Advanced stats and our QBs, 2014 edition

[I Love The Skins]

And yet he threw an interception 33% of the time he threw it downfield tonight.

Wow! do you ever say anything good about RG3?

[TheCoach22]

Stats tell a lot, but they can also be misleading. Cousins will have fewer sacks based on his style and the gameplan... Robert will take more sacks, as will McCoy, as mobile guys looking to extend plays. So Cousins taking fewer sacks but throwing more picks does not indicate the line was fine in those games as at least one poster above claimed.

Also, TD passes in isolation, as another pro-Cousins guys mentioned are not good indicators of QB play. Yesterday, Young has two 1-yard runs for TDs, but the QB play was largely responsible for being in that position.

Personally, I watch Romo get 5+ seconds to throw the ball and I cannot help but think that all of our guys are lucky when they get half-that. People post RGIII will get hurt with his style of play, but he is far more likely to get hurt in a pocket collapsing with bodies flying at his knees. Look at yesterday's one sack where Chester got overpowered in .5 seconds and Robert barely had time to fake the handoff on the RO. I would prefer to see this team build a beast line and allow The best playmaker (RGIII) to flourish behind them. We presently have the worst line in all of football, we need three new starters next year, and neither Jay nor Robert can truly be judged until the fat boys upfront more resemble an OL that can actually make a pocket.

[Tsailand]

QB rating is not a useful stat on a per game basis because the sample size is too low

 

All ratings systems have the sample size problem if you look at just one game.  It doesn't make them worthless, just less reliable on a per-game basis.

 

and biased by the opponent's defensive strategies and capabilities.

 

Most ratings systems have this problem as well.  DVOA doesn't, but I don't think they put out per-game ratings.

 

The actual problems with passer rating have already been posted in this thread, but I'll repeat them for you:

- Rewards completions for no gain

- Rewards TDs too much

- Doesn't penalize ints enough Penalizes ints too much

- Doesn't penalize sacks at all

 

QBR is useless in all situations because it's just a bunch of biased ESPN idiots assigning numbers to their opinions.

 

No, it really isn't.  The formula is secret, but we know that it's based on advanced research into how QB play affects a team's chance of winning.  You can read about it here: http://espn.go.com/nfl/story/_/id/6833215/explaining-statistics-total-quarterback-rating

[redskins59]

Without those two bombs, RG3 threw for close to a 100 yards.  Any advanced stat out there that accounts for outliers? You can't expect RG3 to connect on those long bombs every game. So in all honesty, his other plays did not impress me.  Not to mention that Jackson could have scored a TD or two, if the long bombs were in stride.

[Burgold]

Without those two bombs, RG3 threw for close to a 100 yards.  Any advanced stat out there that accounts for outliers? You can't expect RG3 to connect on those long bombs every game. So in all honesty, his other plays did not impress me.  Not to mention that Jackson could have scored a TD or two, if the long bombs were in stride.

Sure and without Brady's touchdown passes he's never thrown for a touchdown. Without a QB, Calvin Johnson has never caught a ball. Without a football, Barry Sanders never gained a yard.

 

 Stupid. Stupid.  Stupid.

[greenspandan]

No, it really isn't. The formula is secret, but we know that it's based on advanced research into how QB play affects a team's chance of winning. You can read about it here: http://espn.go.com/nfl/story/_/id/6833215/explaining-statistics-total-quarterback-rating

LOL "it's subjective and sciencey and smart! Just trust us!"

Yeah effing right. Give me the real numbers, not the numbers filtered through a black box about which all we know for sure is that it involves human judgement.

Sorry you wasted your time typing a bunch of stuff I was already aware of.

[Destino]

Without those two bombs, RG3 threw for close to a 100 yards. Any advanced stat out there that accounts for outliers? You can't expect RG3 to connect on those long bombs every game. So in all honesty, his other plays did not impress me. Not to mention that Jackson could have scored a TD or two, if the long bombs were in stride.

The standards are different for RG3. When Colt under throws Jackson (even though he caught the ball) fans said he left touchdowns on the field because of his weak arm. When RG3 does the same, we hear about how Griffin is the future and no other QB could have won the game.

It's always been that way. Griffin gets excuses but the other QBs with less reps and less practice get blame.

Griffin had an average game. He didn't light the world on fire, and the defense held the eagles to three after his late game interception. The defense stood tall at the end and Garcon made a huge play at the right time that got an additional 15 added to it because of roughing the passer.

