much time was consumed on short gains, but an impressive 20-yard catch by Terrelle Pryor should have put the ball at the Baltimore 10 with 20 seconds left. However, a defensive holding penalty on Lardarius Webb was offset by a bogus taunting penalty on Pryor, negating the play. All Pryor did was toss the ball without ill intent, but the referee made little effort to catch it before it tapped Webb.

Little things like that should not contribute to deciding games.

On the next play with the ball back at the Baltimore 30, McCown threw a terrible, flat-footed pass on first down into triple-coverage and C.J. Mosley came away with the game-clinching interception. The 37-year-old McCown is an unfathomable 5-29 (.147) at game-winning drive opportunities in his career. Now Cleveland is set to start rookie Cody Kessler in Week 3. If this was the end for McCown, at least we will always remember his first success: the 28-yard touchdown to Nate Poole to knock the 2003 Vikings out of the playoffs.

Clutch Encounters of the Losing Kind

Bengals at Steelers: Have You Ever Seen the Rain…or a Fumble?

Most football purists will always believe the game should be played outdoors in the elements. I would prefer to see a league with 32 retractable-roofed stadiums that can stay closed on days like Sunday in Pittsburgh, where the rain turned the field into a familiar mess and caused havoc with ball security. Both teams were plagued by some bad drops on both sides of the ball, and it was only fitting that a "fumble" led to the young season's most controversial ending in the latest chapter of this heated rivalry.

While Pittsburgh's offense was sloppy for much of the day, a long, fourth-quarter touchdown drive seemed to put the Bengals away at 24-9 with 6:48 left. To Cincinnati's credit, the Bengals came right back with a quick touchdown drive, leaving them trailing 24-15 with a very difficult decision to make.

Now a common argument for teams down by nine points after scoring a touchdown is that they should always go for two right away since they need a conversion eventually, and by doing it earlier, they'll know how to plan the rest of the game. I tend to disagree with this strategy (strongly in some cases), and believe more in extending the game to get to that point where you can tie things up with a two-point conversion. In this particular case, Cincinnati's red zone offense had not looked good, and there was no rushing threat. I would not have felt confident in my offense to convert at a high rate. Then there is the issue of timeouts. If the Bengals had none or one and failed on the attempt, they would have been down nine points with 3:25 left, needing to somehow create two more possessions. That could have meant two onside kick recoveries,when it's already very difficult to recover one when the receiving team knows it's coming. If Pittsburgh had the ball with 3:25 left, it could have run three times and punted, and the Bengals would have been pinned deep with maybe 70 or 75 seconds left, still down 24-15. That is a terrible situation, and a perfect example of why you should not always go for two first.

But the Bengals had all three timeouts, making it more reasonable to go for two first. I still agree with Marvin Lewis' decision to kick the extra point. After forcing a Pittsburgh punt, the Bengals were 75 yards away with 3:00 left, which is plenty of time to score a touchdown and go for two, and then try for an onside kick and game-winning field goal should that two-pointer fail.

But none of those dramatic moments came to fruition after Tyler Boyd, a rookie from Pitt, apparently fumbled at the Pittsburgh 33 with 1:50 left. Pittsburgh's recovery all but ended the game, but the review looked probable to get overturned. The original call on the field was defensible as it looked like a fumble in live action, and we know conclusive evidence must be shown to overturn the call on the field.

I thought the evidence was pretty conclusive that Boyd's knee was down before the ball was loose. James Harrison's right knee contact is what knocked the ball out, but that happened several frames after Boyd's knee was down.

Dean Blandino, the NFL's senior vice president of officiating, disagreed.

Even if Boyd had been ruled down, the Bengals still had some decent work ahead of them just to force overtime, so this was hardly an all-time "the refs stole the game!" moment. But it will be remembered in the rivalry for sure. I personally think the Bengals got a raw deal here.

Colts at Broncos: Whoops, They Did It Again

For as good as Andrew Luck looked in Week 1, the Broncos are the NFL's current litmus test for facing a great defense. It was a test Luck aced better than anyone in 2015, but playing Wade Phillips' unit at Mile High is a different story. After one of the most ineffective halves of his career (43 net passing yards and a 5-of-17 success rate), Luck led a long touchdown drive and had the Colts right back in another tight one, trailing 16-13 to start the fourth quarter.

While Colts coach Chuck Pagano talked about ending his team's "last-minute culture" this week, the Broncos have fully embraced this style of play. We know Denver's not going to score many points, especially now with a young Trevor Siemian at quarterback. Fifteen of Denver's 17 wins under Gary Kubiak have required the defense to shut the door in the fourth quarter with a one-score lead. We have highlighted that table many times in the past, including all of the memorable takeaways and sacks that have put games away for the Broncos.

