The Six-Point Lead That Wasn't: Auditing Sirianni's Field Goal

# The Six-Point Lead That Wasn't: Auditing Sirianni's Field Goal **By Raffaele Cuccinotta | November 2, 2026** Here's the situation: Monday Night Football, Sept. 16, 2024. Philadelphia leads Atlanta 18-15. The Eagles have the ball on the Falcons' 10-yard line, fourth-and-3, roughly a minute and forty seconds left. Nick Sirianni sends out the field-goal unit. The kick is good. Philadelphia leads 21-15. You know what happened next. Atlanta drove 75 yards in nine plays, scored the touchdown, and won 22-21. The postgame takes wrote themselves: Sirianni played it safe and paid for it. But the model says one thing, the coach does another, so let's look at why. ## What the Numbers Said Every major fourth-down model I could find favored going for it in that spot. ESPN's win-probability model said coaches should go for it on anything shorter than fourth-and-9 from there, according to CBS Sports. NFL NextGen Stats showed that kicking the field goal dropped Philadelphia's win probability from **94 percent to 85 percent**. Ben Baldwin's NFL4th.com bot calculated roughly a **3.1-point win-probability gain** for going for it and labeled it a "strong" lean toward keeping the offense on the field. Even Wharton's model, which had a wide confidence interval of minus-5 percent to plus-8 percent, still leaned slightly toward going. The logic is straightforward. A three-point lead means the trailing team plays for a field goal and overtime. A six-point lead means they have to score a touchdown. If you convert the fourth-and-3—and the league converts fourth-and-3 somewhere north of 50 percent of the time—you can kneel out the clock or at least force Atlanta to burn its timeouts. If you miss, the Falcons still need to go 90 yards with no timeouts and about 90 seconds. ## What the Coach Said Sirianni's explanation after the game was simple: he wanted to "go up six points." It's the kind of reasoning that makes sense in a living room and falls apart under a microscope. Six points feels safer than three. It feels like a two-score game. But feelings aren't probabilities. The CBS analysis made a point I think is fair: the model's recommendation depends partly on assumptions about how the trailing team behaves. A team down three plays conservatively for a field goal. A team down six has to throw. If you trust your defense more than your offense, you might prefer the six-point lead and the desperation drive. But Philadelphia's defense had just given up 15 points to a Falcons offense that was hardly elite that night. ## Was It the Wrong Decision? I could be wrong on this, but I think the answer is: probably, but not certainly. The consensus across models is strong enough that kicking looks like a mistake on the ledger. When every independent estimate says you're giving away roughly a tenth of your win probability for the comfort of a six-point lead, that's a costly comfort. But Wharton's wide confidence interval matters here. The true gain from going for it might have been as low as negative-5 percent or as high as plus-8 percent. In other words, the models aren't as certain as the headlines suggest, and the difference between a good decision and a bad one in that specific moment may have been thinner than the postgame pile-on implied. What tips me toward calling it a mistake is the conversion probability. Fourth-and-3 from the 10 isn't a gimme, but it isn't a Hail Mary either. The Eagles had Jalen Hurts, A.J. Brown, and an offensive line that was supposed to be the strength of the team. If you don't trust that group to gain three yards with the game on the line, what exactly do you trust them to do? ## The Outcome Trap The hardest part of this job is keeping decision quality separate from outcome quality. Atlanta's touchdown drive makes the field goal look like a disaster. But if Marcus Mariota had thrown an interception on the ensuing possession and the Eagles had won 21-15, we'd be calling Sirianni a conservative genius who trusted his defense. The result doesn't retroactively validate or invalidate the process. That said, the result does matter for one thing: it puts the decision under a microscope. And under that microscope, the process looks shaky. Sirianni wasn't choosing between a field goal and a punt. He was choosing between a field goal and a chance to end the game. He chose the field goal, and the game didn't end. ## What Remains Unknown One thing none of the models can tell me is what Sirianni knew that I don't. Was Hurts nursing something? Was the left tackle struggling? Did Sirianni have reason to think the Falcons' two-minute defense was vulnerable to a deep shot but not a short conversion? I doubt it, but I wasn't in the headset. The model estimates the average case. The coach lives in the specific case. The gap between those two is where journalism lives. ## The Ledger Decision: Kick 28-yard field goal on 4th-and-3 from opponent's 10, up 3, ~1:40 left. Model recommendation: Go for it (strong consensus, with Wharton as a cautious lean). Coach's reasoning: "Go up six points." Outcome: Atlanta touchdown, Eagles loss. Decision quality: Likely suboptimal, but not a fireable offense on its own. The bigger pattern—Sirianni's repeated conservative choices in high-leverage moments—is what the seat gets hot from. Small sample, but it's worth noting: one call doesn't make a coach, but it can expose the philosophy. And this philosophy looked like it was playing not to lose. --- *Sources: CBS Sports; The Philadelphia Inquirer; Wharton Sports Analytics and Business Initiative; ESPN; NFL NextGen Stats; NFL4th.com.*