Share this article

print logo

Fourth-down analysis: It’s hard to argue with Bills’ coaching choices

There is a revolution coming in professional football, even if it’s moving at the speed of a glacier. There will be a day when NFL teams recognize their fourth-down decisions must be better. In short, they need to go for it more.

Academic research articles have been written about it, including one in 2006 that appeared in the Journal of Political Economy that stated that “teams’ choices on fourth downs depart in a way that is systematic and overwhelmingly statistically significant from the choices that would maximize their chances of winning.”

Documentaries have been made on those who go against the grain, like Arkansas high school coach Kevin Kelley. A folk hero among those in the sports analytics field, Kelley’s teams at Pulaski Academy in Little Rock, Ark., have won four state championships since he took over in 2003. Kelly almost never punts and almost always onside kicks after scores. He’s compiled a record of 138-25-1 heading into the 2015 season.

Publications like the New York Times have blogs devoted entirely to the subject, with every fourth-down decision analyzed by the “NYT 4th Down Bot.” The Twitter feed @NYT4thDownBot tweets out reactions to fourth-down decisions made by NFL teams in real time.

The website Advanced Football Analytics offers users a fourth-down calculator into which they can plug the score differential of the game, time remaining in a particular quarter, field position and yards to go for a first down, and using historical win and point expectancy models, get a comparison of the total expected value of each alternative.

Last season, the Bills’ fourth-down decisions under former coach Doug Marrone were heavily scrutinized. This year, The Buffalo News is planning to track every fourth-down call made by coach Rex Ryan and his staff.

We’ll lay out the situation, what the Advanced Football Analytics’ fourth-down calculator suggests, what the Times would do, and our own opinion of what the right call would have been. That will be formed by taking into account variables that the math doesn’t always account for – things like weather, quality of opponent and the ever-so-tough to define “momentum” during a game.

Against the Colts on Sunday, the Bills faced fourth down seven times. Here is a recap of each one:

1. Fourth and 4 from the Buffalo 47-yard line, 0-0 score, 8:08 remaining in the first quarter.

Bills’ decision: Punt.

What the math says: A success rate of 53 percent produces a total win probability of 50 percent when going for it (57 percent win probability with a successful conversion minus 43 percent with a failure), versus a 49 percent win probability when punting.

What the Times suggests: Go for it.

Our call: Given that quarterback Tyrod Taylor was making his first career start, it’s understandable that Ryan wouldn’t want to saddle him with trying to move the chains on a fourth down near midfield so early in the game. The Bills hadn’t run the ball well on their first five plays, losing a total of 6 yards. It was a coin flip, but punting was defensible.

2. Fourth and 8 at the Buffalo 49-yard line, 0-0, 4:13 remaining in the first quarter.

Bills’ decision: Punt.

What the math says: A 38-percent success rate produces a total win probability of 49 percent (58 percent win probability with a successful conversion and 43 percent with a failure), versus a 49 percent win probability when punting.

What the Times suggests: Punt.

Our call: Bills punter Colton Schmidt did a nice job pinning the Colts deep on his first punt, which was a fair catch at the Indianapolis 9-yard line. The Bills’ defense had also forced a three-and-out on the Colts’ previous possession. Punting was a good call.

3. Fourth and 6 at the Indianapolis 23-yard line, 7-0 Bills lead, 7:53 remaining in the second quarter.

Bills’ decision: Field-goal attempt (good).

What the math says: Going for it is successful 50 percent of the time, while field goals are successful 76 percent of the time. A successful conversion equals an 85-percent win probability, while a failure drops that to 72 percent, for a total of 78 percent. A successful field goal means an 81 percent win probability; while a missed field goal is a 71-percent win probability, for a total win probability of 79 percent.

What the Times suggests: Field goal.

Our call: A 41-yard field goal was no sure thing after the preseason kicker Dan Carpenter had, but the idea of taking a two-possession lead at that point in the game made sense. Carpenter rewarded Ryan’s faith in him. Another good call by the Bills’ coach.

4. Fourth and 17 from the Buffalo 13-yard line, 24-8 Bills lead, 13:33 remaining in fourth quarter.

What the math says: A success rate of only 10 percent. Total win probability of 92 percent when going for it and 96 percent when punting.

What the Times suggests: Punt.

Our call: An easy call here to punt it away.

5. Fourth and 11 from the Buffalo 28 yard-line, 24-8 Bills lead, 9:56 remaining in fourth quarter.

Bills’ decision: Punt.

What the math says: A total win probability of 97 percent either way.

What the Times suggests: Punt.

Our call: Another easy call. Punting here is logical. As an added bonus, Colts punt returner Phillip Dorsett muffed it, and the Bills recovered.

6. Fourth and 11 from the Indianapolis 27-yard line, 24-8 Bills lead, 8:18 remaining in fourth quarter.

Bills’ decision: Field goal attempt (good).

What the math says: Field goals are successful 68 percent of the time, while conversions are successful just 29 percent of the time. Total win probability is 98 percent when going for it and 99 percent when kicking the field goal.

What the Times suggests: Field goal.

Our call: Trying for the points was right on the money. Carpenter’s 45-yarder gave the Bills a three-possession lead.

7. Fourth and 5 from the Buffalo 47-yard line, 27-14 Bills lead, 2:43 remaining in fourth quarter.

Bills’ decision: Punt.

What the math says: Punting produces a 98 percent total win probability, while going for it leads to a 94-percent win probability.

What the Times suggests: Punt.

Our call: Schmitt had been pinning them deep all game, so punting it away made sense. The Colts would have had to have gone 80 yards for a touchdown and then recovered an onside kick to have a chance at winning.

email: jskurski@buffnews.com

There are no comments - be the first to comment