Ravens at Bengals      Pro Bowl Voting      #RavensPumpkin      Game Rewind   

Jump to content


Photo

Flacco is Underrated because Completion Percentage is Overrated


  • Please log in to reply
26 replies to this topic

#1 callahan09

callahan09

    Pro Bowler

  • Members
  • PipPipPipPip
  • 2,156 posts
  • Gender:Male
  • Location:New Jersey

Posted 05 February 2013 - 02:07 PM

Has anybody else ever thought about how the completion percentage factor in the QB Rating formula is a little bit over-weight? It's actually counted twice, because there is a completion percentage rule as well as a yards per attempt rule. Yards per attempt includes completion percentage inherently, as another formula for yards per attempt could be written as Yards Per Completion * Completion Percentage.

Example: 10/25 for 250 yards is 10 yards per attempt = 25 yards per completion * 0.40 completion percentage.

Wouldn't it be weighted more accurately to use yards per completion instead of yards per attempt as the second rule, since using yards per attempt instead means you're essentially counting completion percentage as twice as important as each of the other 3 rules (yards per completion, touchdown percentage, and interception percentage)?

Here is a made-up example to illustrate the effect that completion percentage has on passer rating:

QB1: 18/35, 303 Yards, 5 TD, 0 INT = 120.6
QB2: 23/35, 296 Yards, 3 TD, 0 INT = 120.7

Out of the same number of attempts, QB2 has fewer yards and 2 fewer TDs, but because they've got a couple of extra completions they have a higher rating? Is it my imagination, or wouldn't you rather have QB1's stats instead of QB2's? I'm just saying, I think the QB Rating formula could be tweaked a bit to diminish the weight of completion percentage.

Now, let me show you a real-life example how this effects actual quarterbacks:

QB1: 481/761, 5679 Yards, 32 TD, 21 INT = 88.4 QB Rating
QB2: 422/761, 5451 Yards, 39 TD, 16 INT = 86.5 QB Rating

In the same number of attempts, QB1 has just 228 additional yards. That's about 0.299 yards per attempt higher, but it's 1.110 yards per completion lower. Additionally, QB2 has a +23 touchdown to interception differential (2.44 to 1 ratio), while QB1 has a +11 differential (1.52 to 1 ratio). QB2 threw 7 more touchdowns and 5 fewer interceptions. QB2 threw a touchdown every 19.5 throws, and an interception every 47.6 throws, compared to QB1 throwing a touchdown every 23.8 throws and an inteception every 36.2 throws.

Would anybody on earth take QB1 over QB2? It seems glaringly obvious which one performed better. And yet due to the over-weighting of completion percentage in the formula, QB Rating says QB1 was better.

In case you were wondering, QB1 represents Peyton Manning's actual career post-season statistics, and QB2 represents Joe Flacco's career post-season statistics extrapolated out to the same number of attempts as Peyton Manning.
  • 1

#2 flynismo

flynismo

    Hall of Famer

  • Members
  • PipPipPipPipPipPip
  • 15,798 posts
  • Gender:Male
  • Location:At work

Posted 05 February 2013 - 02:14 PM

Wouldn't it be weighted more accurately to use yards per completion instead of yards per attempt as the second rule, since using yards per attempt instead means you're essentially counting completion percentage as twice as important as each of the other 3 rules (yards per completion, touchdown percentage, and interception percentage)?


IMO, no, because then you are basically saying "on the occassions that so-and-so does happen to connect on his passes, this is what he does".
An example could be that a QB goes 2 for 10 for 100 yards. Sure, 50 YPC is obviously amazing, but what does that really mean?

But, you have brought up another great example of how flawed QBR is, and how it is virtually impossible to fix one thing without breaking another.

Edited by flynismo, 05 February 2013 - 02:15 PM.

  • 0
Self-appointed President of The Flacco Lovers Foundation (F.L.F)Joe Flacco, Super Bowl 47 MVPFarewell to a legend. Thanks for the memories, Ray Ray!If you woke up this morning, congratulations. You got another chance

#3 The Raven

The Raven

    Hall of Famer

  • Members
  • PipPipPipPipPipPip
  • 12,175 posts
  • Gender:Male

Posted 05 February 2013 - 02:14 PM

Sorry, no. If anything, completion percentage is a very underrated stat. It may be over weighted in the QB rating formula, but I don't buy into either QB rating stat.

