Super No-Brainer

February 2, 2015

For baseball fans, the end of the SuperBowl is generally viewed as the official beginning of the baseball season. But I’d like to take an extra moment or two to mull over what was an extraordinary Superbowl. This is a little off my beaten path but not completely out of my depth as I’ve written about football before. The game certainly had its fair share of exciting and memorable moments. Had it not been for Seattle’s final offensive play, Jermaine Kearse and/or Chris Matthews might have entered the pantheon of unlikely postseason heroes. But their efforts will forever be overshadowed by the play that will haunt Seattle and their fans for a long time: the decision to throw the ball on second down and goal with only a half a yard separating them from victory. The decision has been widely criticized but there has sprung up a kind of cottage industry to defend head coach Pete Carroll’s and offensive coordinator Darrell Bevill’s decision to pass. So was the play call justified? It’s worthy of a deeper look.

But before we get going in earnest, many of Carroll’s defenders are starting the conversation with his resume, which according to them includes one Superbowl win and two college national titles. For the record, Pete Carroll has not won two NCAA championships. In fact, he has not won any. The first he is being credited with was for the 2003 season, but that year USC didn’t even play for the championship; LSU and Oklahoma did. USC was voted by the writers as the nation’s top team, but the actual NCAA championship went to the winner of the BCS title game, which for the 2003 season was the Sugar Bowl winner, LSU. You can whine about the fairness of inviting Oklahoma to that game instead of USC but the numbers favored Oklahoma which is why the BCS ranked them higher. And since the BCS was created solely to obviate the writer’s vote due to the debacle of Brigham Young’s 1984 title, the writer’s vote was irrelevant as the final arbiter. The BCS ranking, which was the voice of the coaches as well, was the only opinion that mattered. The following year USC legitimately played in the national championship game and beat Oklahoma. But that win was later vacated when it was discovered that USC had committed serious recruiting violations. As far as the records are concerned there was no national champion that year. So no, Pete Carroll has not won two national championships. He was voted one by a group that didn’t matter, and his team scored the most points in a championship game in which he was essentially caught cheating.

OK, so now let’s get to the play in question. Should Seattle, as the dominant sentiment has suggested, have given the ball to Marshawn Lynch? Was that a no-brainer? Or was this simply a matter of result; had the pass worked would we be saying it was a great call? Let’s look at the components…

Seattle was the #1 rushing team in the NFL, carrying the ball for 172 yards per game, which was 26 more yards per game than the #2 team, the Dallas Cowboys. They averaged an impressive 5.3 yards per carry as a team and led the #2 team by .6 yards per attempt. The last team to lead the league in yards per attempt by a greater margin was the 1997 Detroit Lions with Barry Sanders rushing for more than 2000 yards by himself. Conversely, the Seahawks were the #27 ranked passing offense in terms of yardage and tied for #22 in passing touchdowns with teams like the Tennessee Titans, who won only two games. Even the three-win Oakland Raiders threw more touchdown passes than the Seahawks. That’s not to say that Seattle did not possess the ability to pass but it is abundantly clear that the passing game was not their emphasis or strength. They were efficient but far from prolific. More than 45% of their yardage came from running the ball.

Defensively, the Patriots were middle of the pack defending the pass (16th in yardage, tied for 12th in turnovers, 12th in scoring). Against the run, they were better: 8th in yardage, tied for 13th in turnovers, 2nd in scoring. So the Patriots were clearly better at defending the run, but they were not historically good. All things being equal, passing was slightly preferred as a way to attack them but it wasn’t a glaring weakness, and certainly not so much that a team that clearly favors the run should change their character for one game.

To the situation at hand, according to Harvard Sports Analysis Collective, the Patriots allowed opponents to score 81% of the time in power situations (runs on 3rd/4th down & less than 2 yards to go or within 2 yards of the goal line). That was the worst mark in NFL. On the offensive side, Seattle was second in the league in those same power situations, getting stopped just 17% of the time. Lynch himself converted 17 of 20 3rd/4ths & short this year. So the situation heavily favored a Seattle run. In fact, there wasn’t a better match-up to be had.

