Should the Patriots have intentionally let the Seattle Seahawks score with a minute remaining in Super Bowl XLIX?
As I’m sure is the case for a lot of New England Patriots fans, the last minute of Super Bowl XLIX was one of the most heart-stopping things I’ve ever seen in sports. Jermaine Kearse’s circus catch looked to start a Seahawks dynasty, but Malcolm Butler’s interception restarted the Patriots’ version instead.
However, in the short time between those plays, the game seemed finished. The Seattle Seahawks, featuring one of the most powerful backs in the game, Marshawn Lynch, had the ball on the one-yard line with 57 seconds remaining. Out of the 100 million+ viewers, not one of them thought that Seattle coach Pete Carroll would do anything but punch Lynch down the middle for a touchdown.
I don’t have to remind anyone of the decision Carroll made, but it made Malcolm Butler a superstar and gave Tom Brady his fourth ring.
Here’s what I’ve always found curious about the end of Super Bowl XLIX: the Butler interception, as spectacular as it was, was basically a one-in-a-million play. What happens in the other 999,999 universes where Pete Carroll hands the ball off to Lynch, with three plays of wiggle room?
I think most fans would agree that they were almost guaranteed the win. Three attempts for Lynch to gain one yard? Almost 100%. This left a theory burning in my mind: what if Bill Belichick called for the New England defense to stand back and let Lynch walk into the end zone, Ahmad Bradshaw-style?
This move is incredibly risky, borderline unsportsmanlike, and flat-out dumb. That said, it rests on a technically solid thesis statement: for this to make sense, Bill Belichick would be betting that it is MORE likely for Tom Brady and the Patriots to be able to score 3+ points in one minute than it is for the defense to stop Marshawn Lynch from gaining one yard in one minute.
If New England lets Lynch score, the ball gets kicked back to New England with about a minute left to either tie the game with a field goal or score a miracle touchdown. If New England tries to stop Lynch (in this case, we assume the attempt results in at least one stop), Seattle could let more time waste, restricting a Brady drive much further.
Fortunately, thanks to a helpful Win Probability Calculator at pro-football-reference.com, we can actually test the likelihood of victory in these scenarios.
The artificial intelligence weighs in for the Patriots.
The Win Probability Calculator lets you input the score, Vegas odds (which were actually dead even), time remaining, field position, down, and yards to go. There are a couple shortcomings, though, that we’ll have to use our imaginations to overcome. For one, the algorithm doesn’t take timeouts into account, as the machine cannot simulate clock-stopping strategy. Perhaps it was programmed by Andy Reid.
The Calculator also can’t simulate the impact of individual players on the field, which are essential to this scenario. On one end, you have the greatest quarterback of all time running 2014’s highest-scoring offense. On the other end, you have what was likely the most effective running back for a one-yard punch in the league.
We crunched the numbers for what would be generically evenly-matched teams. The results will still be revealing, and we can provide a little window dressing by picturing the relevant players in this position. So, what is the win probability of a team down by four, with a minute to go, on second down on the one-yard line?
65.7%, which means New England’s odds of winning would be 34.3%. I was surprised at how relatively low this number was, but there’s more to this setup than three HB Draws. Sure, Marshawn “Beast Mode” Lynch seems like a guaranteed touchdown, but there are also variables such as fumbles or penalties that skew it lower.
Or, per se, a hypothetical Russell Wilson pass to a Patriot.
Now, what if the Patriots let Lynch score a touchdown in order to put the ball back in Tom Brady’s hands? Could it really be more likely for a team to run downfield in a minute than to stuff a running back three straight times? What’s the win probability of a team down by three, with a minute to go, on first down on their own 20 yard line?
13.65%. Hmmmmmm. I’ll be the first to say that a 21% drop in probability isn’t worth surrendering to an opponent’s offense in the Super Bowl. Even if we’re very generous and we assume the Seattle kicker shanks the ball out of bounds on the kickoff and placed it on the 40, we only rise to 17.70%.
I don’t have the capability to tell the AI that the running back is Marshawn Lynch and the quarterback is Tom Brady, but even with tweaking, I don’t think I could get the win probability of the New England offense to match the win probability of their defense.
We’ve reached a conclusion. It’s not that shocking, but in case there was any doubt, the best way to win a football game is by trying to prevent the other team from scoring.
The second-best coaching decision of the night belongs to Bill Belichick, for trying his best to stop what looked like an imminent touchdown in a handoff to Marshawn Lynch. The best coaching decision of the night? Pete Carroll telling Russell Wilson to throw it.