Even though this is a new season, Denver is continuing to win in the same fashion as it did last year. On a third-and-15, Aqib Talib beat Phillip Dorsett to the ball and returned it 46 yards for a pick-six and 23-13 lead. As crushing as that was, the Colts regrouped for an 80-yard touchdown drive, but the pass defense struggled to keep the Broncos from getting into scoring range. Kubiak was faced with a fourth-and-1 at the Indianapolis 17 with 1:55 left and the Colts out of timeouts. One yard and the game is over, which sounds like a favorable situation for the offense. In 2015, non-quarterback runs converted 67 percent of fourth-and-1 plays, and passes are always around 54 percent successful. By kicking the field goal to take a 26-20 lead, Denver was inviting Luck to drive for the game-winning touchdown, utilizing four downs all the way. If they had gone for it on fourth down and failed, the Colts would have taken over down 23-20, and it would have been almost inevitable that the Colts would get conservative at some point and settled for a field goal and overtime. With nearly a full two minutes left and the likelihood of a touchback in this altitude starting the Colts at their 25, I hated the 35-yard field-goal decision.

Of course, the Denver defense immediately rendered all that moot when Von Miller blew past Joe Reitz to get to Luck and cause another fumble return touchdown for his unit. Shane Ray really should have gone down right away to end the game with kneeldowns instead of further risking any injuries, and the most improbable of Luck-led comebacks, but good luck telling a defender not to score when the opportunity is right there. These opportunities just keep coming to Denver at an incredible rate. Miller even capped off the 34-20 win with another sack.

Packers at Vikings: The Other Passing Game Steals the Show

Recent games between the Vikings and Packers have usually been billed as Adrian Peterson trying to keep the ball away from Green Bay's passing attack led by Aaron Rodgers and Jordy Nelson. But in opening Minnesota's new stadium on Sunday night, the Vikings were led by the passing connection of Sam Bradford and Stefon Diggs -- not exactly how things would have been envisioned a year ago. But the two hooked up nine times for 182 yards and a touchdown in what was legitimately the high point of Bradford's aimless career. With Peterson struggling (19 yards on 12 carries) and suffering a major injury, and the timing issues Green Bay's duo is still having, this was an impressive showing and role change for the Vikings.

Still, Mike Zimmer's defense was also on display, forcing Rodgers into one of the more ineffective games of his career, which is starting to become all too common. It could have been even worse as Terence Newman dropped an interception in the red zone two plays before a Rodgers scramble touchdown cut Minnesota's lead to 17-14. Rodgers set a career-high with three fumbles, but only the third one was recovered by the Vikings with 6:53 to play. The defense still needed one more stop at midfield after the two-minute warning, and after getting picked on all night for his terrible ball awareness, second-year cornerback Trae Waynes made his first interception a memorable one by picking off Rodgers with 1:50 left. Minnesota let Bradford throw to Diggs one more time on a crucial third-and-6, and a "favorable home-team call" of pass interference on Damarious Randall all but salted this one away.

Green Bay failing to win a close game is old hat. Mike McCarthy is now 22-46-1 (.326) at game-winning drive opportunities, including a 14-35 (.286) record with Rodgers at quarterback. Both records rank as the third worst among active coaches or starting quarterbacks. The more recent development is the broken passing game that used to be historically efficient. For a change, Minnesota had the better passing game.

Seahawks at Rams: Jeff Fisher Eyes Extension After Quasi-Super Bowl Win

Despite how dreadful the Rams looked on Monday night in a 28-0 shutout loss to the 49ers, this outcome was still fairly predictable if you have been following the NFL closely. In the past four seasons, Jeff Fisher's 7-9 B.S. had split the eight contests with Seattle's DVOA dynasty, and only the 2013 finale was not a close game in the fourth quarter. The Rams treat Seattle like their bi-annual Super Bowl, and with the first game back in Los Angeles and Russell Wilson's high-ankle sprain, a 9-3 victory was not far-fetched at all.

For just the second time in his 76 career starts, Wilson's Seahawks failed to score 10 points. While he limited his sacks to two and stayed turnover-free, his play was ineffective, obviously hampered by the injury and lack of running game to support him against a strong defensive line. But the defense kept the game close and Seattle had three cracks in the fourth quarter at erasing the 9-3 deficit. The final drive, which needed to cover 88 yards, started to look like shades of last week against Miami when Wilson helped pull one out for the Seahawks at what seemed like their lowest point. Down to 1:53 and one timeout, Wilson got things started with a 53-yard bomb to Tyler Lockett, who beat Troy Hill in coverage.

Even with all of the injured skill players and disadvantages in the trenches, Seattle was in good position to deliver a last-minute win. But on third-and-10 at the Los Angeles 35, Wilson's dump pass for Christine Michael was fumbled after an 8-yard gain and hit by Alec Ogletree, who made the recovery with 45 seconds left. Like that, the Rams had their first win in Los Angeles despite not registering a single touchdown in eight quarters. Unfortunately, Seattle can only boast one touchdown (born out of desperation a week ago) in eight quarters itself. While slow starts have happened in three of the last four seasons for Seattle, there was never this much concern with Wilson's health, presenting a new challenge for this team in its attempt to remain a top contender.

