Chip Kelly Should Have Gone For it on 4th Down, History Was On His Side

Last night, Chip Kelly opted for a field goal attempt of 44 yards, trailing 26-24 with 2:32 remaining and both teams having all their timeouts remaining. The Eagles had just faced a 3rd and 1, where Ryan Mathews was stuffed for no gain, setting up the crucial decision.

I would have gone for it, not a second-guess based on the result, but because a successful conversion would have moved Philadelphia closer, allowed them to start draining Atlanta’s timeouts, and maybe run out the clock or score with little remaining, rather than leave time for Atlanta to have its whole offense available if trailing by 1. A failure would have left them in a similar position where they ended up: Atlanta playing more conservatively, and the Eagles having all timeouts left.

But this post is not about that, but a tangential point, based on something Kelly said in his post-game comments about his decision.

“On 3rd-and-1, we didn’t get a lot of movement up front. And it was a 43-yarder,” Kelly said, according to CSN Philly. “We have great confidence in Cody. It’s tough. He’s hit those before.”

Every now and again, I get to go searching for something I wrote almost a decade ago on a bland and uninspiring blog that nobody probably read ever. Once upon a time, I wrote “a yard is a yard”, talking about how it would be illogical to base a decision only on the result of the previous play, and how the evidence showed that teams that went for it after getting “stuffed” were more successful on fourth down.

But that was one season back in 2006. So using the wonderful play finder at pro-football-reference.com, I examined all 3rd and 1 runs between the opponent’s 11 yard line and 40 yard line that were stopped for no gain, leaving a 4th and 1, and then looked at the result of the next play.

There were 132 plays that ultimately met the criteria. Of those, 92 times the team went for the conversion, 34 times they kicked a field goal, and 6 times they punted.

Six of those plays involved penalties: one offsides giving a first down, and 5 offensive penalties nullifying what would have been a first down (or ending it pre-snap).

Of the remaining 86 attempts, 62.8% were converted for a first down. For all other first down attempts in this range over the previous five years, the conversion rate was 61.3%.

Now, some of those plays were the runs were stuffed on 3rd and 1 might have gained a foot, or a few inches, making it a shorter conversion. Still, there is not evidence that teams that go for it after getting stuffed are converting less than otherwise. It’s almost as if the effect of one play is far less important than overall likelihood and broader factors.

Coaches have to fight biases all the time. In this case, a decision has to be made very quickly. Kelly doesn’t have time to sit and ponder, “well, we have been pretty good running the ball and the Falcons aren’t the Rams on the defensive front.” Recency bias can certainly come into play.

The Eagles, by the way, faced only one 3rd and 1 “stuffed” situation last year. That was in the December loss to Washington, where Kelly also opted for a long field goal from Parkey. He missed that one, and the Eagles ended up ultimately losing by a field goal.

Perhaps Cody Parkey’s words can be applied more broadly to a decision like this: “There’s no rhyme or reason to missing. I’m human and I missed. Don’t know what to tell you.” Individual results, as they say, may vary. And getting stuffed on one play is not necessarily predictive of how the next will go, just like missing that one field goal probably doesn’t change Parkey’s career outlook much.