[perez24]

I looked at ANY/A lifetime stats to see how they hold up over a career.  Trent Green rates ahead of Brett Favre (who is comparable to Daunte Culpepper in ANY/A), Dan Fouts, John Elway and a number of other Hall of Famers.  I'm not really sure how relevant/accurate that rating system is. 

 

Here's the link:  http://www.pro-football-reference.com/leaders/pass_adj_net_yds_per_att_career.htm

[HAILSKINSNYC]

RG3 put the team in position to score on 6 out of 9 drives.....that's damn good.    

 

one reason he didn't throw for a ton of yards is because of all the penalty yardage we got.

The standards are different for RG3. When Colt under throws Jackson (even though he caught the ball) fans said he left touchdowns on the field because of his weak arm. When RG3 does the same, we hear about how Griffin is the future and no other QB could have won the game.

It's always been that way. Griffin gets excuses but the other QBs with less reps and less practice get blame.

Griffin had an average game. He didn't light the world on fire, and the defense held the eagles to three after his late game interception. The defense stood tall at the end and Garcon made a huge play at the right time that got an additional 15 added to it because of roughing the passer.

 

yes, the standrards are completely different for RG3.....his mistakes are far more criticized then any other QB

[Fred Jones]

Will RGIII be able to reach a point that he will be able to consistently read NFL defenses?  Execute a pro offense at a high level through the season.

 

So far, the answer is no.  Looking at past history around the NFL, the odds are that he will not develop.

 

That said, it would be great if he did.  And, I think some posters unfairly target RGIII, but many do not.  He connected on two long passes that setup TD's.  In the Tampa game he failed to connect and produced 7 points.  You could argue about hitting Jackson in stride, but I think you are getting picky.  Without those two long passes we lose this game.  RGIII gets credit for that as he should, but that does not negate the other games and any future play.  Teams will take that play away from the skins.  RGIII has not demonstrated he can see the entire field and read what the defense is taking away and what the defense is allowing.

 

Two passes and a win is progress, but let's see if he can repeat the performance in the last game at home.  A quality QB can.

 

Edit:  To RGIII getting treated unfairly, this is the NFL.  You don't like it, produce.  Get the ball downfield.  Win some games.  Or, give back all the money and pick a different profession.  Seriously, it would be nice to actually win some games.  Nice to finally have a franchise QB.  Appears people have settled in on mediocrity as a standard.

[HAILSKINSNYC]

let's remember that in the 3 years RG3 has been here; there isn't one QB in the NFL that's had a worse combination of defense and special teams.  

[bradboyd80]

Griffin had an average game. 

Not pointing you out Destino, but this comment was worth referencing.

 

What seems to be the issue on ES right now is that those who don't believe in RG3 at all, are still nit picking to find something to complain about, just to complain (such as the ridiculous social media argument in the other thread).  They lack the ability to give credit where its due. 

 

The RG3 supporters are "jumping for joy" right now because his average performance yesterday was a vast improvement over what he's shown so far this season.  Not only that, but this is his second decent performance in a row now.  And not only is he making good plays, but he's doing things he hasn't done very well at all this season (going through his progressions, moving the safeties with his eyes, moving around in the pocket, showing confidence, etc.). 

 

What we should ALL be doing right now is hoping that he continues the improvements and wins us another game next week.  

[Destino]

I hope he comes back next season much improved and is featured as the strength of the offense. There is a reason he throws less than Colt and Kirk, and that needs to end. He needs to read defenses and be relied on to lead the team, not simply avoid making mistakes.

He's got to put in the work. You hear that same line from everywhere because no one can make him read defenses, that's something he had to dedicate himself to learn. If he does that I think he'll be fine.

I'm rooting for him. I'm not a hater or a fan of his. This team has tricked me too many times, I'm stuck in "show me" mode with all things Redskin.

[perez24]

I'm rooting for him. I'm not a hater or a fan of his. This team has tricked me too many times, I'm stuck in "show me" mode with all things Redskin.

I'm a fan of any of the 3 coming in and clearly beating out the other 2 AND a bunch of opposing teams as well.

[Tsailand]

I'm going to post some week 17 stats, but first some more comments on ANY/A and how it compares to other advanced stats.