Dolphins at Patriots: Different Tenants, Same House

Injury ended a fantastic game from Jimmy Garoppolo in his first home start, but the Patriots still extended the lead to 31-3 in the third quarter with third-string quarterback Jacoby Brissett filling in the rest of the way. New England is about the last team expected to blow a 28-point lead at home. Since 2001, opponents are 1-49 at fourth-quarter comeback opportunities in New England. Only Eli Manning led the Giants to a win back in 2011.

But if there was a formula for a huge comeback to happen, Miami had most of the ingredients. New England's offensive line started to struggle with sacks and a holding penalty with a rookie quarterback in the game. Ryan Tannehill got hot in the no-huddle offense, leading three consecutive touchdown drives with the receivers making plenty of plays against the secondary to cut the lead to 31-24.

Yet even with several rookies in the lineup and no Rob Gronkowski, the Patriots still looked like the Patriots. Brissett even converted twice on the quarterback sneak. But when Stephen Gostkowski had a chance to put this one away with a 39-yard field goal, the great kicker shockingly pushed the kick to the right. Miami had one last chance, needing to drive 71 yards in 64 seconds. After getting as far as the New England 29, Tannehill was nearly intercepted on one sideline throw before getting picked in the end zone by Duron Harmon with two seconds left to secure another New England victory.

Season Summary

Fourth-quarter comeback wins: 12
Game-winning drives: 14
Games with 4QC opportunity: 24/32 (75.0 percent)
10+ point comeback wins (any point in the game): 6

Historic data on fourth-quarter comebacks and game-winning drives can be found at Pro Football Reference. Screen caps come from NFL Game Pass.

Posted by: Scott Kacsmar on 20 Sep 2016

43 comments, Last at 22 Sep 2016, 4:21pm by PatsFan

Comments

1
by jablo1312 :: Tue, 09/20/2016 - 4:16pm

I don't really understand the logic behind the argument that the Bengals should not have gone for 2. Yes, you're right, if they were down 24-15 with 3:30 or whatever left, they would be in pretty dire straits. But if they kick there, and then score again to get the score to 24-22 with what, like 20 seconds left, there in just as awful straits. Unless you can think of a real reason for why a team would be more or less likely to convert a 2 point try down 9 points with 3:30 remaining as opposed to down 2 points with 20 seconds remaining, I don't see why they shouldn't try to put the 2 on the board as soon as they know they're going to need to do it at some point.

3
by JIPanick :: Tue, 09/20/2016 - 4:37pm

I don't understand it either.

4
by RickD :: Tue, 09/20/2016 - 4:47pm

I think it comes down to this:

In this particular case, Cincinnati's red zone offense had not looked good, and there was no rushing threat. I would not have felt confident in my offense to convert at a high rate.

The decision to go for 2 depends not only on the scoreboard, but the probability of conversion. If you think it's very low, you should avoid it until it's the only possibility.

The flip side is that if you think it's particularly high, you should always go for two.

5
by RickD :: Tue, 09/20/2016 - 4:54pm

I should make that always* - if it's a tie game and there's no time on the clock, obv. you go for the PAT unless, bizarrely, your success rate at the 2-point try is higher (ordinarily the threshold would be at half that probability).

7
by Scott Kacsmar :: Tue, 09/20/2016 - 5:35pm

I think we can all agree that a team down by 15 points late will be playing for the tie (and probably OT). I think we can pretty much agree that your odds of converting that 2PC are basically the same whether you're down by 9 or 2. If there is an argument for it being slightly higher at the end of the game down 2, it would be that we have more proof that you can score on this defense since that means the second touchdown was achieved.

Having said that, I know there are more 8-point comebacks than 9-point comebacks in this game, and I'd rather have that opportunity to be a 2PC away from the tie than to try making up a 9-point deficit. For a league with coaches who fear to be criticized when something goes wrong, this is why the vast majority kick the XP. You do not want to be accused of forcing your team into a hopeless situation sooner than you had to.

But it is also worth pointing out that teams scoring a touchdown when down 15 in the 4Q are 5-97* since 1994. That includes an 0-10 record by the teams that went for two first (3/10 successes). So this is probably a good reminder that no matter when you choose to go for two, being down 15 in the fourth quarter is a bleak situation.

*One of the five winners actually scored with a kick return TD to get to 31-23 after the extra point. That was the 2004 Colts against San Diego. Manning's 49th TD pass of the season and a 2PC run by Edgerrin James forced OT and the Colts won on an opening-drive FG.

8
by Denverite :: Tue, 09/20/2016 - 5:42pm

Isn't there an argument that if you're going to need three scores, you want to know it sooner rather than later, because you can do a lot more to try to engineer a third possession if you know you'll need it with 10 minutes left, as opposed to 10 seconds? (Go hurry-up, try onside kicks, etc.).

12
by JIPanick :: Tue, 09/20/2016 - 6:50pm

It's the only argument that makes sense to me. Unless there is some reason to think that the second potential 2PC would be easier than the first, from a purely mathematical win-chance perspective you need to know ASAP whether you need just one more possession or two more.

