Last Thursday against the defending NFC Champion Philadelphia Eagles, the Minnesota Vikings were able to claw back to within seven points on multiple occasions, before eventually losing 34-28.
On two occasions Minnesota, trailing by 13, scored a touchdown to pull within seven points and kicked the extra point to get to within six. Upon first blush, this makes sense. First, this is how it has been done forever, and there is at least some value in the incumbency of ideas. Secondly, with a stop and a touchdown followed by an extra point, the Vikings would have a lead and a potential victory as a six-point underdog.
However, in both cases it made more sense for the Vikings to attempt a two-point conversion. The math, as computed by Ben Baldwin’s model through nflfastR, confirms this.
Vikings win % when down 27-20 and going for 2: 11.9%
Vikings win % when down 27-20 and kicking the PAT: 10.1%
The numbers were 1.7% and 1.1% respectively when down 27-34. Seemingly small, these edges are not insignificant during a football game. This could be the end of the conversation, but football is much more than numbers.
Before we go further, this is a different situation than the “going for two down eight” idea, which has gained steam in the NFL – and was summarized initially and brilliantly by my former colleague, Kevin Cole, on PredictiveFootball (the link no longer works, but here’s a great piece by Seth Walder at ESPN).
Broadly, going for two down eight after a touchdown offers an increase of 12.5%-win probability conditioned on earning the touchdown, stop, and score needed to cover the two-touchdown deficit (assuming 50% conversion rate of 2-point conversions, 100% PAT conversion and 50%-win probability in overtime).
In the case of being down seven, the same argument does not really work since two consecutive touchdowns with PATs lead to a 100% chance of winning, while going for two on the first try leaves open the possibility of overtime and hence drops that theoretical win probability below 100% conditioned on getting the necessary stop.
However, as was very apparent in the Minnesota/Philadelphia game, the strategy hinges on getting the necessary stops, which is not trivial (also, extra points are not 100% propositions). Minnesota gave up a touchdown the first time they were down six, and the second time the Eagles were close to field goal range before getting stopped in the run game, false starting, and punting it back to the Vikings with just a few seconds to play. Furthermore, being scored against is not the only risk of not having the ball as a struggling run defense could surrender enough yards and time to render the game over if the opposing offense has enough of the field to work with.
To that point, let’s look at the possible outcomes:
Extra Point Succeeds
This is the normal course of events. The defense must get a stop and must do so relatively quickly. A deep kick leaves the whole field for the opposing offense to work and curbs some of the impulses to play conservatively as points are still a few first downs away.
An onside kick is a tough choice given the low rate of success, and the fact that after an unsuccessful one there is almost no room for error for the defense because a field goal puts the point differential at nine points and the game becomes immediately out of reach. Any benefits of the opposing team laying up for a field goal are captured in the ever-decreasing chance it misses.
The benefits are trivial: get a stop, score a subsequent touchdown with a point after, and you’re ahead. However, doing so too quickly opens the door for a backdoor field goal by your opponent to win by two points.
Two Point Conversion Succeeds
In the case the two-point conversion succeeds, there are some real, tangible benefits. First, a stop and a touchdown give you a one-point lead, from which you can attempt a two-point conversion to give you a three-point lead that cannot be beaten by a field goal at the buzzer. This is incredibly valuable.
Additionally, and more importantly, a five-point deficit means that a field goal puts you down eight points, meaning that an onside kick is less of a risky proposition. If you get the onside kick, you’re in the driver’s seat. If not, your opponent is close enough to field goal range that they may be enticed to be less aggressive, which leaves you in the game with a make or a miss.
Lastly, even if they do score on you, there is an outside chance that they fail to make the PAT or two-point conversion and you are down 11, which is tied via a field goal, touchdown, and two-point conversion.
Thus, you’re setting up a situation where you can win even if you have a weak defense that is not reliable enough to trust to kick the football deep and defend a long field (which will often end in you never seeing the ball again).
Two Point Conversion Fails
This is obviously not a great outcome, but in the case that you surrender a field goal on the subsequent drive, it is the same outcome as kicking the extra point (being down nine and being down 10, if you only have one possession left, are equivalent). Surrendering a touchdown is also roughly the same bad outcome for you. You’re likely going to be incentivized to kick deep more frequently if you miss the two-point attempt than if you made it, but no more so than if you kicked the PAT.
Granted, a stop and a touchdown doesn’t immediately beat your opponent in this case, and a PAT is not guaranteed to tie it, but you are open to attempt another two-point conversion for a chance to win, so that door is not completely closed, either.
Going for two when down seven (after scoring a touchdown down 13) is probably a sharp move in most circumstances. The upside is that you have many ways to win even if things don’t go completely perfectly, including in situations where you are unable to stop your opponent from scoring. In a league where, for many defenses, a field goal is a stop, this should be considered part of every NFL team’s end-of-game decision-making process.