 

The good: 

- It makes sense.  A yard is worth a yard (technically, it isn't, but that's another discussion).  You gain yards by completing passes, you lose yards by taking sacks. A TD is worth extra yards.  A turnover is worth negative yards.  You divide by the total number of passing plays, because this isn't fantasy football; we care about how much value the QB adds per play, not the total number of times they let him throw it per game (if that last bit is a difficult concept for you, you should probably be posting in another thread).

- It's easy to calculate and understand what the stat means.

- The numbers behind it -- plus 20 yards for a TD, minus 45 yards for an int -- are based on advanced research.

 

The bad:

- It ignores advanced factors like down and distance and clock and score and the quality of the other team's defense

- It doesn't attempt to separate the value added by the QB's play from the value added by his teammates. 

 

I'll provide an example to illustrate that last bullet point. ANY/A considers these two TD passes to be of equal value:

1. QB throws a perfect deep bomb 50 yards in the air.  The WR catches it in stride and runs the remaining 20 yards to the end zone.

2. QB throws a screen pass to a WR behind the line of scrimmage, who runs 70 yards untouched through the entire opposing defense thanks to his speed, a great block by Trent Williams, and terrible angles by the Cowboys.

 

I think you see where I'm going with this.  RG3's ANY/A for yesterday's game was 5.43.  This is bad but not the absolute worst -- for comparison, out of 33 QBs ranked this season, 28 have better season averages and five have worse.  Cam Newton is 28th and he's going to the playoffs.

 

However, QBR correctly assigns the most of the credit for the 69 yard D.Jax screen and the 47 yard Garcon screen to them, not RG3.  And without those two plays, which were our only pass plays of at least 20 yards yesterday, he had a dreadful day.  His QBR was an abysmal 20.3.  None of the 30 QBs "ranked" have an season average QBR that low; the worst is Blake Bortles at 21.9

 

The QBR stat is telling us what we could see for ourselves on the field -- RG3 was horrible against the Cowboys yesterday.  The sad part is, that was only his third-worst game of the season; his 49ers and Bucs games were even worse according to QBR.

[redskins59]

Advanced stats show a TRUE picture.  People need to realize that these are all negative plays:

1. Interceptions

2. Sacks

3. Fumbles

 

etc.

 

People look at Cousins' interceptions without accounting for the sacks he avoids.  When you account for all the negative plays, Cousins will no doubt appear superior.  Because when it comes to RG3, you need to account for the sacks he took.  

Gruden I think has mentioned he is not a stat guy.  That to me explains why he doesn't see the big picture.  He seems to grade interceptions higher than other negative plays--the reason Cousins got pulled.

Your ANY/A stat also clearly shows that Cousins was the best QB.  

IMO, it is pretty clear to me that the Redskins need an Analytics guy/statistician.  

[Tsailand]

The final QBR Rankings for this season show only 30 QBs, because they require a minimum number of plays to appear on the list.    Because of the way the QB duties for the Redskins were split up this season, none of our passers made the list.  http://espn.go.com/nfl/qbr

 

However, we can go to their individual pages to see how their season QBRs compare to the 30 QBs who made the list. 

 

Kirk Cousins: 46.8.  This would put him at 24th, about halfway betweeen Alex Smith (49.4) and Brian Hoyer (43.1).

Colt McCoy: 46.1.  Also 24th, just a little behind Cousins.  According to QBR,  McCoy and Cousins had approximately equal seasons.

Robert Griffin: 30.8.  30th.  Closer to Geno Smith (35.4) than Blake Bortles (21.9).

 

Remember that QBR does not take injuries and opponent defenses into account, so you are free to argue that some of our QBs had easier or harder schedules.

[50yrSKINSfan]

Cousins threw the ball before the WR's got out of their break which is why he didn't take many sacks.  His bad interceptions were either miscommunication or forced passes.

 

Cousins will never take as many sacks as RG3 or Colt because he understands how to get through his reads and anticipate when the WR is about to be open, we don't have another Qb on the team that can do that.

Agree. Seems to me KC was always behind in most games as they got into the second half and he was forced to throw to get back in the game. Teams loaded up on him and played nickel against him. Our D sucked so much he always had poor field position. Just my take. A lot easier to play QB on the Pats than the Skins.

[TD_washingtonredskins]

 

 

Yeah bro, Cousins was on pace for a weird season.  Something like 5,000 plus yards passing, 29 TDs, and 28 picks.  He's got the dragon for sure.

 

That looks almost exactly like Manning's rookie season (more yards, but Cousins is also in his third season). 