I expect that the default decision to kick is based on concerns with some combination of job security (the later you finish losing, the better the optics) or possibly team morale (you can lose players if you do mathematically correct but unconventional things without buy-in and it doesn't work). The latter is me putting a charitable spin on it; I suspect it is the almost always the former.

30
by techvet :: Wed, 09/21/2016 - 1:15pm

That's been my thought. If I am going to fail on the 2-pt conversation in the 4th quarter, I'd rather know sooner than later.

9
by Scott Kacsmar :: Tue, 09/20/2016 - 5:44pm

Also worth pointing out: one of those three successful 2PC's down 9 was 100% useless. In 2007, Chad Pennington threw a pick-six to fall behind 38-23 in the final minute. He threw a touchdown as time expired, then added a 2PC pass for a 38-31 final. There was really no incentive for the Bengals to even defend the 2PC; the game was over.

10
by Anon Ymous :: Tue, 09/20/2016 - 5:46pm

*Having said that, I know there are more 8-point comebacks than 9-point comebacks in this game*

In order to be down by 9, the missed 2-pt conversion must be certainty, so you can't compare it to a situation where the opportunity to convert remains.

11
by hoegher :: Tue, 09/20/2016 - 6:09pm

That's a good point, what's the W/L record for teams that fail on the 2-pt try after the second TD?

35
by Eddo :: Wed, 09/21/2016 - 5:21pm

This, and also, it stands to reason there are more seven-point comebacks than eight-point comebacks.

What you really need to do is compare the down-seven, down-eight, and down-nine scenarios. If, hypothetically, the likelihood of seven-, eight-, and nine-point comebacks are, respectively, 50%, 40%, 10%, then Scott might be right, as going from eight-down to seven-down isn't as important as going from nine-down to eight-down. But if the likelihoods are 50%, 20%, 10%, then you do want to go for two.

37
by Scott Kacsmar :: Wed, 09/21/2016 - 8:57pm

I didn't read through this again, but there is a bunch of down 7 vs. 8 vs. 9 stuff in here http://www.footballoutsiders.com/clutch-encounters/2013/clutch-encounter...

Seems to be focused on the final 2:00 though, but still useful. Teams getting the ball, down 9-11 in the final 2:00, had a record of 1-204.

2
by ramirez :: Tue, 09/20/2016 - 4:19pm

"Forget last year's 52-49 shootout; we know the Saints are rarely that explosive away from the Superdome"

I wonder what Scott's explanation for that is, given he maintains that domes don't help passing statistics.

6
by Scott Kacsmar :: Tue, 09/20/2016 - 5:09pm

Road teams in general will rarely impress as much as they do at home, but let's be honest. The dome argument has always been a poor attempt to try downgrading the careers of Peyton Manning, Kurt Warner and Drew Brees, all while ignoring the countless dome offenses/QBs that weren't any good. Manning's Colts, Warner's 1999-2001 Rams and the Brees-Payton Saints are really the only three dome offenses that have sustained greatness in this league. The Vikings were also strong in the late 90's with arguably the most gifted WR ever (Randy Moss) as well as players like Cris Carter, Robert Smith and Randall Cunningham/Daunte Culpepper. The 1998 Vikings scored 274 points on the road, 5th most in a season.

But it really comes back to Manning, Warner and Brees. Meanwhile, if you look at the 25 highest scoring teams in home games and road games, there are more indoor teams in the road list (7) than there are in the home list (6). Manning went to Denver and continued to run some of the most prolific offenses on record, home and away. The GSOT Rams rank in the top 25 in road points scored in a season for 1999-2001. The 2009 Saints rank 13th on the road, but in recent years, things have cooled off for the Saints. When you combine the road performances with the absurd home performances Brees has in prime time (high visibility games), it creates the home-road perception the Saints have. When they have games like Sunday, or the stinkers last year in Philadelphia and Washington, it gives support to that perception.

13
by ramirez :: Tue, 09/20/2016 - 8:47pm

As I knew he would, Kacsmar has made an unconvincing argument that domes don't matter. He does this because he knows that if he admits that domes help passing numbers, it will hurt Peyton Manning's case to be considered the best QB of all time. All I have to do to show that domes matter is to show that QBs, when playing in a dome, will produce better stats than in outdoor games. The fact that there have been dome teams with weak offenses doesn't have anything to do with it. I never said that every dome team has a great offense.

The crux of his argument in the past has been that Manning put up strong numbers while playing outdoor home games in Denver. But it's obvious why this is so. It's because when we look at the outdoor home split of Manning's career, we're looking exclusively at his time with the broncos, when his overall numbers were great. We can observe a similar effect when we look at the career of Drew Brees. Brees has a ypa figure in home outdoor games of just 6.75, and a ypa figure in road outdoor games of 7.26. Presumably, Kacsmar believes that is because Brees has some fundamental inability to play well in outdoor home games. But that's not the real reason. It's because the home outdoor split for Brees only includes his time in San Diego, when his overall numbers were weaker. When you're looking at a split that only covers a portion of a guy's career, you can run into this problem.

