David Romer

David Romer is a Berkeley economist whose study of fourth-down strategy challenged professional football’s customary preference for kicking. Using data from 700 NFL games and dynamic programming, he compared the expected value of alternative plays across field positions, arguing that teams often sacrificed scoring opportunities by choosing the conventional, cautious option. His work made football a clear example of a broader problem: experienced decision-makers can persist with established practice even when systematic analysis points elsewhere.

Research Question and Method

Romer set out to answer two related questions: when a team should try for a first down rather than punt or attempt a field goal, and when a team inside its opponent’s ten-yard line should pursue a touchdown rather than kick. He analyzed nearly every game from the 1998, 1999, and 2000 NFL seasons—700 games in all—using dynamic programming. The method allowed him to estimate the value of a first down at each point on the field, rather than treating every successful possession or scoring opportunity as equivalent.

His estimates made field position consequential in specific, comparable terms. A first-and-ten at a team’s own twenty-yard line was worth a little less than half a point; at midfield it was worth about two points; at the opponent’s thirty-yard line, three. He also calculated how often teams facing fourth down would gain a first down if they went for it, and how often they could make a field goal. That gave him a common scale for comparing the choices, including what a failed attempt would cost in field position.

The Case for Going for It

The comparison could reverse what looked like the safer play. On fourth-and-three at the opponent’s thirty-two-yard line, Romer’s example gave a team a 60 percent chance of gaining a first down. Since a first down at the opponent’s twenty-nine was worth three points, going for it had an expected value of 1.8 points. A field-goal attempt from the thirty-one-yard line, by contrast, was worth barely more than a point. On those assumptions, the aggressive play was the better one.

Romer’s analysis also included the consequences of failure. A long field-goal attempt was not simply a contest between three potential points and no points: if it missed, the opponent would take possession at a favorable spot, such as its own thirty-five-yard line. Romer’s calculations made it possible to compare that field-position loss with the potential value of a first down. The argument was not that risk could be ignored, but that the risks of kicking and going for it had to be assessed together.

His recommendations were striking by NFL standards. When a team needed three yards or fewer on fourth down, he generally advised trying for the first down; between midfield and the opponent’s thirty-yard line, he thought teams should be more aggressive still. Inside the opponent’s five-yard line, he argued, a team should always go for a touchdown. For example, on fourth down at the opponent’s two, a field goal offered three points, while a touchdown attempt succeeding 43 percent of the time had roughly the same expected scoring value. But a failed attempt would leave the opponent pinned at its own two-yard line, an additional advantage to pursuing the touchdown.

At midfield, his reasoning likewise turned on what success and failure would mean in context. On fourth-and-three, if the attempt succeeded half the time and failed half the time, the teams would have the ball at roughly the same field position after failure. Success, however, would bring an average gain of six yards. The successful outcome therefore left the team better positioned than the failed outcome left its opponent, supporting a more aggressive choice.

Conventional Wisdom and the Rams

The study’s implications illuminate the decisions of Mike Martz in the Super Bowl. With the Rams at the Patriots’ thirty-two-yard line and needing three yards, Martz sent out the kicking team rather than attempt to gain a first down. Romer’s calculations suggested that going for it was justified even if the Rams needed nine yards at that point, because a missed field goal was costly and the field-position penalty for failing to convert was relatively small. With the Rams’ offense, the value of trying for the first down might have been higher than the league-wide average.

Six minutes later, after another Rams drive stalled at the Patriots’ thirty-four, Martz again chose a field-goal attempt, this time with five yards needed for a first down. The kick went wide left and the Rams scored nothing. The book presents the choices as consistent with NFL convention: coaches typically preferred taking available points. Romer’s analysis instead suggested that both decisions were poor. It does not establish that either decision caused the Rams’ eventual loss, but it shows how conventional practice could steer a team away from the higher-value option.

Findings, Limits, and Significance

Romer’s broader review found many instances in which teams could have benefited from trying to convert. Among first-quarter fourth-down plays across the three seasons, he identified 1,100 plays where the teams would have been better off going for it. Instead, they kicked the ball 992 times. The contrast sharpened the puzzle: coaches were experts with strong incentives to win, yet a strategy with substantial expected benefits was not widely adopted.

The analysis had limits. It could not fully account for momentum, and league-wide averages might not capture the particular chances of an individual team. Football is also a dynamic game in which skill, strategy, emotion, and luck can be difficult to separate. Those qualifications leave room for adjustments, but the book argues that the potential gains were too large to dismiss as a statistical fluke. Romer’s work thus serves not only as a proposed change to fourth-down tactics, but as an example of how numerical analysis can expose the distance between what a professional group treats as established wisdom and what its own objectives may call for.