Why is passer rating still being used as a stat?

[Todd Pence]

Passer rating is a fundamentally flawed statistic that sets arbitrary standards and gives equal weight to unequal categories. It should be given credit, when first conceived, as an attempt to express an overall rating of a team's passing game that went beyond total yards, which was a watershed idea at the time in formulating new and better statistical models that gave us more effective measurements than just the traditional basic numbers. But our understanding of football statistics since then have long since left this initial primitive and clumsy model behind and given rise to far more useful formulas such as AYPA. Why is passer rating still being cited?

[Bryan]

When are we switching to the metric system?

[Ronfitch]

Passer rating is a fundamentally flawed statistic that sets arbitrary standards and gives equal weight to unequal categories. It should be given credit, when first conceived, as an attempt to express an overall rating of a team's passing game that went beyond total yards, which was a watershed idea at the time in formulating new and better statistical models that gave us more effective measurements than just the traditional basic numbers. But our understanding of football statistics since then have long since left this initial primitive and clumsy model behind and given rise to far more useful formulas such as AYPA. Why is passer rating still being cited?

Years ago, I posted to the forum asking what is the most important single stat in determining the winner of a game. A few people cited Allen Berra's work in the '80s, which determined it was either AYPA or AYPC (I don't recall which and it was the old forum, so is now gone). My gut tells it it was AYPA, because that would possibly mean that there is success on a mixture of short, medium and long passing attempts.

But, I love the stat that Ralph posted earlier this week about the effect of a pass rush and a sack:

http://www.profootballresearchers.org/f ... f=5&t=5744

[Throwin_Samoan]

When are we switching to the metric system?

1980. Better start studying.

There probably IS a better way to do this now. (ESPN's Total QBR or whatever it is feels unnecessarily complicated, but I have no idea if it actually is.)

When the passer rating became the new metric in 1973, it was based on the performances of "all qualified passers" from 1960-1972. (And Len Dawson instantly became the league's all-time leading passer under the formula, with an 83.9 rating. Gardner Minshew was at 91.2 last year.)

Given how the passing game has changed over time and with completion percentages far higher and interception percentage half what it was then, it makes sense to at least re-calibrate the calculation. (I actually asked Steve Hirdt this several years ago: "Shouldn't there be an updated formula for this?" He didn't have an answer for me.)

The league average rating in 1973, the first year of the new calculation, was 64.9.

One of the benefits of the switchover at the time was that the prior system was good for ranking passers within the league for that year, but was not good for making year-to-year or era-to-era comparisons. Now that's out the window because if the list of all-time leaders is dominated by players from this era (with Young and Montana, at least, representing), seems like something is off.

[RyanChristiansen]

I see passer rating as a metric that tries to put a single number on how a QB performed in one game, perhaps to compare a QB to himself, and that it really can’t be used to compare QB vs. QB, especially when some QBs’ wild card might be their feet.

[Throwin_Samoan]

There's a reason it's called Passer Rating and not Quarterback Rating. For years, the disclaimer appeared in the NFL Record and Fact Books that it deliberately was NOT an attempt to rank quarterbacks (though, obviously, it's used to do that), because they could not incorporate things like "guts" and "determination" and "he called his own plays" and "all he does is win" and all the other intangible nonsense that people quantify in their heads but can't put an actual number on. It also can't distinguish between a Randall Cunningham, whose ability to run gave him another dimension, and a Tom Brady, whose inability to run does not.

It's purely about how you throw the ball, nothing else. Like any stat, the influences and unquantifiables don't invalidate it, but just demand that we all use it as just one piece of information, and not the entire determination.

[Fitzclarence]

The passer rating formula might be arbitrary, flawed and in need of updating, but if it is applied to all quarterbacks in a given season, it does a good job of separating the great, good and average performers.