Virtually all QBs post better numbers indoors, even guys like Brady and Rodgers, who have never played a home game in a dome. When you look at the entire careers of guys like Warner and Brees, they have much better numbers indoors than outdoors. Cherry-picking examples from individual seasons isn't going to change that fact.

14
by Scott Kacsmar :: Tue, 09/20/2016 - 9:29pm

I had several paragraphs done for a long reply, but screw it, it's not worth it right now, so let's condense.

No one has put together a convincing dome study. What you threw together here sure as hell doesn't pass for one. A legit dome study would include air yards and YAC. It would acknowledge the fact that only a few of the same teams play in a dome year after year, tainting the sample, including some of the NFL's more consistent losers in Detroit and Atlanta (and New Orleans prior to 2006). It's not like we get to see 32 teams alternate between indoors and outdoors. If you have 8 primary indoor teams (circa 1995-2015) and 3-4 of them are only relevant with a HOF-caliber QB and 4-5 of them have routinely poor defenses, then how the hell is that not going to skew the indoor passing stats? For the 24 outdoor teams, do they have 12 HOF-caliber QBs and 14 lousy defenses? Of course not. A legit study would look at the difference in success the defense has on the road versus at home, and then compare that to how outdoor teams fare on the road vs. home. Yes, dome teams usually have bad defenses, but I bet you they're still bad on the road too and it's a personnel issue more than anything. The Vikings aren't struggling right now. The Saints were great with the Dome Patrol era in Jim Mora's tenure.

This topic is much better suited for the offseason, and a stat like passing plus-minus would be very beneficial to a dome vs. outdoors study.

19
by carybird :: Wed, 09/21/2016 - 2:25am

"The crux of his argument in the past has been that Manning put up strong numbers while playing outdoor home games in Denver. But it's obvious why this is so. It's because when we look at the outdoor home split of Manning's career, we're looking exclusively at his time with the broncos, when his overall numbers were great. [...] When you're looking at a split that only covers a portion of a guy's career, you can run into this problem."

Is this for real? First of all, his overall numbers were great his entire career, not just in Denver. Second of all, what's your point? That overall he was a great QB in Denver? Yes, he was. Where he didn't play in a dome. So him playing in a dome was irrelevant to him being great because he was just as good when he moved outside. So what Scott has been saying this whole time. You literally just proved Scott's point: Peyton was a great QB in Denver, therefore the dome isn't what made him great.

21
by ramirez :: Wed, 09/21/2016 - 8:39am

You're not getting my point. You need to ask why it is that so many other players have better numbers in road indoor games than in home outdoor games. If the dome effect isn't real, we would expect that not to be the case. The reason why Manning's home outdoor splits look good is because they only cover a small portion of his career, during which he was performing at an exceptional level overall. Manning's numbers in Denver, particularly 2012-2014, were better than his last 3-4 years in Indy. Look up the numbers for yourself. Read my other posts, take a close look at the evidence, and you'll see that my argument makes sense.

22
by dmstorm22 :: Wed, 09/21/2016 - 9:50am

Your Manning point at the end so is ridiculously stupid.

Scott pointed out that Manning was great in Denver when he was playing 8+ games outdoors.

Your rebuttal was that the reason we can't use this as evidence that playing in a dome did not make Manning great was that he was so great in that period?

So basically, if Manning was slightly worse than he actually was in 2012-2014 it would be hurt your argument?

As for Brady, he's played 20 games in domes/retractable roof stadiums. We also have to look at who those teams were. That is a very small sample size, and includes games against the expansion era Texans, the Manning / Luck Colts (only once did he play a great Colts defense in Indy, and he struggled in 2007), the Vikings one time, and a few games here or there. If he was playing the dome patrol and the early-90's Vikings in those 20 games, his numbers probably wouldn't look so great.

25
by ramirez :: Wed, 09/21/2016 - 10:19am

Look, Manning has better numbers on the road indoors than outdoors, like many other QBs. This suggests that he gets a boost from playing indoors. If he had played outdoor road games at the beginning of his career, when his overall numbers were weaker than when he was in denver, it's very likely his home outdoor splits would be weaker. But they're not, because those weaker seasons at the start of Manning's career don't figure in his home outdoor splits, only his time in Denver. It's not complicated. You're forgetting that my argument isn't solely based on Manning's numbers. It's based on putting Manning's splits into the wider context of other QBs and their career splits. The vast majority of QBs have better numbers indoors than outdoors. Manning's home outdoor splits appear to be an exception, but what I am arguing is that it's a mirage, created by the circumstances of his career. The home outdoor split for Manning only counts his time in Denver, when his overall numbers were fantastic. He didn't maintain that rate of production for his entire career. That's why Manning's home outdoor splits look good compared to so many other players and their home outdoor splits.

