passer rating

[John Grasso]

Why is the passer rating such a convoluted formula so that the best a player can have is 158.3?

Wouldn't it make more sense to divide the final result by 158.3 and multiply it by 100 (or divide by 1.583)
so that the maximum score would be 100? It would be a lot easier to compare results if it were
on a scale of 0-100 rather than the way it is now.

[Steviek]

Agree, but at the time it was instituted in 1973, the historical passing data suggested that achieving a 100 passer rating would be much more difficult to accomplish than it is in today's pass happy league. I believe it was a pretty rare occurrence in the 1970s - Ken Stabler and Bert Jones accomplished it, but not too many others.

In 1978 I believe the top rated passers in both the AFC and NFC had ratings in the 80s.

[NWebster]

Why is the passer rating such a convoluted formula so that the best a player can have is 158.3?

Wouldn't it make more sense to divide the final result by 158.3 and multiply it by 100 (or divide by 1.583)
so that the maximum score would be 100? It would be a lot easier to compare results if it were
on a scale of 0-100 rather than the way it is now.

The mathematical reason is the interception component of the equation. It subtracts a transformed interception percentage from 2.375. The resulting figure is then added to the comp %, YPA and TD% figures and divided by 6 to get the rating. This is set with a floor of 0, i.e. if your interception percent is 100% the pure math gets you negative value, but the floor is set at 0. By definition, for int the cap is at 2.375. All other catagories with no "natural caps" we're also capped at 2.375. So a player with a 100% TD% would mathematically get (1.0 × 20) which would be added to the other components and divided by 6, but on TD's alone they'd have 20 6ths or a 300 and change. Basically they figured if the best you can do on Int% is get 2.375, they figured the best you should be able to do in any category is 2.375. 2.375 ×4 = 9.5, divide that by 6 and you get 1.583333, or 158.

All this was math to solve for average being 66.6. The net effect is that a TD% above 11.87% doesn't increase your passer rating. Because average was 5% or .05 and that portion of the equation is TD% × 20. 2.375 (the cap) divided by 20 (reversing the math) is 11.87%. Similarly YPA stops adding value when it gets extremely high as well.

That's a really long winded way to say there are caps (2.375) and floors (0) for each subcomponent. The Int% has a natural cap in the equation which is enforced across all other components and the others have a natural floor at 0 which is enforced on INTs. For this to hold the range must be 0 to 158.

There's a test later.

[Mark L. Ford]

Why is the passer rating such a convoluted formula so that the best a player can have is 158.3?

Wouldn't it make more sense to divide the final result by 158.3 and multiply it by 100 (or divide by 1.583)
so that the maximum score would be 100? It would be a lot easier to compare results if it were
on a scale of 0-100 rather than the way it is now.

My theory is that some rogue algebra teacher came up with the formula. From what I can tell, if your adjusted figures for completion percentage, yards per attempt, touchdowns and interception avoidance add up to 9.498, they get divided by six and multiplied by 100 to get anywhere between zero and 158.3. Seems like they could have had a scale of 1 to 100 if they had just said "divide a+b+c+d by 9.498" instead of dividing it by six. You'd still have a weird number to deal with, but it wouldn't be in the final result.

[Saban1]

I think that the passer rating favors short passers over passers that throw more long passes. JMO.

[BD Sullivan]

Agree, but at the time it was instituted in 1973, the historical passing data suggested that achieving a 100 passer rating would be much more difficult to accomplish than it is in today's pass happy league. I believe it was a pretty rare occurrence in the 1970s - Ken Stabler and Bert Jones accomplished it, but not too many others.

Stabler and Jones were the only starters to achieve that mark between 1973-77--both in 1976.

Before 1973, the only QB's to start 10 or more games and maintain that number:

Milt Plum, 1960 (the standard for ridicule directed at this rating)
Bart Starr, 1966
Y.A. Tittle, 1963
Roger Staubach, 1971
Len Dawson, 1966

Others of note who had most of their team's starts:

*Otto Graham, 1946
Frank Filchock, 1939
Sammy Baugh, 1945
*Otto Graham, 1947
Bart Starr, 1968
*Frankie Albert, 1948
Charlie Conerly, 1959

*AAFC

[bachslunch]

Milt Plum, 1960 (the standard for ridicule directed at this rating)

Not familiar with this -- what was the issue? Just curious.

[BD Sullivan]

Milt Plum, 1960 (the standard for ridicule directed at this rating)

Not familiar with this -- what was the issue? Just curious.

Just the fact that Plum was never considered an elite QB, yet he somehow ranks at the top of the list for the greatest season ever (110.4)--in which the Browns finished at 8-3-1, 1.5 games behind the Eagles. Note that the other four in his group either won a title that year or reached the title game.

Here are Plum's numbers for that year: 151-250 (60.4%), 21 TD's, 5 INT

[bachslunch]

Milt Plum, 1960 (the standard for ridicule directed at this rating)

Not familiar with this -- what was the issue? Just curious.

Just the fact that Plum was never considered an elite QB, yet he somehow ranks at the top of the list for the greatest season ever (110.4)--in which the Browns finished at 8-3-1, 1.5 games behind the Eagles. Note that the other four in his group either won a title that year or reached the title game.

Here are Plum's numbers for that year: 151-250 (60.4%), 21 TD's, 5 INT

Thanks. Two thoughts:

1. it's entirely possible to have a non-elite player have a fluke one-off season accomplishment which they don't approach again, which looks like it happened in Plum's case. For baseball examples, see Roger Maris's 61 HR season in 1961 or Brady Anderson's 50 home run season in 1996. They're tremendous outliers compared to the rest of their careers, and neither player has any legit claim to elite status. Neither does Plum, of course.

2. was curious to see why the 1960s Browns underachieved. There's always the possibility of bad luck since their three losses were close games. But they seemed to have a large number of rookies and second year players starting for them, including Hickerson and Schafrath on their o-line, SE Rich Kreitling, half their d-line (Floyd Peters, Jim Marshall), and their entire secondary. That's a lot of inexperience to be relying on, and it might have cost them a few of those close games. So the problem may -- or may not -- have been Plum specifically.

[bachslunch]

Milt Plum, 1960 (the standard for ridicule directed at this rating)

Not familiar with this -- what was the issue? Just curious.

Just the fact that Plum was never considered an elite QB, yet he somehow ranks at the top of the list for the greatest season ever (110.4)--in which the Browns finished at 8-3-1, 1.5 games behind the Eagles. Note that the other four in his group either won a title that year or reached the title game.

Here are Plum's numbers for that year: 151-250 (60.4%), 21 TD's, 5 INT

Just checked the all-time single season leaderboard for this stat over at pro-football reference. According to them, Plum's 1960 was never the greatest season ever using this stat at any time. According to them, it was Otto Graham's 1946 season, listed as 112.1. Were you excluding the AAFC for some reason? Also, his 1960 season currently ranks 14th all time, with all except Graham's coming since 1989, so he's nowhere near being the record holder now.

http://www.pro-football-reference.com/l ... season.htm