[Birdlives]

Our QB's belong in Oz. One lacks brains, one lacks courage, and one lacks heart.

[Tsailand]

Football Outsiders (DVOA) has their final QB rankings for the regular season, but first I want to talk about another one of their advanced stats: Defense-Adjusted Yards Above Replacement (DYAR).  This stat is in comparison to a hypothetical "replacement player" at any given position -- the first player who comes off the bench when a starter is injured or cut.

 

DYAR is a per-game/per-season stat, not per-play.  So it's much like fantasy football points.  To get a high DYAR, you need to be good and get a lot of touches.  A very low (negative) DYAR requires you to be bad and still get a lot of touches. 

 

For example, the receiver with the best DYAR this week was Eric Decker, at +86.  This number is meant to represent yards of field position, but it's not directly based on his statline (10 receptions on 11 targets, 221 yards and a TD), but on many factors, and the formulas used are secret.  On the other hand, Marqise Lee was targeted 8 times, but only caught 2 of them, for a grand total of 8 yards. In a PPR fantasy football league, this would still be good for a few points, but according to DYAR this is worth -39, meaning 39 yards worse of field position for this team than you would expect a bench player to get from those same 8 targets.

 

Because QBs have the most impact on the game, they tend to have the best and worst DYARs.  The QB with the highest DYAR this season was Big Ben, with +1598.  The list goes down from him (in order) Rogers, Manning, Brees, Romo, Brady, etc.  At the very bottom is poor Blake Bortles, -995 DYAR. The QB closest to zero, the hypothetical "replacement QB", was Ryan Lindley.

 

 

The Week 17 DYAR numbers: http://www.footballoutsiders.com/quick-reads/2014/week-17-quick-reads

34 QBs got significant playing time this week. Charlie Whitehurst is at the very bottom, -118.  RG3 is 33rd, -101.  Blake Bortles' DYAR was -90, meaning he hurt his team less this week than RG3 hurt the Redskins.  This doesn't mean that Bortles played better, just that his coach didn't give him as many opportunities to throw the ball as Gruden gave RG3.

 

 

The full-season DYAR numbers: http://www.footballoutsiders.com/stats/qb

There are 44 QBs in the main ranking, plus 19 others who threw between 10 and 99 passes and don't appear on the main ranking.

Of those 63 passers, the worst DYAR is Bortles (-995), second worst is McCown (-665), and third worst is RG3 (-371).  Again, this does not mean that RG3 was the third-worst QB in the NFL, just that the ones worse than him weren't given as many opportunities to hurt their teams as he was given to hurt the Redskins.  For example, our old friend Jason Campbell was horrific, but he only got 20 dropbacks all year.

 

Think about this for a minute:  RG3 only played 1/3rd of a season, and yet he hurt his team more this year than any other player, out of 1700+ in the league, except for McCown and Bortles.

 

 

Colt McCoy 2014 total DYAR: -44.  This is close enough to zero that we can say McCoy is what we knew he was: replacement level.  Your basic backup QB.

 

Cousins: 2014 total DYAR: 224.  This is quite respectable.  He's closest to Andy Dalton (232), who got to start the whole season, and Marc Sanchez (209), who got half a season.  Cousins did as much to help us this year as those QBs did for their own teams, yet he did it in only five-and-a-half games.

[Tsailand]

So, Football Outsiders' DVOA stat.  This stat is per-play, best for getting an idea how well (or poorly) a QB played.  It is in comparison to an average player, not the hypothetical "replacement player" of DYAR, but all we care about is the ranking anyway.

 

http://www.footballoutsiders.com/stats/qb

44 QBs ranked.  We're going to ignore the 19 QBs with 10-99 passes this time, because small sample size.

 

Cousins: 15th best, out of 44.

McCoy: 35th.

Griffin: 41st.

 

Not much to add to that.

[redskins59]

So essentially, DYAR, DVOA, ANY/A and I would argue even QBR all suggest Cousins was our best QB.  Someone should send Gruden a link to this thread. It pretty much tells you who should be our starter.

Unfortunately though, TK mentioned that Gruden wants Cousins to be traded.  

[pjfootballer]

Our QB's belong in Oz. One lacks brains, one lacks courage, and one lacks heart.

Maybe if we click our heels three times, a franchise QB will appear.

[Dont Taze Me Bro]

Maybe if we click our heels three times, a franchise QB will appear.

And this thread will disappear......