27
by Scott Kacsmar :: Wed, 09/21/2016 - 10:29am

Honest question: do you know how many NFL teams play in a dome, and which ones?

Manning played 15 road indoor games with the Colts, and 5 of them were the Texans (years before J.J. Watt). If you can't see how that sample is filled with bias, then I don't know what to tell you. Yeah, he tossed 6 TD on the 2003 Saints and 2004 Lions. He was also in his prime and those defenses stunk. But those games are going to inflate the totals too.

And I'd love to hear your explanation for why two of Manning's worst games with Denver in 2012-14 were in domes (2012 Falcons and 2014 Rams).

28
by ramirez :: Wed, 09/21/2016 - 12:06pm

Scott's explanation for why Brees and Warner have a huge indoor/outdoor split is that they are great passers, which doesn't even try to account for the disparity. If it was just because of their skill level, why would they look better in a dome to begin with? Perhaps it's just a statistical anomaly, but I seriously doubt it.

And when discussing Manning, Kacsmar is just doing what he always does. He's cherry-picking individual games where Manning played poorly in a dome, which utterly fails to explain why Manning, along with many other QBs, have better numbers indoors than outdoors. The fact that 3 people have dismissed my points just shows that they refuse to address what I have said directly. They haven't convinced me that my interpretation of the numbers is wrong. They've just belittled me for having the audacity to ask why, for example, Manning looks so good in outdoor home games.

It's at moments like this that I shake my head, and ask myself why I even try to reason with people like Kacsmar and his ilk.

32
by Scott Kacsmar :: Wed, 09/21/2016 - 2:20pm

Will Ferrell and Adam Sandler should take notes. /\ /\ This is what comedy looks like.

Did you really just say I cherry-picked by pointing to the only two dome games Manning played in 2012-14?

15
by ramirez :: Tue, 09/20/2016 - 9:48pm

Dude, if you're not even going to acknowledge the specific points and stats I discussed, then this is not worth pursuing further. You haven't acknowledged the things I said about Manning and Brees' career splits. If domes don't matter, why do indoor and outdoor QBs virtually all put up better numbers indoors? And why did they call the Rams the Greatest Show on Turf, if not as an acknowledgment that domes help offense? Also, when did I say that dome teams have bad defenses? That doesn't explain the better numbers for guys like Brees and Warner, unless you're arguing that playing in a dome causes those guys' opponents to play weaker defense. And if that's the case, isn't that just another way of saying that domes help offense? I don't get it.

16
by Scott Kacsmar :: Tue, 09/20/2016 - 10:51pm

Do I really need to go point by point when your history shows you want threads to devolve into the Irrational, and you put words in my mouth about Brees' SD tenure? And you seriously just came up with the worst explanation for the GSOT nickname ever.

Like I said, a complete study would factor in what happens to the defense too.

17
by ramirez :: Tue, 09/20/2016 - 11:00pm

Not only is this guy a terrible analyst, but he's a jerk with an attitude as well. I don't know what your problem is with my point about Brees, who has weak numbers in outdoor home games from his SD tenure. That's a fact. Again, I guess your argument is that Brees and Warner faced a slate of bad defenses whenever they played home dome games. Otherwise, how do we explain the disparity in the numbers?

18
by Scott Kacsmar :: Tue, 09/20/2016 - 11:25pm

You don't have a Brees point. Why would anyone care about what he did before 2004 when he wasn't a good QB? Of course his splits would be lousy then. In 2004, Brees was great in any split. In 2005, he played one game indoors (the one that knocked off 13-0 Indy) and struggled. So what is San Diego supposed to prove?

20
by ramirez :: Wed, 09/21/2016 - 8:33am

Let me take the chance to explain my points again, since Kacsmar and carybird aren't getting it. Kacsmar argues that Manning's strong performance in outdoor home games proves that the dome effect doesn't exist. But this is wrong, because Manning's strong numbers outdoors in Denver are a function of the fact that his overall numbers in Denver were very good. The point about Brees is that he has a similar effect in his career with regard to outdoor home games. Brees has better numbers in outdoor road games than outdoors at home, but that's because his outdoor home split only covers his time in San Diego. Again, when you're dealing with splits that only cover a portion of a player's career, you can get results like this that point to incomplete conclusions.

Now, let's take a look at some more splits for Brees and Kurt Warner, two QBs who have played much better in a dome than outdoors. Warner is at 8.17 yards per attempt in home games, 7.74 on the road a difference of 0.43 ypa. Brees is at 7.76 ypa in home games, and 7.31 on the road, a difference of 0.45 ypa. So we can see that both guys have a similar home/road split. But when we look at their indoor/outdoor splits, and count retroroof games as indoors, we get a much bigger disparity. Warner is at 8.43 ypa indoors, and just 7.41 outdoors, a difference of 1.02 ypa. Brees is at 7.91 ypa indoors, and 7.12 outdoors, a diff of 0.79 ypa. So, it's clear that the home/road split alone isn't enough to explain the better indoor numbers for Warner and Brees.

