On Saturday, I talked about Sports Illustrated’s piece on the self-described “Most Important Stat” in football being Passer Rating Differential. I was skeptical of it, not because passing offense and defense are not important, but because it didn’t make the case that the specific passer rating formula was optimal.
The Passer Rating formula is unwieldy and convoluted. Brian Burke’s criticisms laid out here are valid. It still has a strong relationship with winning. This doesn’t mean it is the optimal formula, though. The current passer rating utilizes four different categories: yards per attempt, completion percentage, touchdown percentage, and interception percentage. They are capped and scaled (I won’t bore you with the details) with each category comprising 1/4 of the final number. Well, except completion percentage is partially double counted, because yards per attempt is equal to yards per completion x completion percentage, so that teams with higher yards per attempt already tend to be teams with higher completion percentages.
My argument that the case hadn’t been made yet, is that it hadn’t been compared to other passing measures or tested to see if different weights would provide better results. My guess was that Yards per Attempt would be more important for winning and championship teams than completion percentage, even though they both count equally in the formula.
So, I went through the champions since 1940, though I will present just those since 1970. The reason: before 1970, there were between 10 and 16 teams, and for most of the period no playoffs that introduced more randomness, so finishing in the top 2 was easier for a champion. Anyway, I found the relative rankings for each Super Bowl Champion since 1970 in the eight sub-categories that are part of Passer Rating Differential. Here are the relative order for the Champions.
1. Defensive Yards per Pass Attempt. (Average Ranking for SB Champ: 6.5, 35 of 41 better than league average)
2. Defensive Touchdown Percentage (Average Ranking for SB Champ: 6.6, 37 of 41 better than league average)
3. Offensive Touchdown Percentage (Average Ranking for SB Champ: 7.0, 37 of 41 better than league average)
4. Offensive Yards Per Attempt (Average Ranking for SB Champ: 7.1, 37 of 41 better than league average)
5. Defensive Completion Percentage Allowed (Average Ranking for SB Champ: 8.7, 33 of 41 better than league average)
6. Defensive Interception Percentage (Average Ranking for SB Champ: 9.0, 31 of 41 better than league average)
7. Offensive Completion Percentage (Average Ranking for SB Champ: 9.2, 32 of 41 better than league average)
8. Offensive Interception Percentage (Average Ranking for SB Champ: 10.3, 30 of 41 better than league average)
As you might suspect, yards per attempt on both offense and defense are the categories where our champions were relatively better on average, along with touchdown percentage. Touchdown percentage is reflective measure, that is influenced by other things, such as field position and special teams. The more throws you have to make to reach the end zone, the lower your touchdown percentage. These also tend to show, though, that the championship teams were pretty good at avoiding giving up big chunks of yards and big touchdown plays.
Completion percentage and Interception percentage rank lower. Some teams rank highly in these categories, but there are more counter-examples. When the West Coast teams won titles, completion percentage seems more important, until you look at vertical passing teams like the Steelers from the 1970′s, or the Redskins of 1987 and 1991. Both of Parcell’s Super Bowl teams, the Cowboys of 1993, and the 49ers of the late 1980′s ranked below average in completion percentage allowed on defense, so that Passer Rating Differential undersold how good they were by overweighting completion percentage. (The Colts in 2006 were 30th in completion percentage allowed also, despite being 8th in yards per attempt, showing teams tended to throw short to avoid pressure).
I wouldn’t base a measure solely on looking at these 41 champions, but they show what other teams have tended to show–yards per attempt is most important and should be weighted more heavily, while completion percentage is overrated in the current formula. The most important stat would be better if the weights were changed.
[photo via Getty]