You can't just look at any one stat when looking at QBs. I don't believe in looking at the QB rating. I care about four stats in a quarterback. Completion percentage, yards per attempt, touchdowns, and interceptions. That's all that matters. If you give me a formula that takes a balanced look at these stats, that'd be great.

Low completion percentages stall drives. Fact.

Edited by The Raven, 05 February 2013 - 02:14 PM.

  • 0

#4 flynismo

flynismo

    Hall of Famer

  • Members
  • PipPipPipPipPipPip
  • 15,798 posts
  • Gender:Male
  • Location:At work

Posted 05 February 2013 - 02:19 PM

I care about four stats in a quarterback. Completion percentage, yards per attempt, touchdowns, and interceptions.


Agreed, I think that the only stats that truly mean anything are those that measure effeciency, because effeciency is not a stat that will be distorted by one or two huge games; it is something that will remain pretty consistent over a long period of time.
  • 0
Self-appointed President of The Flacco Lovers Foundation (F.L.F)Joe Flacco, Super Bowl 47 MVPFarewell to a legend. Thanks for the memories, Ray Ray!If you woke up this morning, congratulations. You got another chance

#5 callahan09

callahan09

    Pro Bowler

  • Members
  • PipPipPipPip
  • 2,156 posts
  • Gender:Male
  • Location:New Jersey

Posted 05 February 2013 - 02:26 PM

Sorry, no. If anything, completion percentage is a very underrated stat. It may be over weighted in the QB rating formula, but I don't buy into either QB rating stat.

You can't just look at any one stat when looking at QBs. I don't believe in looking at the QB rating. I care about four stats in a quarterback. Completion percentage, yards per attempt, touchdowns, and interceptions. That's all that matters. If you give me a formula that takes a balanced look at these stats, that'd be great.

Low completion percentages stall drives. Fact.


My whole point was that completion percentage is effectively counted twice, because it is a factor in yards per attempt. That's why my proposal is that maybe it'd be better to use yards per completion instead of yards per attempt. Do you disagree that Flacco's extrapolated stats deserve a higher rating than Peyton Manning's in my example above?
  • 0

#6 flynismo

flynismo

    Hall of Famer

  • Members
  • PipPipPipPipPipPip
  • 15,798 posts
  • Gender:Male
  • Location:At work

Posted 05 February 2013 - 02:28 PM

Do you disagree that Flacco's extrapolated stats deserve a higher rating than Peyton Manning's in my example above?


I agree that in your example, Flacco outplayed Manning (AGAIN, teehee).

But those stats also are far from the whole picture -- but nearly all stats usually are.

To expand on that, look at it like this.

How was their performance in situational play? I don't care if Flacco overthrew Torrey on an attempt fifty yards downfield, because those plays usually occur on first or second down and are not going to kill the drive.
But how did Flacco do on third down? How did he play in the red zone? And so on

Edited by flynismo, 05 February 2013 - 02:33 PM.

  • 0
Self-appointed President of The Flacco Lovers Foundation (F.L.F)Joe Flacco, Super Bowl 47 MVPFarewell to a legend. Thanks for the memories, Ray Ray!If you woke up this morning, congratulations. You got another chance




#7 The Raven

The Raven

    Hall of Famer

  • Members
  • PipPipPipPipPipPip
  • 12,175 posts
  • Gender:Male

Posted 05 February 2013 - 02:34 PM

My whole point was that completion percentage is effectively counted twice, because it is a factor in yards per attempt. That's why my proposal is that maybe it'd be better to use yards per completion instead of yards per attempt. Do you disagree that Flacco's extrapolated stats deserve a higher rating than Peyton Manning's in my example above?


I guess I'd like that. If QB rating factored in yards per completion Joe would have the highest QB rating in history, lol.

Edited by The Raven, 05 February 2013 - 02:34 PM.

  • 0

#8 RavensFanInCA

RavensFanInCA

    Pro Bowler

  • Members
  • PipPipPipPip
  • 2,420 posts

Posted 05 February 2013 - 02:35 PM

I don't think completion percentage is completely overrated. If you have someone who can be hot and cold at any time, it can hurt you. Imagine if you play cold against a bad team and hot against a good team. What if that ended up being just too little to win either? I'd rather have a player that played "warm" for both games if it means I won one of the two.

Plus, you could say that you make those big plays to get close to the end zone, but if you can't punch it in, it's worth nothing. You need consistency to make sure in those 4 chances inside the red zone, you get your TD.
  • 0

#9 RavensAllTheWay

RavensAllTheWay

    MVP

  • Members
  • PipPipPipPipPip
  • 4,078 posts
  • Gender:Male

Posted 05 February 2013 - 02:37 PM

No one complained about Joes 52.2 and 52.9 percentages against Indy and Denver, and for good reason. He averaged 12.26 and 9.74 yards/attempt with 5 TDs and no picks. If you can make your completions count, no one's going to complain.