When you also factor in that players like Brady and Rodgers have better numbers at home than on the road, but also have much better numbers in a dome than outdoors, despite never playing any indoor games at home, it points to the conclusion that domes help passing stats. Perhaps that's not the case. I agree with Scott that someone needs to do a comprehensive study to know for sure. Kacsmar claims that these splits are the result of the fact that dome teams usually have bad defenses, but I don't think that's a sufficient explanation. At a minimum, it's silly of him to dismiss the evidence and claim that domes don't matter, when Kacsmar agrees that we need more data.

23
by Scott Kacsmar :: Wed, 09/21/2016 - 9:53am

This is the problem. You start with something and continue repeating it even after people have already explained why it doesn't make sense.

I'm just going to leave this here from the data I compiled during the SB study. Let's forget about the retractable roof teams for a second and focus on the consistent indoor teams in the DVOA era. I ranked them in their indoor years in DVOA rank compared to the rest of the league.

Average defensive DVOA rankings since 1989
21. 1995-2015 Rams, 16.8
23. 1989-2013 Vikings, 17.1
26. 1989-2015 Saints, 18.9 (2005 excluded)
27. 1989-2015 Colts, 19.0 (19.1 if we only go RCA Dome years, 1989-2007)
30. 1989-2015 Lions, 20.2
32. 1992-2015 Falcons, 21.9

Yes, we would want to look at home/road splits, pass/run splits, but this definitely suggests that dome teams do not field good defenses. So when you tell me about Brady and Rodgers having great numbers in domes, you're basically telling me they're good against these select (mostly crappy) defenses. And it's still a small sample size of games for them. We have seen them look like crap in a dome before too, including Sunday night for Rodgers. Go figure, the Vikings actually look to have a defense.

829 teams since 1989, right? 201 qualify as indoors, including the retractables.

I only count 19 top 5 defenses from the indoor teams since 1989

Seahawks - 1991, 1992
Oilers - 1993
Rams - 1999, 2001
Vikings - 1989, 1990, 1992, 1993, 2008
Saints - 1991, 1992
Colts - 2005, 2007
Cardinals - 2013, 2015
Texans - 2012
Lions - 2014
Falcons - 1998

If we assume every team in the league has an equal shot at finishing as a top 5 unit, no matter if they're indoor or outdoor, then take a look at this.

Top 5 Defense
Indoor - 9.5% of teams
Outdoor - 18.4% of teams

Top 5 Offense
Indoor - 12.0% of teams
Outdoor - 17.6% of teams

I've never said domes don't matter. I've said it's downright stupid to diminish a QB because of it.

Correction: 200 indoor teams. I left the 2005 Saints in there by accident. Doesn't change anything significantly though.

24
by ramirez :: Wed, 09/21/2016 - 10:07am

Again, you won't address my points. Warner and Brees have great dome numbers because they always face weak defenses when playing indoors? Is that your official position? Because what you posted doesn't explain why Warner and Brees have much better ypa figures indoors than outdoors.

26
by Scott Kacsmar :: Wed, 09/21/2016 - 10:23am

Three people have already dismissed your points for foolishness. Warner and Brees have great stats because they were/are great passers. That's my official position.

31
by tammer.raouf :: Wed, 09/21/2016 - 2:11pm

Point of clarification here:

Are you saying that Peyton's numbers in Denver can't be used to argue that he is good outdoors because he was just playing very well during that period compared to his career at large?

33
by dmstorm22 :: Wed, 09/21/2016 - 2:30pm

Yes, he is arguing that we cannot use the fact that Peyton Manning was really good at home when 'home' meant outdoors because he was, in fact, really good at home. And why? Because he was too good at home.

34
by ramirez :: Wed, 09/21/2016 - 4:02pm

dmstorm22, this is now the 2nd time you have misrepresented my postion. I assume you are just here to troll, and don't have an answer for my actual point. But I'll be generous and give you one final chance to grasp the distinction between what I actually said, and what you claim I said.

Most QBs perform better in dome road games than home outdoor games, which suggests that domes inflate passing stats. One of the exceptions is Peyton, who has really good home outdoor splits. So that raises the question, why does Peyton do better by that particular split? I believe the answer is because when we look at Peyton's home outdoor split, we're only seeing the portion of his career from 2012-2015. His overall numbers in that era were fantastic, which is why his home stats look so good. Peyton didn't match those numbers in earlier portions of his career like 1998-2002. So Peyton's home outdoor split represents only a portion of his career, when for many guys that split represents their whole career in home games. So, in effect, Peyton's home outdoor numbers are artificially inflated by the fact that for much of his career, he didn't play any home games outdoors. That's why Peyton's home outdoor split doesn't work as evidence that domes don't matter.

If all you guys have in response is an effort to misstate my position, or spout nonsense about Ferrell and Sandler, stop wasting my time.

