In any loss, there are plenty of "what if?" moments that allow for the Monday second-guessing to which we've all become so accustomed. What if Dez Bryant didn't drop the two-point conversion? What if Torrey Smith had gotten flagged for pass interference before halftime instead of Morris Claiborne?
For me, the biggest "what if?" came just prior to the start of the fourth quarter with Dallas down 24-20. On a third-and-5 at Baltimore's 35-yard line, the Cowboys lined up in "Ace" formation. The Ravens blitzed and Tony Romo immediately went deep to Miles Austin down the sideline on a fade. An incompletion set up fourth-and-5.
From the 35-yard line, a field goal attempt would have been around 53 yards – not outside Dan Bailey's range but certainly not money in the bank. Nonetheless, I figured Jason Garrett had two decisions: attempt the long field goal or go for a first down. He bypassed both options, instead instructing Romo to take a delay of game penalty to give the team more room to punt.
To determine which call was correct, we can examine how many points teams have typically scored with the ball in various situations. That is, how many points could the Cowboys be expected to score if they converted a first down, kicked a field goal, or punted?
Let's work in reverse order. Following their punt, the Cowboys left the Ravens with a first-and-10 at their own 8-yard line. Over the past decade, NFL teams starting at or near their own 8-yard line have scored 1.15 points per drive. Thus, the expected points for the Cowboys following their punt was -1.15.
Had the Cowboys attempted the field goal, there's a solid chance Bailey would have connected (despite the length). His last-second field goal miss aside, Bailey and other NFL kickers have become extremely accurate all over the field, hitting on around 55 percent of field goal tries just over 50 yards over the past decade. Had Bailey missed, however, it would have provided Baltimore with a first-and-10 at its own 43-yard line, a situation in which NFL teams have averaged 2.09 points per drive over the past decade.
Thus, simple math shows us the overall expected points for a field goal try from the 35-yard line are 3 (0.55) - 2.09 (0.45) = 0.71. That is, even after adjusting for the chances of a miss and the good field position it would provide the Ravens, a field goal attempt was "worth" 0.71 points to the Cowboys.
Finally, what if Garrett had kept his offense on the field? Well, had Romo and Co. acquired exactly five yards for a first down, they would have been left with a first-and-10 at Baltimore's 30-yard line, a situation in which NFL teams have averaged 3.32 points per drive. But could we expect the Cowboys to convert on that play?
Since Garrett began calling plays in Dallas and I began tracking them, the Cowboys have converted 47.1 percent of their plays on either third-and-5 or fourth-and-5. The league average during that same period is 49 percent. With Dallas moving the ball against Baltimore effectively all day, we'll label their chances of converting as an even coin flip at 50 percent. Had the 'Boys failed on their fourth down attempt, they would have left the Ravens with a first-and-10 at their own 35-yard line, a starting point that results in an average of 1.76 points per drive. We can again easily deduce the overall expected points of going for a first down with some math: 3.32 (0.5) - 1.76 (0.5) = 0.78.
Ultimately, the correct decision for the Cowboys with their fourth-and-5 at the Ravens' 35-yard line would have been to go for it. They could have been expected to "gain" 0.78 points, on average, had Garrett kept the offense on the field, compared to 0.71 for a field goal try. Meanwhile, Dallas handed over 1.15 expected points by punting, meaning the overall net loss from the punt was 1.93 expected points, nearly two points I think the team could have used later in the game. Going for a first down was the best option and kicking the field goal was a relatively close second, but punting really shouldn't have been on the mind of Garrett, especially down four heading into the final quarter.
Of course, I think the entire situation could have been avoided by calling a different play on third-and-5. Statistically, rushing the ball on third-and-5 leads to around the same percentage of first downs as passes, primarily because defenses don't anticipate it. By running the ball on third-and-5, the Cowboys would have surprised Baltimore's defense while still doing what they had done best all day. Plus, a run that didn't go for a first down could have still put the Cowboys in a better position on fourth down, increasing the worth of either going for it or kicking the field goal.