How about the actual players involved? As mentioned, Lynch was pretty strong in important short yardage situations. If the statistics are adjusted to context neutral, he was the #1 running back in the NFL last year when it came to yardage from rushing. In addition, he led all backs in touchdowns scored and lost only one fumble all year. In the game he had rushed 24 times for 102 yards. In 22 of those 24 attempts he had rushed for at least 1 yard and only five times had been slowed for less than 3. Basically he had a 91.6% chance of scoring on the next play if they handed him the ball, and Seattle still had two more downs to try again if he had been stopped. He was about as safe a bet to score a touchdown from a half yard away as there is in the NFL.

Seattle also had a quarterback who had rushed for 849 yards during the regular season (7.2 yards per carry), a total that tied for 16th best among all NFL players. To give an idea of how impressive a runner Russell Wilson is, in the history of the game only five other QBs have rushed for more than 800 yards in a season: Colin Kaepernick, Robert Griffin III, Michael Vick (twice), Randall Cunningham and Bobby Douglass. Wilson had already run for 39 yards on three carries during the game. He clearly had the wheels to cover half a yard, especially on a misdirection play or on a simple quarterback sneak.

So it’s acceptable to suggest that giving the ball to Marshawn Lynch wasn’t completely a no-brainer; they could have let Russell Wilson run it, too.

But that’s not what they chose.

Seattle decided it would be better to let Russell Wilson throw it. Unfortunately for the Seahawks, he’s not nearly as impressive a passer as he is a runner. His completion rate for the game to that point was under 60%. During the regular season his efficiency was 63.1% which ranked 23rd among QBs who had thrown at last 100 passes, behind such passers as Mark Sanchez and Jay Cutler, both of whom are better known for throwing interceptions than touchdowns. Not to say he was a bad QB; no, his QB rating was solid, ranking 11th overall and his yard per attempt ranked 12th. However, much of that was due to his ability throwing the long ball, where he ranked 7th overall in pass plays of greater than 20 yards. This discrepancy would then seem to indicate that short passes were not his forte. That theory is supported by the fact that Wilson ranked 19th in the number of passes for a first down, and 20th in percentage of his passes for first downs and 33rd in attempts per game. His value as a passer derives almost entirely from throwing a few well-placed long balls.

And he was just coming off a miraculous win for the Seahawks in which he threw 4 interceptions. That was by far the worst performance of his career, but it is something that Carroll and company should have kept in the back of their collective minds, that asking him to throw was not as safe an option as it might have been had Seattle had a different QB. And in general, throwing the ball over the middle into traffic was probably not a great idea either. It is the pass play with the greatest risk of bad things happening. A tipped ball likely results in an interception with so many people near the ball. If the receiver catches it, he’s open to especially jarring collisions that could cause a fumble with defenders flying from the opposite direction. And if he does not get into the end zone, precious seconds tick off the clock.

What about his intended target, Ricardo Lockette? He had caught only 11 passes during the regular season and was arguably their 5th best receiving target behind Doug Baldwin, Jermaine Kearse, Chris Matthews and Lynch. The latter three had been making big catches all game. One could also throw in tight end Luke Wilson as a better option since the Patriots were 30th best in the NFL at defending against tight ends but apparently Seattle forgot they could throw to the tight end because they did not attempt it even once during the game.

So Seattle’s brain trust opted to forsake the highest percentage chance to score the winning touchdown, as well as the variation on that option by running the quarterback, but instead chose let a quarterback who wasn’t even in the top half of the league in passing efficiency attempt a pass that he’s not particularly practiced or adept at completing, throwing to the team’s 5th or 6th best receiving option. Sure, that’s a surprising strategy for the opposition in much the same way that asking Bartolo Colon to hit and run in the ninth inning of Game 7 of the World Series would be. Or as one popular internet meme put it, telling your team to give Bill Cartwright the final shot of Game 7 when Michael Jordan and Scottie Pippen are on the team. Surprising, for sure. But smart? 

As for Carroll's excuse that the Patriots had stacked the box and that's the reason they decided to pass, that's pretty weak sauce. Over the last three years, Seattle made its reputation by rushing the ball, even against teams that had stacked their defenses against the run. They prided themselves on that ability even to the point of arrogance. But with the season on the line and a chance to win a second consecutive championship by going with their bread and butter, they suddenly decided to go in the opposite direction and become a West Coast offense like Bill Walsh's 49ers? Thanks, but no sale.

