Once again, with the Indianapolis Colts' early exit in the playoffs, we're seeing (on various websites and in the Twitterverse) and hearing (on this or that TV and radio show) that Peyton Manning(notes) has a "losing record" in the playoffs. In his 19 postseason appearances, Manning apparently has a 9-10 record, which is supposed to mean something. Never mind his status as quite possibly the greatest quarterback ever to play the game. Never mind the fact that he does have a Super Bowl ring, and managed to take his 2010 Colts team to the three-seed despite injuries to several of his skill players, a running game that could charitably be described as "iffy," and an offensive line with major holes.
And never mind the fact that in his postseason history, Manning has completed 435 passes in 718 attempts (a 63.1% completion rate) for 5,389 yards, 29 touchdowns and 19 interceptions. But in Manning's "loss" to the Jets on Saturday night, he had the fourth-best game of his season if you subscribe to the quarterback rating statistic. Somehow, despite all his accomplishments, his "quarterback wins" stat in the playoffs somehow denigrates his legacy and puts him behind Tom Brady(notes) in the Pantheon.
More and more in the last few years, we've heard quarterback wins thrown out there as a supposed indication of player performance. And as ridiculous as wins and losses have always been for pitchers, they're far worse measurements for quarterbacks. The same analysts who use "wins" as a bar for performance will turn right around and point to the fact that football is the ultimate team sport, which hardly makes sense. More reasonable users of the wins concept will admit that while it isn't a statistical catch-all, it does tell you a lot about a quarterback's value to a team. After all, they say, the quarterback is the most important player on the field, and it's every team's job to win, right?
Well, sticking with Manning, let's go back to his championship season of 2006, when the Colts beat the Chicago Bears in Super Bowl XLII. In that season, the Bears actually finished the season with a better record (13-3) than did the 12-4 Colts. If we are to hold to the QB Wins metric to any degree, we then have to put one Rex Daniel Grossman III in the same league as Manning for that season.
Why? Because Grossman "won" games that season in which he posted quarterback ratings -- another flawed chestnut analysts love to use -- of 64.9 (against the Minnesota Vikings), 10.2 (against the Arizona Cardinals in the infamous "They are who we thought they were" game), and 1.3 (against the Vikings again). In fact, if you include the postseason, Grossman's 15-4 "record" puts him on Manning's level, though we should probably ding him a point or two for not winning the big one.
After all, it's all about wins, right? Stats are for losers!
Let's take a closer look at the two quarterbacks. Manning had an MVP season in which he was a primary component of his franchise's first Super Bowl win since 1970 despite the worst regular-season run defense of the expansion era. Grossman, who finished 29th in Football Outsiders' efficiency metrics for quarterbacks in 2006 while Manning finished first, had a great defense and a completely amazing special teams array to buttress his often mediocre performances.
Younger quarterbacks today are just as affected by this selection bias. You'll hear that Baltimore's Joe Flacco(notes) has already "won" three playoff games, while Aaron Rodgers(notes) won't be a great quarterback -- no matter how totally and completely Rodgers has outperformed Flacco to date -- because Rodgers hasn't "won" a playoff game at all. [UPDATE: Until Sunday, when Green Bay beat Philadelphia.] Well, Flacco has participated in three Ravens postseason wins, and he's thrown one touchdown and one interception total in those three games. You might argue that Flacco did a lot more to help lose two playoff games in which he threw a total of zero touchdowns and five picks.
Flacco's total postseason numbers go something like this: 57 of 120 for 660 yards, one touchdown and six interceptions. In Rodgers' one playoff game (against the Arizona Cardinals in 2009), he completed 28 of 42 passes for 432 yards, four touchdowns, and one pick. The only reason he "lost" that playoff game was because Kurt Warner(notes) put up what may have been the greatest single-game passing performance in playoff history. Rodgers did everything he possibly could to help his team win -- it's just that Warner did a little bit more. To tack a loss on Rodgers' performance is absolutely moronic.
Just as baseball sabermetricians finally came up with more relevant numbers to accurately gauge pitcher performance (a primary reason that Zack Grienke and Felix Hernandez were able to win the Cy Young awards they deserved despite their pedestrian won-loss records), football statheads (and I count myself as one) need to continue to think outside the box and figure out ways to assess just how much quarterbacks have to do with the outcome of each individual game.
Until then, the wave of talking heads putting far too much importance on one little number will continue to skew the true value of the game's most crucial position.