You may wonder what I think of the passer rating statistic. In order to even address that sentence, let me clarify something first.
A statistic measures one of two things. One is production/output, and the other is efficiency. There could be something beyond that, something unique like Bill James’ Fibonacci Win Scores (which is pure genius); to the best of my knowledge, nothing like this actually exists in the world of football. This renders it impossible to actually combine production and efficiency. My own systems that I employ in do it, but I recognize this and plod on in trying to improve what I have already committed to print.
The lovely thing about quarterback/passer rating is that it was based on a bunch of numbers that theoretically applied at a certain period of time. But it's an idiotic mathematical algorithm based on arbitrary numbers and guesses. It's one of the dumbest creations in the world of sports ever devised, and it boggles the mind to see anyone who fancies themselves to be a smart individual use it to settle anything. You might as well just junk it and go to "Well, I think that..." because that has as good of a chance as QB rating does of decisively making a statement.
Have you ever wondered what goes into this rating system and how it’s calculated? Well, look no further, because I’m going to tell you once and for all how it’s done at the NFL level in a very easy step-by-step process.
1) Take the number of completions and divide it by the number of attempts
2) Multiply that number by 100
3) Subtract 30 from that number
4) Divide that total by 20. Call this letter “C” and set aside
5) Divide the number of passing yards by the number of passes attempted
6) Subtract 3 from that number
7) Multiply the total by ¼. Call this letter “Y” and set aside
8) Divide the number of touchdown passes by the number of passes attempted
9) Multiply that number by 20. Call this “T” and set aside
10) Divide the number of interceptions by the number of passes attempted
11) Multiply that number by 25.
12) Take the number you’re left with and subtract it from 2.375. Call it “I” and set aside
13) Almost done! Now just add C, Y, T, and I together.
14) Divide that sum by 6
15) Multiply that last number by 100
16) In just 15 easy steps, you’ve calculated a passer rating!
The reason that I am against using this in any way, shape, or form, is because it’s just dumb. It creates an entire system where the Custodian quarterbacks get a huge boost because of the fact that they don’t throw the ball very much and the throws they do make are the “safe” ones. It skews toward West Coast-style quarterbacks as far as style goes.
But now you ask yourself, "What is a better way?" A better way is one that measures what a quarterback produces (and costs) for his offense in a basic number that correlates to actual points on the scoreboard. For this, it's necessary to understand differing levels of efficiency. In my explanations on OER and DER, I referred to a certain threshold as being average, with a number exceeding it being more efficient than normal and one below it being inefficient. In the case of my Points Generated system that I am about to describe, it works the same way.
Points Generated is a basic mathematical algorithm that correlates to actual points scored leaguewide. Trust me, I tested it. By jamming a bunch of numbers into a formula, I've managed to place the number of theoretical points extremely close to the actual points for every single year of professional football history. The widest variance comes in years in which the caliber of play suffered a decline; this is most obvious in the WWII years. But over the entire span of NFL history, the difference between theoretical points and actual points is 0.533 points per game. Quite frankly, that's insane.
Quarterbacking is about more than passing the ball, and it's certainly about more than being able to complete a high number of passes. The Points Generated formula takes this into account and includes rushing totals.
What is the formula? Here it is for quarterbacks. (Total yards / 22.5) + (total touchdowns * 3) - (interceptions) = Points Generated. There's another one that measures overall efficiency, which involves Points Generated / Offensive Opportunities (offensive opportunities defined as total passes plus total rush attempts). I greatly prefer the efficiency model, as it correlates very strongly to winning and losing games; it also gives quarterbacks on run-heavy teams their due and removes the inherent bias toward those on awful teams who simply throw for a ton of yards in garbage time.
What I and most coaches care about is whether offensive yardage and points get produced. The two aren’t mutually exclusive of each other, either. I can live with seeing a quarterback have a 45% or a 50% completion percentage as long as the yards are being produced. If it takes running the ball, great. If it takes passing the ball, great. The bottom line, and the only real important thing, is that the yards are being produced.
Yes, it’s irritating to see a quarterback chuck a ball askew from an open running back six yards downfield. It’s a wasted down. But, on the other hand, if he comes right back and hits a receiver on the next play for 55 yards, do I really care about what happened the previous play? It may warrant extra practice time or a closer look in film sessions (after all, the object is to make your offense more efficient), but the yards are produced and the points are on the board.
Points Generated doesn’t even address any efficiency ratings. Completion percentage, touchdown/interception ratio, and touchdowns per pass ultimately mean nothing. I want a quarterback to lead a scoring drive. I want the field position gained, I want the offense on the field, I want the points on the board. Period.