Quarterback rating

[Pats Fan]

That's a pretty solid "ain't broke" post right there.

Unless one of yous arguing against QB rating can come up with a QB stat that is a better predictor of winning %, then this argument should pretty much be over.

We are all different people. Some people need stats to justify what they are saying. Me, I equate that with bureaucracy. If there is anything I like less than a bureaucrat, well I am not sure who that would be.

As a football fan, can you not just look at a player and see what you see?? The only use that stats are is that the season is more than one game. So, one game does not tell a story and we cannot follow every team, every player. So stats then become someone important.

But as for your team, stats mean nothing. Do I need you to tell me how good Brady is. No. Do you need me to tell me how much better you are with Romo. No.

Can't we just think outside the box and LOOK. That is about all you need to do.

X to every bureaucrat of the world.

 

[burmafrd]

trying to balance completion percentage while not rewarding dinkers should be a prime feature of any rating system. That is why I think 3rd down completions for 1st downs of over 5 or 7 yds should be part of the system.
That is where the best QBs shine; and where the worst look the worst.
Also why the system should also require a minimum number of passes to qualify- that is why I said 20 per game average.

 

[AbeBeta]

We are all different people. Some people need stats to justify what they are saying. Me, I equate that with bureaucracy. If there is anything I like less than a bureaucrat, well I am not sure who that would be.

As a football fan, can you not just look at a player and see what you see?? The only use that stats are is that the season is more than one game. So, one game does not tell a story and we cannot follow every team, every player. So stats then become someone important.

But as for your team, stats mean nothing. Do I need you to tell me how good Brady is. No. Do you need me to tell me how much better you are with Romo. No.

Can't we just think outside the box and LOOK. That is about all you need to do.

X to every bureaucrat of the world.

Many relationship are readily detectable to the naked eye -- also, many relationships that we can "see" fall within the bounds of what we might observe in a random process.

Simply rejecting statistics as bureaucracy is silly. Bureaucrats follow stupid and pointless rules, whereas properly applied statistics tell you something meaningful.

Frankly, in my experience the only people who reject statistics are those who don't have a full understanding of statistics.

 

[Pats Fan]

Many relationship are readily detectable to the naked eye -- also, many relationships that we can "see" fall within the bounds of what we might observe in a random process.

Simply rejecting statistics as bureaucracy is silly. Bureaucrats follow stupid and pointless rules, whereas properly applied statistics tell you something meaningful.

Frankly, in my experience the only people who reject statistics are those who don't have a full understanding of statistics.

Oh, I understand them. I just don't need them. Put a guy on the field and I can see very clearly how he plays. I don't need a stat to tell me anything.

Some players players play for OK teams, but you can still see how they play the game.

I personally do not need any stat to tell me how a player is playing. All I need to do is watch. That tells me everything I need to know.

 

[AbeBeta]

Oh, I understand them. I just don't need them. Put a guy on the field and I can see very clearly how he plays. I don't need a stat to tell me anything.

Some players players play for OK teams, but you can still see how they play the game.

I personally do not need any stat to tell me how a player is playing. All I need to do is watch. That tells me everything I need to know.

So you are going to watch every single game, every team, every season?

 

[kmd24]

The AER contains stats for the entire team, so it only stands to reason that it'd be a better indicator of success than any single stat for any single position.

You're mixing up terminology. The AER is composed of data for the entire team. It is a single statistic.

The passer rating is composed of data for a single QB (which is also typically all the passing stats for a single team). It is also a single statistic.

Anyway, the fact that the passer rating correlates well with winning should be no surprise. It comprises TD's and INT's, among other things, so it will be significantly correlated with points scored and turnover margin. Points scored and turnover margin are also strong indicators of wins and losses.

As a final point, the QB rating should discern between passers, not teams. The fact that it correlates well to winning is only coincidental if you consider the history of its design.

 

[THUMPER]

I have my own QB Rating system that I have used for years and it is much more accurate (IMO) than the one the NFL uses in determining a QB's relative performance.