36
by Eddo :: Wed, 09/21/2016 - 5:47pm

It's fine to dismiss Manning's career as merely being one data point, but it's still a data point that goes against what you're asserting. Manning - post-major injury, and in his late 30's - moved from a dome stadium to an outdoor stadium, and his numbers got better. That singular data point might not prove that there is no advantage to playing in a dome, but it certainly moves the needle in that direction, if only slightly.

Therefore, it's on you to provide more data to show that playing home games in a dome inflates your numbers (the reasonable null hypothesis is that it has no effect). You've provided Brees, but he, like Manning, is just one data point.

38
by nat :: Thu, 09/22/2016 - 10:02am

Hmmm...

Wasn't Manning's whole point of getting surgery to get healthier and thus play better? Wasn't the point of choosing Denver to have better weapons and to ditch the depleted Colts?

Manning's move from the Colts to Denver is a poor case to examine when looking for dome effects or their absence. Instead, it's an example of the kind of noise that makes this kind of analysis difficult.

It would have been a stunner if his numbers hadn't improved. He went from injured, worn out, unprotected, and weaponless to healthy, rested, protected, and armed to the teeth.

BTW: DVOA already recognizes that playing indoors skews offensive performance.

Offense gets a slight penalty and defense gets a slight bonus for games indoors. (see http://www.footballoutsiders.com/info/methods#DVOA)

40
by dmstorm22 :: Thu, 09/22/2016 - 12:08pm

I won't argue the subjective of weapons etc. No one was calling Denver a team with great weapons when Tebow was yukking it up the year before.

However, to say he picked Denver and 'ditched' the Colts is a bit wrong given that he was cut and didn't ditch anyone. If anything, he probably would have rather stayed in Indianapolis over anything.

41
by Eddo :: Thu, 09/22/2016 - 12:30pm

Thanks for the info on how DVOA treats it.

re: Manning

I don't think you can expect a player in his mid-30's, coming of a serious neck injury - even if he has surgery - to put up better numbers than he did in his 20's, regardless of the weapons around him. It's possible, but to expect it? No way.

He had surgery to play better compared to when he was injured. You wouldn't expect a veteran running back coming off surgery to repair a torn ACL to come back and put up better numbers than ever before, would you?

29
by techvet :: Wed, 09/21/2016 - 1:11pm

Paul Allen's call of the Nate Poole TD never gets old in Packerland.

39
by Red :: Thu, 09/22/2016 - 11:56am

Ramirez / WR / Will: The official world record holder for starting Manning / Brady arguments in threads that have nothing to do with either. When someone goes out of their way to convince others how right they are about a given subject matter, it means that person is insecure about their own belief, and is really tryimg to convince himself more than anything.

42
by nat :: Thu, 09/22/2016 - 4:15pm

Scott was lamenting the lack of a study of dome/roof effects on passing that accounts for indoor teams possibly being just weak on defense for reasons unrelated to weather.

Here we go:

I use completion percentage, because it's an easy stat to get, and correlates very well with VOA. It's a good (not perfect) proxy for passing effectiveness.

All stats are from PFR's database for away games in the period 1994-2016. The sample sizes are very large, so I think we can trust all of these as not being flukes.
Outdoor - 58.5%
Indoors - 61.2%
diff ---- 2.7%

As Scoot would note, that could be because of factors other than weather, such as indoor teams just having sucky defenses. So let's look at September, when indoor and outdoor conditions are most similar, and then at December, when the benefits of playing indoors would be at their highest point of the regular season. If Scott is right about playing indoors not inflating passing stats, these should look the same.

Sep+Outdoor - 59.4%
Sep+Indoors - 60.8%
diff -------- 1.4%

Dec+Outdoor - 57.7% (down 1.7%)
Dec+Indoors - 61.0% (+0.2%, essentially unchanged, as we would expect)
diff -------- 3.3% (up 1.9%)

So, how much could be weather effects? If we assume no weather effects in September, then we get 1.7% to 1.9% for the benefit in December of playing indoors.

Now it's generally thought (known, really, regardless of Scott's protestations) that wind is an important factor in the passing game, not just cold and rain and snow. So it's not really a given that September has zero weather effects. Some of that 1.4% difference in September could be weather effects, too.

That means that the regular season benefit of playing home games indoors should range from 1.3% to 2.7% in completion percentage, depending on what you think of September weather effects. It's dependent on the time of year. It would be an even greater effect in January. But that's a different study.

There's more detail that could be done. We could also look at pass counts month to month indoors and outdoors, which could affect cumulative stats like DYAR. [EDIT: it looks like teams don't change their play mix indoors to outdoors differently between September and December.] But this is one study that accounts for the defenses of indoor teams possibly being different.

BTW: Football Freakonomics has done several simple studies of this, too. They see similar drops in completion percentage from month to month for outdoors games. This isn't really ground-breaking.

Weather matters. That means playing indoors matters. Playing indoors inflates your passing stats. Get over it, guys.

43
by PatsFan :: Thu, 09/22/2016 - 4:21pm

Damn your inconvenient facts!