When the Eagles' Nick Foles and the Patriots' Tom Brady take the field as the starting quarterbacks at U.S. Bank Stadium on Sunday, they will represent one of the most lopsided QB matchups in the 52-year history of the Super Bowl. But it's probably not close to being the worst pairing ever.
So which one is? We could compile a list of what we think are the worst matchups and then ask you to vote (which we'll also do; just scroll down), but that approach would be tainted by our and your biases. How do we remove that bias from the equation?
The question is whether there is a way to statistically measure the gap that exists between Brady — who is in the 17th season of the most accomplished, and likely best, quarterback career in the history of football — and Foles — who has never been a starter for a whole season in six years of pro football and has been on three teams the last three years — and compare that gap to the previous 51 quarterback pairings in the Super Bowl.
There is no modern-day catchall statistic for measuring football performance like baseball's wins above replacement or basketball's win shares. Football Outsiders has a metric called DYAR, defense-adjusted yards above replacement, to measure QBs, but the results go back to only 1986, leaving the first 20 Super Bowls out of the mix. ESPN's total QBR statistic goes back to only 2006. Using those numbers would exclude a matchup many would predict to be one of the worst ever.
That leaves us with the NFL's passer rating. It was adopted by the league in 1973, but because it's calculated using a player's common statistics — completions, attempts, yards, touchdown passes, and interceptions — passer rating is available going back before the Super Bowl era. It's far from a perfect statistic, especially when applied to small samples. (It's unlikely Falcons receiver Mohamed Sanu, with his perfect 158.3 passer rating from one pass, was the best quarterback in the NFL this season.) But the larger sample we take, the better picture passer rating provides. And if we use that number and consider factors such as usage, experience, and performance that season, we can rate the pairings.
We can't simply compare passer rating numbers, however. It's a rate statistic that doesn't consider volume. There are other factors fans and pundits typically consider when comparing quarterbacks, particularly experience and usage.
If you stack up two quarterbacks with similar ratings against each other, the more-veteran player is typically considered more valuable. And a quarterback who has thrown more passes is usually thought to be more important to his team's offensive performance.
So, to use passer rating as a proper tool, the number needs to be weighted for those factors. For each Super Bowl, we multiplied the career regular-season passer rating before the game (because we want to know what the general feeling was that day) for both starting quarterbacks by games played and pass attempts. By comparing those numbers for each quarterback, we can compute a percentage difference, then average those results to give us an index to compare to other games.
Of course, the single season a quarterback is having going into the game is a consideration, too. We can create a number based on the same formulas for a single season, and then put that into the mix, too. The question then is, how much consideration is given to career vs. season? We produced three sets of results based on a single season being worth 40 percent of the index, half of the index, or 60 percent of the index.
How does it all come together? Let's look at Super Bowl XV as an example. To that point in his career, the Eagles' Ron Jaworski had a career 72.3 passer rating over 1,693 pass attempts in 86 regular-season games. The Raiders' Jim Plunkett had a 62.1 rating, with 2,329 attempts in 104 games. That season, 1980, Jaworski had a 91.0 rating over 16 regular-season games with 451 attempts, compared to Plunkett's 72.9, 13, and 320. Here are the calculations:
The Jaworski-Plunkett matchup is one of the closest in Super Bowl history, by our measurement.
The Bradshaw-Ferragamo matchup in January 1980, which the Steelers won, 31-19, is by far the "winner." If you consider a player's career is worth 60 percent of the score and the lead-in season 40 percent, the gap between Bradshaw and Ferragamo is more than twice as large as the one between Brady and Foles.
There are a lot of similarities between the Super Bowl XIV quarterback pairing and the one coming up on Sunday. Bradshaw was in the 10th season of a Hall of Fame career, and Ferragamo was in his third season and had started only five games in his career — all in 1979. His career passer rating of 51.4 going into the game is the worst ever for a Super Bowl starting quarterback.
When we adjust the career-season balance to an even split, the top two spots remain the same, with Bradshaw-Ferragamo still having a gap more than twice that of Brady-Foles.
If we adjust the balance so the lead-in season is worth 60 percent of the advantage index and a career is worth 40 percent, the top two remain the same, but the difference is not as great. In that case, the Super Bowl XIV mismatch is only 1.9 times as big as the Super Bowl LII gap.
Finally, to get a composite of the three, we can apply a ranking to each matchup in the three scenarios (52 for the biggest mismatch, 1 for the smallest), and then average those rankings.
A few notes about the three scenarios:
The Eagles' plan wasn't for Foles to be playing today. Carson Wentz might have been on his way to a league MVP award before tearing the lateral collateral and anterior cruciate ligaments in his left knee during the Birds' 13th game of the season. Would the Patriots enjoy such a large advantage at quarterback if Wentz had been healthy all year?
No. But it would still be one of five biggest mismatches in Super Bowl history.
Why? This is only Wentz's second season, and while he's had a terrific start to his career, it doesn't come close to measuring up to what Brady has accomplished. And when comparing the single seasons Brady and Wentz just compiled, Brady gets the edge even once you prorate Wentz's performance to a full 16 games. Brady had a slightly better passer rating, and his 581 pass attempts were more than the prorated 541 Wentz would have thrown.