Over the years, I've tended to side with statistical analysts who believe NFL coaches are too skittish about going for it on fourth down. More often than you might think, data built up from years of NFL play runs counter to conventional wisdom about the right time to punt, kick a field goal and/or "trust your defense."
Once such opportunity seemed to arise Sunday for the Minnesota Vikings late in their game at the Chicago Bears. Leading 27-24 with three minutes, 18 seconds remaining, the Vikings had a fourth-and-goal from the Bears' 4-yard line.
Coach Leslie Frazier faced a dilemma. He could kick the short field goal, take a six-point lead and leave his team vulnerable to a regulation loss if the Bears answered with a touchdown. Or he could go for a touchdown that would force the Bears to score twice to win.
In the worst-case scenario, if the touchdown attempt failed, the Bears would have gained possession pinned deep in their own territory with at least 60 yards to gain -- and only one timeout available -- for a chance at a game-tying field goal.
We all know what happened: Frazier chose Blair Walsh's 22-yard field goal. The Vikings took a 30-24 lead, but the Bears scored a touchdown with 10 seconds remaining for a 31-30 victory.
Watching the game live, I thought Frazier should have handed the ball to Adrian Peterson and tried to score a touchdown to seal the game. I figured the Vikings' chances to seal the game with the reigning MVP were at least as good, if not better, than the likelihood that their defense would stop Bears quarterback Jay Cutler in a fourth-quarter frenzy.
Years of hard data, however, mutes that sentiment. Frazier made the most informed choice, either intentionally or by chance. ESPN analytics specialist Alok Pattani graciously researched and passed along the details.
The ESPN win probability model contains data from all possible down-and-distance situations dating back to 2001. Over that period, NFL teams have scored touchdowns on 27 percent of their attempts on fourth-and-goal between the three- and five-yard lines. Perhaps the Vikings' chances were better than that with Peterson, but 12 years of precedent is hard to ignore.
Four yards takes more than an offensive line push and a lunge from the running back. You need either a well-blocked hole or an exceptional play by a running back who had averaged less than four yards per carry in the game.
As a result, statistical analysis gave the Vikings a better chance to win by converting the field goal (81.5 percent) and defending against a touchdown drive than going for it on fourth-and-goal (74 percent chance) and facing the likelihood of failure.
Frazier is as traditional of a game manager as there is, and I'm not sure he would have gone for it even if the ball at been at the 1-yard line, but this exercise emphasizes that knee-jerk reactions in the balance between convention and statistical analysis can work both ways.