Reading Super Bowl Box Scores Properly
Most people look at a final score and stop there. They see 38-35 and move on. The actual story is buried in the play-by-play data, EPA per snap, success rates, and red-zone efficiency numbers that nobody bothers to pull together. If you want to actually understand what happened in a championship game, you need to go deeper than the scoreboard.Here's how I approach game film combined with modern advanced metrics when breaking down a Super Bowl. The phrase sounds like something a fantasy football app invented for a headline, but the principle behind it is real. It's about taking the final number and reverse-engineering every drive, turnover, and situational failure that produced it. A 24-10 score looks dominant until you see that the defending champions scored three consecutive fourth-quarter touchdowns because the leading team's offensive line allowed six pressures inside the pocket. The score tells one story. The underlying data tells the truth. I keep a personal spreadsheet that tracks expected points added (EPA), success rate on offense, pressure rate on defense, and red-zone conversion percentage for every Super Bowl going back to 2010. When I cross-reference those numbers, patterns emerge that nobody discusses on broadcast television.
The Metrics That Actually Matter
Yards per play is useless in isolation. It gets gamed by big pass plays and draws to the five-yard line. EPA per play measures the value of each individual down based on where the offense was, what the score differential was, and how much time was left. A third-and-eight conversion that moves the chains in the fourth quarter of a close game is worth far more than a twenty-yard scramble that happens when you're up by thirty. Situation-success rate filters those plays down to standard downs. On third and fourth down plus runs inside the red zone, the metric shows which teams actually execute when it counts. Super Bowl LVII showed this clearly. The winning offense had a 62% situation-success rate while the defense allowed just 31%. That gap was wider than the point differential suggested. Pressure rate from the defensive side gets misused constantly. Total sacks are one thing. Pressure rate as a percentage of dropbacks tells you how consistently a defensive front disrupts timing. A team might record two sacks but generate pressure on forty percent of dropbacks, which collapses the pocket enough to force hurried throws and negative plays that never show up in the box score.
A Real Problem I Ran Into
After Super Bowl LVIII, I was pulling league-wide EPA data from public sources and noticed a consistent discrepancy. The official league stats showed the winning offense averaging 4.8 EPA per play while several independent tracking services listed it closer to 5.4. The difference came down to how each source counted play action passes. Some counted play action as a run for EPA purposes when it should have been categorized as a pass. Play action is a pass concept even when the quarterback hands the ball off, and misclassifying those plays skews the entire offensive efficiency picture. My workaround was straightforward. I pulled the raw play-by-play CSV from the NFL's data portal, filtered for every play labeled as play action, reclassified it as a pass concept, and recalculated EPA using the standard EPA formula with the adjusted play type. The revised number for that game matched the advanced tracking model I trust. It took about twenty minutes and required zero special tools beyond a spreadsheet and basic filtering.
Get the Full Details
Building Your Own Breakdown
You don't need a paid subscription to do this. The NFL's data API and the play-by-play CSVs are publicly available through stats.nfl.com. You also have access to Pro Football Reference's game logs, which include adjusted net yards per attempt, completion percentage over expected, and takeaways-plus giveaways for each game. Start with the final drive. Trace every play backward from the last scoring possession. Note the down, distance, field position, and result. Then layer in the pressure data and EPA values for those same drives. You'll immediately see which team won the situational battles and which side got lucky on field position or turnovers. The process usually takes about forty-five minutes for a full game breakdown if you're working from raw data. First time through it'll take longer. After five or six games you'll have the workflow down to under thirty minutes.
Where This Approach Falls Short
Advanced metrics are not a crystal ball. EPA and success rates describe what happened, not what will happen next season. Sample size in a single game is tiny. A quarterback throwing three interceptions on negative EPA drives could simply have gotten unlucky on ball placement that defensive backs tipped at the line. One game does not override a full season of production. Another limitation is that these numbers do not capture scheme adjustments in real time. Coaches change coverages and run calls based on what they see. The stat sheet records the result, not the adaptive decision-making that produced it. For that you need film study, which no metric can replace. If you want something faster and less manual, sites like Sharp Football Analysis and NFL Next Gen Stats provide pre-calculated EPA and success rate data for every game. They cost money or require accounts. The free data works fine if you don't mind doing the calculations yourself.
One Counter-Intuitive Finding
Teams that win the Super Bowl by double digits frequently lose the EPA battle. This sounds backwards until you examine the data. A lopsided score often comes from defensive touchdowns, punt returns, or interceptions that swing momentum without reflecting sustained offensive efficiency. The winning team might have generated fewer positive EPA plays but forced three turnovers that resulted in thirty combined points. The score was accurate. The underlying offensive performance was average at best. This is why looking at score alone is misleading. The next time you see a blowout result, check the EPA differential before drawing any conclusions about which team was actually better.
