<
>

Inside Slant: Innovating after touchdowns

(Another in an Inside Slant series that will appear regularly during the 2014 offseason.)

By now you're aware that the NFL has taken the dramatic -- for this league, anyway -- step of initiating a two-week preseason experiment with extra points. During the first two weeks of the 2014 preseason, teams will line up at the 20-yard line for extra points -- turning what had been a near-automatic play into something slightly more difficult.

The experiment is meant to identify any unintended consequences of adding at least some drama and strategy to what is otherwise time for viewers to check their fantasy results. Even if all goes well, it probably will take years of discussion before 24 of the NFL's 32 owners approve a permanent change. But for now I think it's worth evaluating just how much a 38-yard extra point -- or thereabouts -- would change the game.

Namely: Might we reach a day when it makes more sense to go for two points (at the orignal spot on the 2-yard line) than an extra point?

I enlisted the services of Alok Pattani, a senior ESPN analytics specialist, to walk me through the scenarios. My conclusion based on Alok's advice: During the part of games when maximizing points is the prime objective -- basically the first three quarters -- a credible argument could be made to choose a two-point conversion over a 38-yard extra point. The fourth quarter, however, requires a shift into maximizing the chance of winning -- which in some cases means an extra point would be the best option.

Here's how we looked at it, using statistics from the past two seasons given the recent and sharp rise of kicking accuracy:

During the past two seasons, the average conversion rate of a 38-yard kick is 89 percent. In that same span, teams have been successful on 49 percent of their two-point attempts.

Traditional thinking -- and I admit it is hard to get away from this approach -- suggests teams should take the one point they would get on nearly nine out of every 10 tries rather than the two points they would get on less half of those attempts.

But basic analytics requires us to consider the tradeoff in a different way: expected value over time, which equalizes the comparison. Getting one point on 89 percent of attempts gives us an expected value of 0.89. Two points on 49 percent of attempts is an expected value of 0.98.

In other words, you would score more points over a given time period by going for two every time -- even accounting for the higher degree of difficulty and likely failure -- than if you only kicked PATs.

That's a difficult point to absorb for traditionalists -- be it coaches, fans or media members -- especially if a failed isolated instance proves the difference in a defeat. The key to this kind of thinking, as cutting edge Pulaski (Ark.) Academy coach Kevin Kelley said at last month's MIT/Sloan Sports Analytics Conference, is understanding that even the better bet doesn't work every time.

Ultimately, I doubt that the difference in expected value -- about five points on 50 touchdowns scored -- would be enough to convince many coaches to change their fundamental philosophy. Based on current success rates, at least, I imagine the NFL would have to set extra points at around the 46-yard line -- where the success rate the past two years drops to 75 percent -- to prompt serious consideration for an emphasis on two-point conversions.

There are some other factors to consider, of course. Bad weather might push some coaches away from kicking a 38-yard extra point. And in the fourth quarter, everything shifts into maximizing point differential. It makes no sense to go for two, for instance, if an extra point would give you a two-score lead with five minutes remaining or if the score is tied with one minute remaining.

I think the NFL is on the right track in providing a new template for strategy decisions, at least for those coaches who would appreciate it. My guess is that a 38-yard extra point will still be too attractive for most coaches to diverge from, but it's a start.