Super Bowl XXXVI

Super Bowl XXXVI, played between the heavily favored St. Louis Rams and the New England Patriots, becomes a concrete example of the gap between conventional wisdom and sound collective judgment. Rams head coach Mike Martz twice chose a field-goal attempt rather than trying to gain a first down on fourth down. A later statistical analysis by economist David Romer suggested that both choices were excessively cautious, turning a high-profile game into an illustration of how expert convention can resist better strategy even when winning is the stated goal.

Expectations and Stakes

Before the game, the Rams were fourteen-point favorites. Their case looked formidable: St. Louis had one of the most potent offenses in NFL history, had led the league in eighteen statistical categories, and had outscored its opponents 503 to 273 during the regular season. Victory seemed close to assured, making the decisions that followed especially revealing: the contrast was not between a weak team and a desperate gamble, but between an accomplished offense and a cautious approach to converting its opportunities.

Martz’s Fourth-Down Decisions

Midway through the first quarter, the Rams made their first major drive, advancing from their own twenty-yard line to the Patriots’ thirty-two. On fourth down, with three yards needed for a first down, Martz sent kicker Jeff Wilkins onto the field. Wilkins made the kick, giving St. Louis a 3–0 lead. The choice secured points, but surrendered the chance to extend the drive and pursue a larger gain.

Six minutes later, a second Rams possession stalled at the Patriots’ thirty-four-yard line. This time St. Louis needed five yards for a first down. Martz again chose the kicking team. Wilkins’s attempt went wide left, and the Rams received no points. The two situations differed in distance and result, but shared the same basic decision: Martz favored the conventional field-goal option over trying to keep the offense on the field.

Romer’s Strategic Reassessment

In the book’s account, the choices aligned with standard NFL practice. Coaches almost always took a field goal when one was available, following the conventional view that teams should take points when they can. Romer’s analysis offered a systematic challenge to that habit. Using dynamic programming, he examined nearly every game—about seven hundred in total—from the 1998, 1999, and 2000 NFL seasons, calculating the value of a first down at different positions on the field and evaluating when teams should go for a first down rather than punt or kick.

The examples clarify why a field goal is not automatically the prudent choice. On fourth down at an opponent’s two-yard line, a field goal yields three points, while a touchdown attempt that succeeds 43 percent of the time has an expected value of roughly three points. But a failed attempt also leaves the opponent pinned at its own two-yard line, so the field position strengthens the case for trying to score a touchdown. At midfield, a fourth-and-three can likewise favor aggression: even if the attempt succeeds only half the time, success gains an average of six yards, while failure leaves the teams with the ball in essentially the same field position. In each case, judging the decision requires accounting for what happens after either outcome, not merely comparing the immediate certainty of a kick with the possibility of a failed conversion.

Romer’s calculations suggested that Martz would have been justified in going for a first down at the Patriots’ thirty-two even if the Rams needed nine yards, rather than the three they actually needed. At that field position, the chance of missing a field goal was high and the cost of the resulting field position relatively small. The case for attempting the conversion would presumably have been stronger still for an offense as potent as St. Louis’s. Romer’s framework therefore made the Rams’ two choices look not merely conservative in retrospect, but difficult to defend as the better route to maximizing their chances.

Limits, Outcome, and Significance

The analysis had acknowledged limits. Romer could not incorporate every factor, notably the possible effect of momentum, and his numbers averaged performance across the league; individual teams would need to adjust the calculations to their own chances of success. Football is also a complex, dynamic game in which skill, strategy, emotion, and luck are difficult to separate. Still, the book presents the gains from fourth-down aggression as large enough that these caveats were unlikely to explain away the result as a statistical fluke.

The Rams lost the Super Bowl, but the account does not claim that either one of Martz’s decisions caused the defeat. Its point is more careful: the loss was not surprising in light of choices that Romer’s work judged to be poor. Nor was Martz an isolated case. Across the first-quarter fourth-down plays in the three seasons Romer studied, the analysis identified 1,100 occasions when teams would have been better off going for it; instead, they kicked 992 times. The game thus anchors a broader puzzle: coaches were experts with strong incentives to win, yet conventional caution persisted despite evidence that greater aggressiveness could provide a competitive edge.