Passer rating (QB rating) explained

Have you ever looked at the formula for the passer rating and asked yourself "What the heck?" Computing the thing is very tedious (so use a computer), but not difficult (it only involves addition, subtraction, multiplication, and division). After I first found out about it, I wondered where the formula come from? So I went a-hunting for an explanation and eventually, I found this article

I think it does a pretty good job at explaining the origins of the formula, although I did have to read a couple of paragraphs more than once and it still leaves some questions unanswered (like why they cap each of the 4 components of the formula at 2.375). (Also, it's not a quadratic formula, as the article says. But that's beside the point).

After reading the article, I'm a bit sceptical of the use of the formula to rate CFL quarterbacks, since, according to the article, it's based on the average performance of NFL quarterbacks up to the early 70s, before that league implemented rules to encourage passing. Thus, we're judging quarterbacks in the more pass-oriented CFL by a formula that was developed for a less pass-oriented NFL in an era when passing was even less important for that league than it is now.

So, although I doubt anyone is going to do this, I think it would be interesting to see what would happen if the CFL adopted the same (or similar) methodology as the NFL did, but used CFL data instead of NFL data, to come up with a "made-for-CFL" formula, so to speak. It makes more sense to me to judge CFL quarterbacks against other CFL quarterbacks rather than NFL quarterbacks, given the differences in rules and game play between the two leagues.

Thanks, PiCat for that cool info on QB ratings. I knew it was a math formula based on NFL stats, but I never knew it was based on stats from the 70s.

I'm not sure it's really necessary to rework the formula to make apples to apples type comparisons, though. If a CFL quarterback has a rating of 110, he is compared to other CFL quarterbacks, who have ratings of, say, 90 and 120. These guys are all playing in the same league, under the same rules, so the comparison is valid because it uses the same formula.

As for comparing between NFL and CFL, they are different games, so I don't think that you can make an apples to apples comparison to begin with. Using a different formula in the CFL than in the NFL would only complicate matters further, IMO.

Now, if the goal is to update the formula, so that a "perfect" qb has a rating of 100, in order to make it more understandable to a casual fan, then THAT would justify rejigging the stats and equations. But the only difference would be the final number, but not the relative rankings. The only difference is that their new numbers would be 80, 64, and 88, respectively. The guy with the 110 rating on the old scale would still be between the guys with 90 and 120 ratings in either system.

Even current NFL QBs have "inflated" numbers due to the older rating system, but I think noone wants to change formula since doing so would make harder to compare current players to those from older days

The NFL system is flawed in other ways besides the age of the data, which is a huge flaw. A "perfect" rating is 158.3 but its possible to have this rating in a game, then go out and throw a 35 yard touchdown pass with no further impact to the rating. Announcers love to calculate the game passing rating for players during games, but it never really was designed for that. It was designed to be a seasonal rating.

I'm sure that someone could easily come up with a good system that would address the "data age" flaw. As the OP said, there have been countless changes to the rules that have ALL had the effect of encouraging passing over the years. This started in about 1974, when the NFL was under at least mild competitive (read: "entertainment" ) pressure from the old original WFL and its all out passing attacks.

Comparing across eras is almost impossible in the NFL. Too many changes in playing personnel, equipment, size/shape of the ball, rules, stadia, domed stadia versus the old weather affected games that were once commonplace, jet travel (time zone changes), the list is endless.

What I would do is calculate the average performances in a game and the season for the league, and with that as a benchmark see how each player (passer/QB) did against those benchmarks. This has a disadvantage in that accurate ratings can't be compiled until seasons end, but you could always use the previous seasons numbers as the benchmark for THIS season, then update each year. This would give an excellent picture of how the QB's compare against contemporary competition.

The only way to compare across eras is to see who was the most dominant &/or the farthest above their contemporaries in terms of performance. In the NFL, that would probably make Benny Friedman (I can hear it now... "Who???!!!" lol!) of the 1920's NFL the greatest NFL passer of all time, as his performances were so far above those of his peer competition as to be totally ridiculous.

In the CFL... I dont know enough about the evolution of rules and conditions to say. But the statistical record only goes back to 1953 or so, I believe I have read. With only 9 teams, it should be a relatively easy project for someone that is both historically and statistically minded to rate all CFL passers back to 1953 (or more if the numbers are there) based on the stats and come up with pretty reliable ratings.

Great comments by all here.

After reviewing countless profiles of NFL QBs via over the years to compare QBs in the "Super Bowl era," including many a profile of a Hall of Famer and any QBs on a team with more than one Super Bowl win, I have used an NFL QB rating of 80 as a good rule of thumb to designate successful seasons by any given quarterback.

It's a rule of thumb so don't take that as definitive please, as in the case for example when a QB throws almost as many TDs as interceptions though somehow gets a rating of more than 80 due to high yards and high completion percentage and so forth.

However, due as explained below the fact that the calcuation is antiquated and based on the old rules, which included amongst other differences more allowable hits on the QB as well as the fact that DBs could bump receivers all over the field until the ball was thrown instead of the 5-yd maximum in place since 1978, a rule of thumb for a comparable rating for a modern QB is more around 85 compared to a peer from the era prior to 1978 around 80.

Also noteworthy is that the rating has in some seasons had little relevance to a team's overall performance, including winning the Super Bowl, when especially the team won most games through the use of a ball control strategy with tough defence and few turnovers more than the stereotypical and always hyped drop-back passing game.