I think completion percentage may be a tad overrated. Phil Simms continually says when he calls our games that we don't have nearly enough easy throws and drive starters, especially compared to other offenses around the league. Not to mention, I'm sure Joe and the offense take way more deeps shots than the average team.
  • 0

#10 Gordo52

Gordo52

    52 G.O.A.T.

  • Members
  • PipPipPipPipPipPip
  • 18,444 posts
  • Gender:Male
  • Location:In Your Mind

Posted 05 February 2013 - 02:38 PM

Lol so Joe isn't in the elite section....he's in the young rising stars. (Flacco,Luck,RG3,Wilson,Kap) No Matt Ryan on that list either... Tim Hassleback chose everyone over Flacco besides RG3.

His argument was based around QBR.


No words.
  • 1
2dc45a6e-0473-4d16-b247-f9f9608f28d3_zps

#11 callahan09

callahan09

    Pro Bowler

  • Members
  • PipPipPipPip
  • 2,156 posts
  • Gender:Male
  • Location:New Jersey

Posted 05 February 2013 - 02:39 PM

IMO, no, because then you are basically saying "on the occassions that so-and-so does happen to connect on his passes, this is what he does".
An example could be that a QB goes 2 for 10 for 100 yards. Sure, 50 YPC is obviously amazing, but what does that really mean?

But, you have brought up another great example of how flawed QBR is, and how it is virtually impossible to fix one thing without breaking another.


Out of curiosity, in your example, is it 2/10 for 100 yards, 0 TD, 0 INT, or what? If so, then you're looking at a 68.8 QB Rating. Interestingly enough, that's the exact same passer rating you'd get if you went 8/10 for 100 yards, 0 TD, 2 INT. So you've thrown the ball the same number of times, for the same amount of yards, but in one case you didn't turn the ball over, and in the other case you turned it over both times you didn't complete the pass. Same rating? I dunno about you, but I'd rather my quarterback throw for 100 yards and no turnovers, instead of 100 and 2 turnovers, on the same number of attempts, regardless of completion percentage.

I mean, if you break it down into each QB having two drives, as an example, then you could have something like this...

QB1's Drives:

1st Down: Incomplete
2nd Down: 50 Yard completion
1st Down: Incomplete
2nd Down: Incomplete
3rd Down: Incomplete
4th Down: Field Goal

(x2) = 6 Points

QB2's Drives:

1st Down: 25 yards
1st Down: 5 yards
2nd Down: 10 yards
1st Down: 10 yards
1st Down: Interception

(x2) = 0 Points, and probably additional points for the other team

It's obvious to me when you look at it like that that I'd rather have QB1 do what he does than QB2 moving the ball for a while then turning it over. But QB Rating is telling me that these performances are equal, which I disagree with.

For what it's worth, I just re-tooled my QB Rating calculator to utilize yards per completion instead of yards per attempt and the result is that QB1 goes up to a 79.2 rating and QB2 goes down to a 64.2. Personally, I think those numbers better reflect the quality of those stat lines.

And for what it's worth, Peyton Manning's line in my initial post would change to a 91.5, and Flacco's would go up to a 95.4, which I also think reflects the stat lines better.

Edited by callahan09, 05 February 2013 - 02:58 PM.

  • 0




#12 Ravensfan23

Ravensfan23

    Hall of Famer

  • Moderators
  • PipPipPipPipPipPip
  • 6,995 posts
  • Gender:Male
  • Location:Baltimore

Posted 05 February 2013 - 02:41 PM

Agreed, I think that the only stats that truly mean anything are those that measure effeciency, because effeciency is not a stat that will be distorted by one or two huge games; it is something that will remain pretty consistent over a long period of time.


Agreed but I think there should be a better way of determining what's effecient. If a guy drops back 10 times but only completes 4 passes, that's not every efficient. However what if out of thos 6 incompletions, 2 were dropped passes, 2 were throw aways to avoid quick pressure and 2 were because of great defense, is that QB not efficient? All his passes were accurate and now of the incompletions were a result of his poor play, yet the numbers will show he's not having a good day.

