One of the big issues with “individual NFL stats” is that there is so much entanglement because of teammates, coaching, and opponents (not to mention randomness and luck). A short, benign screen pass can turn into a long touchdown with a great play or missed tackle. An interception can bounce right off a receiver’s hands and go to a defender.
Today, I’m going to brainstorm a rough test of how much of a stat belongs to the quarterback, versus how much of it is because of things outside his control. For the purposes of what I am trying to look at here, we are talking about apportioning responsibility a competent NFL starter playing with a supporting cast of NFL players. If I suddenly took over at quarterback for a NFL team, we might conclude that the difference in quarterbacks was 100% responsible for the stats, because teams would go to roughly 0.0 yards per attempt. Conversely, if we put Peyton Manning in a commercial with 8-year olds, going against a real NFL defense, we might conclude that it’s all about the bodies around the quarterback that make the difference.
So here’s what I came up with: let’s look at the range of outcomes for NFL quarterbacks who played enough seasons to get a good range. That range shows us the variability for individual QBs, which is largely going to be because of things like teammates, scheme and coaching changes, opponents, and luck. Then, let’s compare that average range, and the average lows and average best years in categories, to the absolute worst of the worst of NFL results.
I used the last 30 years, and took all quarterbacks who had at least six different seasons where they qualified for a passing category by having enough attempts (224).
Here’s what our fictional “worst of the worst” quarterback (I took the bottom 10 seasons in each category, out of nearly 1,000) would look like:
47.0% completion percentage, 5.1 Yards per Attempt, 1.1% Touchdown Rate, 6.0% Interception Rate, 13.3% Sack Rate
So let’s start with sack rate, which I talked about Friday as clearly being a Quarterback Stat. But let’s see what that means.
There were 66 qualified passers who have thrown 224 or more passes in at least six seasons since 1988. On average, the best sack seasons for this group was 4.0%, and the worst sack seasons were at 9.1%. That gives us an average range of 5.1%. I then subtracted 13.3% (our fictional worst realistic NFL QB performance) minus 9.1%. That gives us 4.2% as the average good quarterback contribution to the sack rate, over the baseline worst possible result. Divide that amount by the difference between the worst possible and the average best contribution (9.3%), and you get the following:
Sack rate is at least 45% the responsibility of the quarterback, and about 55% everything else. In this case, everything else includes offensive teammates, coaching/scheme/offensive philosophy, opponents, game scenarios, injuries that affect performance, aging/inexperience (as a subset of the worst seasons in categories tend to come in the first or last year starting), and randomness/luck.
Does that sound low or high?
Well, if you question whether it is a quarterback stat because the quarterback himself may only be responsible for about half of it, then let’s move on to the other quarterback stats that are regularly cited and used in passer rating.
By the same technique, here is the percentage of quarterback responsibility for each stat:
Completion Percentage: 44%
Interception Rate: 40%
Yards Per Attempt: 38%
Touchdown Rate: 34%
Consistent with all the research on what happens when quarterbacks change teams, when teams change quarterbacks, and which stats show the highest variation among individuals, Sack Rate and Completion Percentage come out on top.
Let’s try to put these numbers in some context. What is a commonly recorded stat in another sport that is less susceptible to teammate contribution? I’d say Free Throw Percentage in the NBA. I would expect free throw percentage to show a smaller range of outcomes, or variation. There’s still some luck involved because a few bad games or good bounces can swing a percentage 5-10% over any season.
When I used the same method as above for our NFL stats, Free Throw Percentage comes in at 76% as being the base responsibility of the player, with variation largely due to randomness, and improvement of the player due to experience, or aging.
So if you want to say that things like aging/decline and improvement in a performance category due to experience are also the responsibility of the player, then the answer to how much is the quarterback is responsible for changes.
But at minimum, I’m comfortable giving an estimate that a quarterback is actually responsible for at least 34-45% of the stats in the various categories, and at least 40% of the overall passing production of an offense. The worse you think the reasonable possibilities are with a quarterback in a NFL offense–something that Nathan Peterman is testing–then the more you will think a quarterback is responsible for. The more the spread between replacement level quarterbacks and top starters grows, the more the quarterback will seem responsible for the stats. This isn’t offered as a definitive answer but as a way to advance looking at, so feel free to message me with any thoughts.