The formula is as follows:

((Completion%/10)+(Avg per att*1.5)+(Avg per comp/2)+((TD%-INT%)*2)-(Sack%/2)+((Won-Lost)*1.5))*1.8

My top ten come out as:

T Brady - 105.11
P Manning - 71.91
T Romo - 65.52
Ben R - 58.36
J Garcia - 56.55
D Anderson - 53.97
D Garrard - 53.48
B Favre - 51.22
E Manning - 50.69
K Warner - 50.32

I know most will complain about the numbers which seem to imply that Brady is twice as good as Favre but don't let the range scare you. What is important is not the range but the ranking itself. Guys like Brady, who is having an incredible year, will throw the scale out of whack but most of the others are within proper range.

It is heavily weighted towards winning games which is a QB's primary job IMO. So Brady & Manning are at the top having won all of their games while Bulger, Harrington, and the rest who have yet to win are at the bottom. A QB's job is to win games, not simply to throw the ball.

They are QB ratings, not simply passer ratings. I tried including rushing stats but there is too much variation between QBs and it wasn't worth the trouble.

I can e-mail the spreadsheet to anyone who wants it and you can play around with the formula to make it more to your liking. Just PM me.

 

[burmafrd]

I would argue that winning games should not be a factor. Look at just over the last few years how many of our games were lost due to factors BEYOND the QBs control. Blocked FGs, etc. While a QB is usally the biggest single factor in how a team does, he is by far not the only factor.

 

[AbeBeta]

As a final point, the QB rating should discern between passers, not teams. The fact that it correlates well to winning is only coincidental if you consider the history of its design.

So having a good completion % combined with more TDs than picks is only coincidentally related to winning percentage?

Fascinating.

 

[Pats Fan]

So you are going to watch every single game, every team, every season?

Well, I think I already addressed this point. Stats are important for players you don't see. But, but, but, that tells you maybe 50% of what you need to know.

We don't have the time, but the coaches of the league do. That is their job.

Me, I cannot stand stats. I USED to be one of "them". But then many years ago I was forced to take a philosophy course. I didn't like it one bit. But then, finally I started listening. It opened up my world.

The world is not black or white. It is gray. Being hard nosed on either side of the fence you choose will lead to a life of stress.

Just look. See and think.

 

[AbeBeta]

The world is not black or white. It is gray.

Anyone who thinks that stats are black or white doesn't understand them well enough to be using them.

 

[theogt]

You're mixing up terminology. The AER is composed of data for the entire team. It is a single statistic.

The passer rating is composed of data for a single QB (which is also typically all the passing stats for a single team). It is also a single statistic.

Of course, I never said that AER wasn't a single stat.

 

[AdamJT13]

Oh, I understand them. I just don't need them. Put a guy on the field and I can see very clearly how he plays. I don't need a stat to tell me anything.

Some players players play for OK teams, but you can still see how they play the game.

I personally do not need any stat to tell me how a player is playing. All I need to do is watch. That tells me everything I need to know.

Is it not useful to keep track of how often you see a player play well and how often you see him play poorly? Or which players you see play well and which players you see play poorly?

Or would you just trust your memory to know which player has played better after watching each of them for, say, 16 games?

 

[Pats Fan]

My last thoughts on this.

Moss is a cancer. He has lost a step and can't play anymore. Stats, Oakland.

Welker is just a midget, who in their right mind would give up a second round draft pick for this nobody??

Stallworth, fast, yeh, but always injured and really was never any good anyway.

On and on. Stats play a role, but it only tells you a small amount about what you need to know.

My world will always be gray and open to something more than stats.

Stats are an indicator, not what is.

 

[sonnyboy]

I've brought this up before. After the Giants game in NY when people/media insisted Romo had a bad game.

Any formula will be flawed. I Cant take the surrounding talent into the equation.

However the current formula can easily be improved.

In the current formula you credit for both COMP% and YPA. Its stupid. YPA is all that matters. You have to QB's with YPA of 8.0 and the guy with a 65% COMP gets a higher rating than a QB with 60%COMP.

QB's should also get credit for all those FG drives.

Formula should look something like this:

(50%YPA) (12.5%Passing points) (12.5%Total points) (25%INT)

Something like this. It gives credit for all points and double credit for the points scored actually passing. It also takes comp% out and gives a lot of weight to the actual yards a team gains every time they pass.