I look at ESPN's QBR for Kaep in the Super Bowl at 46.1 or something and I laugh. Yes he had his issues, but that team doesn't have a shot without the plays that guy made and the adjustments he forced the Ravens to make. Their QBR suggests that Kaep played terrible and that's not close to being true.

So I agree with you, but I think that even measuring a QBs efficiency will become flawed the closer you look at it. So i guess you just that the bad with the good.
  • 0

mj5v69.jpg


#13 flynismo

flynismo

    Hall of Famer

  • Members
  • PipPipPipPipPipPip
  • 15,798 posts
  • Gender:Male
  • Location:At work

Posted 05 February 2013 - 02:51 PM

Agreed but I think there should be a better way of determining what's effecient. If a guy drops back 10 times but only completes 4 passes, that's not every efficient. However what if out of thos 6 incompletions, 2 were dropped passes, 2 were throw aways to avoid quick pressure and 2 were because of great defense, is that QB not efficient? All his passes were accurate and now of the incompletions were a result of his poor play, yet the numbers will show he's not having a good day.

I look at ESPN's QBR for Kaep in the Super Bowl at 46.1 or something and I laugh. Yes he had his issues, but that team doesn't have a shot without the plays that guy made and the adjustments he forced the Ravens to make. Their QBR suggests that Kaep played terrible and that's not close to being true.

So I agree with you, but I think that even measuring a QBs efficiency will become flawed the closer you look at it. So i guess you just that the bad with the good.


This is exactly why Ray himself says that stats are for losers. You evaluate with your eyes, not stats. It's essentially the reason behind the Flacco is/is not elite wars over the last couple years.

But that leads me to my response to this thread:

Flacco isn't underrated due to completion % being overrated, he is underrated because people are too concerned with a pretty stat line than what takes place on the field.

The most memorable plays of the SB to me was:

1. Flacco nearly getting sacked, but dumping off to Leach at the last possible instant. That play resulted in negative three yards passing, but salvaged four or five yards by avoiding the sack.

2. Flacco audibling out of the run on third down and making that short completion to Boldin.

3. Flacco running like a mad man, then running to the sieline, and made an insane 18 yard completion to Boldin. I thought for certain Flacco was going to throw it away that play, but instead bought time with his feet and made a huge first down.

And those first two plays actually reflect poorly on Flacco's stat sheet, even though they were awesome plays in reality!

Edited by flynismo, 05 February 2013 - 03:01 PM.

  • 3
Self-appointed President of The Flacco Lovers Foundation (F.L.F)Joe Flacco, Super Bowl 47 MVPFarewell to a legend. Thanks for the memories, Ray Ray!If you woke up this morning, congratulations. You got another chance

#14 gabefergy

gabefergy

    Hall of Famer

  • Members
  • PipPipPipPipPipPip
  • 15,788 posts
  • Gender:Male
  • Location:West Side

Posted 05 February 2013 - 02:59 PM

I agree that in your example, Flacco outplayed Manning (AGAIN, teehee).

But those stats also are far from the whole picture -- but nearly all stats usually are.

To expand on that, look at it like this.

How was their performance in situational play? I don't care if Flacco overthrew Torrey on an attempt fifty yards downfield, because those plays usually occur on first or second down and are not going to kill the drive.
But how did Flacco do on third down? How did he play in the red zone? And so on


I think ESPN tried to do this with their bogus "QBR" but they inserted way too much variability and subjectivity into the equation which makes it complete crap.

However some of the concepts are good, like considering down and distance, drops, penalty yards, and rushing yards.

I think there should be an actual formula that incorporates traditional stats like YPA, comp %, TDs, and INTs along with EPA and WPA. That would be the best IMO.
  • 0

@gabefergy


#15 Dline

Dline

    Practice Squad

  • New Members
  • Pip
  • 3 posts

Posted 05 February 2013 - 03:04 PM

Since Jim Caldwell took over as the OC Flacco has 15 TD's to 1 INT. Flacco is still a young quaterback and he needed a good OC to help take his potential to the next level and Cam I think not on purpose was a wall to Joe's development.

Edited by Dline, 05 February 2013 - 03:08 PM.

  • 0

#16 flynismo

flynismo

    Hall of Famer

  • Members
  • PipPipPipPipPipPip
  • 15,798 posts
  • Gender:Male
  • Location:At work

Posted 05 February 2013 - 03:08 PM

I think ESPN tried to do this with their bogus "QBR" but they inserted way too much variability and subjectivity into the equation which makes it complete crap.

However some of the concepts are good, like considering down and distance, drops, penalty yards, and rushing yards.

