In recent years, Rutgers statistician Harold Sackrowitz has continually pushed for a more enlightened approach to the two-point conversion. In the 2000 paper “Refining the Point(s)-After-Touchdown Decision,” he comes to the same conclusion as Porter: If the success rate when going for two is 38.2 percent or greater, you’ve gotta do it. Statistician David Annis of the site SportsQuant says the same thing. So does J. Denbigh Starkey, who assumes two-pointers succeed a mere 43 percent of the time (a dated estimate circa 2004) and that kickers make just 94 percent of extra-point attempts. The outcome using those figures: Going for two gives you a winning percentage of 54.0 compared to 45.5 for kicking the extra point. (If you’re wondering about college football, the 2012 NCAA record book reveals that college teams made 41 percent of two-pointers and 95 percent of PATs between 2000 and 2011. That means going for two makes sense for college teams, but it’s a much closer call.)
Despite the unanimity of the quants, NFL coaches are unmoved. According to the fantastically useful Pro Football Reference Play Finder, there were 12 separate NFL games this season (including Cowboys-Saints) in which a team scored a touchdown with less than six minutes on the clock to reduce its deficit to eight points. None of those 12 teams dared to go for two.
In fact, if the Play Finder is correct, there has been just one instance since 2000 in which an NFL team took the mathematically correct, go-for-two-when-down-14 approach. On Oct. 14, 2001, the Baltimore Ravens trailed the Green Bay Packers 31-10 late in the fourth quarter. The Ravens scored a touchdown with 2:56 left to close to 31-17, recovered an onside kick, then scored again with 38 seconds to go. This time, Baltimore coach Brian Billick eschewed the extra point. Randall Cunningham’s two-point pass fell incomplete, though, and the Ravens’ follow-up onside kick failed. The Packers held on to win 31-23.
In a Baltimore Sun story a few days later, Billick explained that he wanted to “let the next score give you a chance.” The implication: He made the unconventional go-for-two move because he believed his exhausted team would be worse than an even bet in overtime. In his 2000 paper, Sackrowitz cites a similar case from 1999 in which Detroit Lions coach Bobby Ross went for two with his team trailing 23-19 in the back half of the fourth quarter. “We didn't want to go into overtime—not with three games in 11 days and as banged-up on special teams as we were,” Ross explained. As in the Ravens-Packers game, this shows that a coach’s end-game calculus may change if he doesn’t believe (as we’ve been assuming) that overtime will be a 50-50 proposition.
After that game in 2001, Billick didn’t get badgered by the press about his unsuccessful two-point maneuver. That’s most likely because the Ravens didn’t recover that final onside kick, making the whole thing moot. Ross, by contrast, was attacked mercilessly after the Lions missed their two-point conversion, then—down four instead of down three—were unable to kick a tying field goal on a subsequent, last-gasp drive. Even the neutral-to-a-fault AP got on Ross’ case: “Next time, maybe Detroit Lion Coach Bobby Ross should go for the sure one, not the risky two.” (Detroit Mayor Dennis Archer did have the coach’s back. "One thing I want you all to appreciate: risk-takers, people who make a difference," Archer said, adding that “you don't always win when you take risks; you get a lot of pats on the back when you're successful.”)
The cases of Billick and Ross aren’t exactly parallel—taking the tying field goal out of play makes the down-by-four two-pointer the significantly worse percentage move. According to Football Commentary’s two-point conversion chart, it makes sense to go for two when trailing by four with six minutes left if you’re likely to make the two-pointer more than 59 percent of the time. (That’s a general figure that doesn’t take into account the strengths of one’s field-goal kicker and various other factors.)
Even so, the huge disparity in the backlash to Billick and Ross reflects that football fans and football coaches confuse outcome with process. The PAT-kicking Cowboys should not be chided because they lost to the Saints in overtime. Rather, they deserve criticism for deploying a strategy that reduced their chances of winning. So, let me also censure Gary Kubiak and the Houston Texans for their behavior against the Jacksonville Jaguars on Nov. 18. In that game, the Texans pulled within 34-26 with 5:39 to go, kicked the extra point, tied the game on another TD and extra point with 1:34 remaining, then won the game in overtime. The Texans got the right outcome from the wrong process. Shame on you, Houston.
If going for two when down eight is a winning strategy, why doesn’t anyone do it? Because—just as in the case of punting on a short fourth down—kicking the PAT minimizes the chance you’ll look dumb. As SportsQuant’s David Annis explains, the extra-point strategy decreases your chances of losing in regulation while increasing your overall chance of losing the game. As a consequence, Annis writes, it seems the typical football coach “really isn't trying to maximize his chances of winning. Rather he's attempting to prolong his defeat.”
My correspondent Walter Sun suggests there’s a simple way to get risk-averse, play-not-to-lose coaches to come around to the right way of thinking. “If you go for two when down eight, you not only improve your win probability (the sound mathematical reason), you also control your own fate (the reason fans/management will accept),” Sun says. Even in the age of the NFL’s new, fairer overtime system, the team that wins the OT coin toss still holds the advantage. “In a sport where the phrase ‘controlling your own destiny’ is overused,” Sun continues, this is the one time the cliché makes sense. Come on, NFL coaches—do you really want your fate to depend on whether a coin comes up heads?