As for the cottage industry, the most well-written defense was probably offered here. First, there are several significantly flawed assumptions in this piece. For one, it assumes that the Seahawks are like other NFL teams. They clearly are not. The breakdown of how their offense succeeds reveals they are not a typical offensive team. How often has a team had such a dominating running game featuring the best running back in the league and one of the top rushing quarterbacks in history? In 2012, the Redskins featuring Alfred Morris and RGIII led the NFL in overall rushing yardage and finished 2nd in yards per carry. In 2004, the Michael Vick-led Atlanta Falcons led the NFL in overall rushing yards and yards per attempt but their top rusher, Warrick Dunn, rushed for roughly 500 yards fewer than the league’s top RB, Curtis Martin. And that’s it. One has to go back before the 1967 merger to find rushing RB/QB tandems as potent. And the last time a Superbowl contender depended on rushing the ball for as much as 45% of its yardage? The 1986 Chicago Bears. The 2008 New York Giants came close but basically one has to go back more than 30 years to find similar successful rush-dependent offenses like Seattle’s. So applying general assumptions about how the rest of the league currently prosecutes offense is a poor fit to model how Seattle should.

Secondly, the Patriots would not have had 20 seconds left as the writer asserts. Had Lynch run the ball and not scored, Seattle controlled when the clock would have been stopped. They could have run it down as far as they wanted and then called a timeout. Again, they had two opportunities to run the ball regardless if they needed a third play. But let’s assume for the sake of argument that they scored with Lynch on second down. First of all, running plays take more time off the clock than passing plays so more seconds would likely have ticked off. Additionally Lynch’s no-surrender style of running would have undoubtedly run off an extra second before the clock was stopped unless the Patriots completely blew their scheme and allowed him to coast in. But even giving in to the wildly optimistic assumption that he scored with 20 seconds left, the Seahawks would still have had to kick the extra point and kick the ball off back to the Patriots. Unless the Patriots downed the ball on the kick-off, a few more seconds would have run off. The Seahawks could have forced that issue with a squib kick down the field, ticking off a few more seconds. Depending on where the ball ended up being downed it’s not unreasonable to suggest that the Patriots could have been left with 15 seconds or less.

That would have been enough time to possibly run two plays and then attempt a field goal. The Patriots had two time-outs. If they had started their final drive from the 35 (which seems pretty optimistic) they would have had to cover 25 yards in two plays in order to attempt a 57-yard field goal, which would have broken the current Superbowl record for longest field goal by 3 yards. Brady had completed only 2 of 6 passes for 20 yards or more so they would have likely needed a couple of short-medium distance passes. That would have been an interesting challenge given that Julian Edelman, Brady’s top target for the game, was possibly out from a blow to the head suffered during the previous drive. It’s possible for a well-disciplined team like the Patriots to accomplish what was needed even without their top receiver, but even under the most ideal circumstances the writer acknowledges that even with 20 seconds left the Patriots would have had only had a 5% chance to attempt a tying kick. Not for a win. A tie. Given that Brady had already thrown 2 interceptions in the game, I don’t think it’s unreasonable to suggest that the Seahawk defense wasn’t facing Brady in his prime and that he was less than infallible, even in crunch time. So even under the most ideal circumstances, had Seattle scored without any clock manipulations the Patriots had less than a 5% chance to tie the game and force overtime.

Here is the most important point that needs to be made, a point which appears to be lost on the 538 writer: it does not matter when the winning points were scored, only that they get scored. It is impossible to win a game with fewer points than your opponent has. Far more often than not in professional sports, one does not get to choose when points will be scored. So when given an opportunity, you better score them. That Pete Carroll felt he could waste a play and still win reveals the magnitude of his hubris, for which he deserves this comeuppance. His team was competing against the only team left in the competition for the title of best team in the league, yet he believed his team could score whenever the whim suited him, despite more than 59 minutes of game play evidence to the contrary. Wasting a play should not have even been a consideration. Like the writer at 538, the Seahawks offensive brain trust simply over-thought this one. Sometimes there is a right answer and it’s the obvious one.

My own opinion is that had Seattle scored on the pass we might have termed it a gutty call (and an incredibly lucky result), but it was not a great call, like, say, an onside kick to start the second half. Effectively, this choice reduced Seattle’s chances for winning the game from 99% to 1% in one play (in the interest of precision, the actual win probability went from 85% - having the ball on the half yard line on second down with 26 seconds left down 4 points - to 0.2% - not having the ball and down by 4 points with 20 seconds left). Taking a risk that great when there is a less risky path that gets you to the same place is never a great call, regardless of the outcome. That is the real no-brainer.