I think there should be an actual formula that incorporates traditional stats like YPA, comp %, TDs, and INTs along with EPA and WPA. That would be the best IMO.


Theoretically, it would be ideal, but really, no matter what formula we come up with, it will never tell the whole tale.
But the main issue in your idea would be determining the gravity of each stat (example, how much weight do we give completion % on third down? What if that completion on third down was not enough to move the chains? How many TD's erase one INT? Did that INT even matter in the context of the game? ETC...)
  • 0
Self-appointed President of The Flacco Lovers Foundation (F.L.F)Joe Flacco, Super Bowl 47 MVPFarewell to a legend. Thanks for the memories, Ray Ray!If you woke up this morning, congratulations. You got another chance




#17 Ravensfan23

Ravensfan23

    Hall of Famer

  • Moderators
  • PipPipPipPipPipPip
  • 6,995 posts
  • Gender:Male
  • Location:Baltimore

Posted 05 February 2013 - 03:44 PM

This is exactly why Ray himself says that stats are for losers. You evaluate with your eyes, not stats. It's essentially the reason behind the Flacco is/is not elite wars over the last couple years.

But that leads me to my response to this thread:

Flacco isn't underrated due to completion % being overrated, he is underrated because people are too concerned with a pretty stat line than what takes place on the field.

The most memorable plays of the SB to me was:

1. Flacco nearly getting sacked, but dumping off to Leach at the last possible instant. That play resulted in negative three yards passing, but salvaged four or five yards by avoiding the sack.

2. Flacco audibling out of the run on third down and making that short completion to Boldin.

3. Flacco running like a mad man, then running to the sieline, and made an insane 18 yard completion to Boldin. I thought for certain Flacco was going to throw it away that play, but instead bought time with his feet and made a huge first down.

And those first two plays actually reflect poorly on Flacco's stat sheet, even though they were awesome plays in reality!


Couldn't agree more. That pass to Leach was great and everyone at my party thought I was crazy for cheering it. Until I explained to them exactly what you just said.

The scramble pass to Boldin almost reminded me of the scramble pass against the Steelers in 08' he completed to Mason, don't know if you remember that play or not. Anyway, that single play it what separates Flacco apart from the Brady, Breesm and Manning crowd imo. Because Flacco was not only on the run, but he had not time to set up. It looked like Flacco just threw a jump ball, but in reality, he threw the ball to a spot and led Boldin to it.

Most QBs wouldn't have attempted and if they didm the ball would have fluttered in the air. Joe's arm strength and Brett Favre type boldness to attempt any throw on the field is special.
  • 0

mj5v69.jpg


#18 A Fish Called Yanda

A Fish Called Yanda

    Hall of Famer

  • Members
  • PipPipPipPipPipPip
  • 8,252 posts
  • Gender:Male

Posted 05 February 2013 - 03:56 PM

I think we should just take out Yards after Catch - because that's the receiver's stat, not the QB's.

Or just get rid of QBR altogether. That's far simpler.
  • 0

#19 gabefergy

gabefergy

    Hall of Famer

  • Members
  • PipPipPipPipPipPip
  • 15,788 posts
  • Gender:Male
  • Location:West Side

Posted 05 February 2013 - 03:59 PM

Theoretically, it would be ideal, but really, no matter what formula we come up with, it will never tell the whole tale.
But the main issue in your idea would be determining the gravity of each stat (example, how much weight do we give completion % on third down? What if that completion on third down was not enough to move the chains? How many TD's erase one INT? Did that INT even matter in the context of the game? ETC...)


That's for smarter people than me to figure out ;)

Edited by gabefergy, 05 February 2013 - 03:59 PM.

  • 0

@gabefergy


#20 Ed_Reed20

Ed_Reed20

    Crenshaw Pete

  • Members
  • PipPipPipPipPipPip
  • 24,748 posts
  • Gender:Male
  • Location:UK

Posted 05 February 2013 - 04:11 PM

I think we should just take out Yards after Catch - because that's the receiver's stat, not the QB's.

Or just get rid of QBR altogether. That's far simpler.


But there are numerous situations where a QB's throw allows the WR to pick up YAC.

For the simple fact that QBs don't play in a vacuum, I don't think there will ever be a perfect formula to gauge their play.
  • 0
"...when Sean Taylor is in the neighborhood, you will feel him." - Al Michaels

Posted Image

ER20 takes off and then gets distracted by a debate going on...





0 user(s) are reading this topic

0 members, 0 guests, 0 anonymous users