 

[Yakuza Rich]

As a semi-football stat freak I'm not a big fan of the QB rating system, but it does have some merits. In particular, the correlation coefficient of QB rating Margin (Team's QB Rating - Team's QB Rating Allowed) to wins is extremely high. In other words, good or bad those teams with the higher QB Rating *Margin* typically win more games. I think they should try and come up with a formula that accounts for sack yards lost and fumbles lost as well. But generally when I look at QB's I look at YPA. The higher that is, generally the better the QB is legitimately playing.





YAKUZA

 

[percyhoward]

Part of the problem here is that many times, if the QB leads his WR correctly, he'll get more YAC. So you're negating positive yardage that are at least partially the result of the QB.

That sounds good as a theory, but doesn't stand up to analysis.

Going on the assumption that the QB's with the most YAC are the ones who are the best at leading the receiver correctly, then the most accurate passers should be the ones with the highest YAC. If that's the case, then Peyton Manning, Carson Palmer, and Tony Romo are among the least accurate passers in the game.

This is from last year:
YAC per attempt
McNabb 4.4
Brees 3.8
Rberger 3.3
Bulger 3.3
Kitna 3.1
Losman 3.1
Rivers 3.1
Romo 3.1
Palmer 2.9
PManning 2.7

YAC doesn't have near as much to do with accuracy as it has to do with completions to running backs. McNabb and Brees don't get their YAC from the way they precisely place the ball--they get it from throwing a ton of passes to Westbrook and Bush.

So if you eliminated YAC from passer rating, you'd have to concentrate on WR and TE YAC, or only include YAC on passes thrown more than 10 yards, etc.

 

[theogt]

That sounds good as a theory, but doesn't stand up to analysis.

Going on the assumption that the QB's with the most YAC are the ones who are the best at leading the receiver correctly, then the most accurate passers should be the ones with the highest YAC. If that's the case, then Peyton Manning, Carson Palmer, and Tony Romo are among the least accurate passers in the game.

This is from last year:
YAC per attempt
McNabb 4.4
Brees 3.8
Rberger 3.3
Bulger 3.3
Kitna 3.1
Losman 3.1
Rivers 3.1
Romo 3.1
Palmer 2.9
PManning 2.7

YAC doesn't have near as much to do with accuracy as it has to do with completions to running backs. McNabb and Brees don't get their YAC from the way they precisely place the ball--they get it from throwing a ton of passes to Westbrook and Bush.

So if you eliminated YAC from passer rating, you'd have to concentrate on WR and TE YAC, or only include YAC on passes thrown more than 10 yards, etc.

Where in this analysis was my point negated?

 

[Zaxor]

if a WR was suppose to cross the CB face but did not resulting in an INT it goes on a QB's stat... or

if a WR falls down
runs the wrong route
tips the ball

they all can end up as an int...

which goes on the QBR... in order to have an accurate read on the passer you have to be able to determine what the play was and who should get credit or fault...I just don't believe that is entirely possible without the coaches help and no coach in the NFL wants to throw his WR/TE/RB or QB under the bus...

so any system designed is going to have inaccuracy

 

[THUMPER]

if a WR was suppose to cross the CB face but did not resulting in an INT it goes on a QB's stat... or

if a WR falls down
runs the wrong route
tips the ball

they all can end up as an int...

which goes on the QBR... in order to have an accurate read on the passer you have to be able to determine what the play was and who should get credit or fault...I just don't believe that is entirely possible without the coaches help and no coach in the NFL wants to throw his WR/TE/RB or QB under the bus...

so any system designed is going to have inaccuracy

The same is true about TD passes:

The CB may have fallen down
The WR broke a tackle
The WR outjumps the DB
The ball is tipped

Aikman was penalized in the passer ratings because we handed the ball off to Emmitt when we were inside the 10 yard line so he didn't get as many passing TDs as some other QBs... like Steve Young.

That's why I believe wins and losses should count for the QB as well as sacks and fumbles (although I didn't include fumbles in my rating system I plan on adding it soon). Overall, the QB has more to do with winning or losing the game than any other player although there are always individual exceptions.