Much has been said about the Seattle Seahawks last offensive play of Super Bowl XLIX. If you need a reminder or are a masochistic Seahawks fan here it is. Since the Super Bowl there's been tons of articles breaking down why Pete Carroll should have called a run play there. However, given the clock situation and the probability of a touchdown on a pass/run play suggests that Pete Carroll's decision to pass the ball is not just defensible, but the outright correct call.
Before I start spitting out numbers a couple assumptions need to be made:
|
Empirical data
I only went back to 2012 partially because I felt the amount of samples I had (324 pass plays, 690 run plays) was sufficient and partially because the way I was scraping data was very manual. As I (hopefully) get better at extracting data I'm planning on having more intensive analytics posts. I also decided to look at the entire NFL as teams generally perform equally well in these situations, and because of the small sample size of the Seahawks in these situations.
Anyways, the takeaway here is that difference between a turnover on a run play vs a pass play is very minimal. Pete Carroll was not significantly increasing the risk of a turnover when he decided to pass. Using these probabilities I determined the likelihood of failure of the 7 different play call scenarios Pete Carroll had for 2nd-4th down.
For example, lets say Pete decided to call pass, run, run on 2nd-4th down. The scenarios in which that fails are:
The last scenario is considered a failure after 3rd down because of the when 2nd down starts (0:26) & the assumption that the time in between plays is 15 seconds.
Using those assumptions the 7 scenarios are:
Anyways, the takeaway here is that difference between a turnover on a run play vs a pass play is very minimal. Pete Carroll was not significantly increasing the risk of a turnover when he decided to pass. Using these probabilities I determined the likelihood of failure of the 7 different play call scenarios Pete Carroll had for 2nd-4th down.
For example, lets say Pete decided to call pass, run, run on 2nd-4th down. The scenarios in which that fails are:
- 2nd Down: Interception
- 2nd Down: Incomplete, 3rd Down: Fumble
- 2nd Down: Incomplete, 3rd Down: No Gain/Negative Gain, 4th Down: No Gain/Negative Gain/Fumble
- 2nd Down: Pass for No Gain, 3rd Down: No Gain/Negative Gain/Fumble
The last scenario is considered a failure after 3rd down because of the when 2nd down starts (0:26) & the assumption that the time in between plays is 15 seconds.
Using those assumptions the 7 scenarios are:
From here on out the analysis is more qualitative, I don't have any other empirical data to add. Before we move forward with the qualitative analysis its important to understand the assumptions this model this makes
- Given this model uses data from all NFL teams, it does not take into account how good Seattle's goal line pass/rush offense and New England's goal line pass/rush are. Given the capabilities of each team I would wager that this assumption favors passing on 2nd down
- This model does not distinguish between types of pass plays. Slants, like the one thrown in the Super Bowl, likely have a higher interception and completion %, while a fade route's would be much lower. Its difficult to determine what this assumption favors.
- This model assumes that a negative gain (both pass and rush) have no negative effect on success percentages I laid out earlier. Obviously losing a couple yards would lower a team's TD%, however given the complexity of modeling how many yards a given play would lose I assumed negative gains were the same as no gain. Since rush plays are significantly more likely (10 vs 5%) to result in negative gains, this assumptions running on 2nd down.
analysis
sLooking at the 7 scenarios we can safely eliminate Scenario 1 (PPP) & Scenario 5 (RR) because of their higher failure rates. Scenarios 2-4 are in support of passing on 2nd down so there's no analyze them further. Scenarios 6 & 7, running on 2nd down, certainly seem to have the numbers on their side, however ultimately are non-optimal choices because of 3 major non-empirical reasons.
- After the 1st down Lynch run the personnel that the Seahawks had on the field were: 3WR, 1TE, 1RB. Changing personnel would take too long or burn a valuable timeout, so the Seahawks are forced to run 2nd down with the players they have on the field. 3WR/1TE/1RB does not lend itself to running the ball, especially against a 8 defenders on the line.
- If the Seahawks ran on both 2nd and 3rd there would be no time to have a 4th down play (Scenario 5). Consequently the Seahawks are forced to run a pass play - a fact the Patriots undoubtedly know.
- If the run resulted in negative yardage that likely forces the Seahawks to run pass plays on 3rd and 4th. This predictability allows the Patriots to game plan better.
Closing thoughts
Based on the empirical evidence coupled with the situation at the time I would conclude that Pete Carroll's decision to pass the ball was the correct call and 99 out of a 100 times (or 88.32% to be more accurate), the decision to pass is the correct one in such a situation.
Bonus
The video shows exactly how perfectly the Patriots need to address the play in order to stop it.
If any one of those 5 things doesn't happen the play likely results in a touchdown and at worst is an incompletion. Even if you don't take stock in the empirical data above, hopefully this video sheds some light on why Pete Carroll's decision wasn't a bad one.
- At the 0:04 mark Malcolm Butler recognizes the formation and lets Browner know that its a pick play designed to take them both out of the play.
- At the 0:10 mark Lockette fakes outside but Butler reads the direction of Jermaine Kearse's pick and doesn't fall for the fake.
- Also the 0:10 mark Browner plays the pick perfectly and jams Kearse effectively stopping the Seahawks ability to pick Butler.
- At the 0:11 mark Butler makes a hard jump on the route.
- At the 0:12 mark Butler actually beats Lockette to the spot of the throw and gets the pick.
If any one of those 5 things doesn't happen the play likely results in a touchdown and at worst is an incompletion. Even if you don't take stock in the empirical data above, hopefully this video sheds some light on why Pete Carroll's decision wasn't a bad one.