WhatIfSports: Pack the World Champs

The website Whatifsports has played out the entire NFL season, Monte Carlo style, and has established the Green Bay Packers as the favorites to win the Super Bowl over the San Diego Chargers.

Before you go online to check the price of February flights to Dallas, last season Whatifsports had New England playing Philly in the Super Bowl, and they had the Packers going 7-9 and missing the playoffs.

I’m not picking on them.  I’m just saying the NFL is very hard to predict, especially given its one-off playoff system and its homefield advantages.  In fact, WhatifSports 71% accuracy mark is what I believe to be the best anyone can do without getting lucky.

If you accept the premise, the worst record the Packers can have this season is 9-7, because they have the Packers winning 13 games and losing 3 (13 * .70 = 9.1).

I just went through WhatIfSports predictions from last season and they missed on average by 1.4 games.  The most common miss was 2 games.  They missed by 3 or more games only 4 times.

Thus, my prediction of 11 wins looks good.  Now, as for the Super Bowl, well that takes a lot of luck.

The “true” strength of the Packers 2010 schedule

Just messing around, I calculated the 2009 home/away adjusted point differential for the Packers 2010 schedule to try to guage this season’s schedule’s relative strength.

Adjusted as mentioned above, the Packers opponents combined point differential would have been -20.6 points, which, if you figure that 43 points is worth one win above average, then the schedule strength is -0.5 wins above average.

Meaning, if the Packers played average football against this schedule, their record would be 8.5 wins and 7.5 losses.  So its a schedule of slightly below average strength.

If the Packers play to last season’s level, then they should once again win 11.2 games.  (I figured the Packers 2009 “level” by adjusting for last season’s -2.9 strength of schedule, as posted by pro-football-reference.com)

If you throw out the Detroit Lions, and replace them with an average team, then the schedule becomes slightly harder than average, but not by much.  I suspect its that way throughout the NFL, given the League’s parity.

How inaccurate is Mason Crosby’s FG kicking?

Last season there was a growing sentiment among Packer Backers that Packers K Mason Crosby needed to be replaced, or at least scared with a little exhibition known as a “kicker tryout”.  It seems he had missed one too many field goals. 

Crosby’s detractors could cite to Crosby’s low 75.0% field goal accuracy as prime evidence that he was substandard.  Indeed, that accuracy mark helped rank the Packers in the bottom half of the NFL in terms of FG accuracy.  The NFL average FG accuracy is around 85%, so Crosby was, in general terms, about -10% below average.

But you cannot simply cite to blanket accuracy statistics when referring to field goals.  You have to account for the distance on each attempt.  What I wanted to know is this:  just how far below average was Crosby if we adjust for distance?

A promising beginning, a downward spiral

Even after adjusting for the distance of each attempt, it appears the Crosby detractors are correct when they say he is below average.  And more disturbingly, his distance adjusted accuracy appears to be in freefall. 

Adjusting for the distance on each of his attempts, in 2009 Mason Crosby should have made 81.6% of his field goals.  As mentioned, he made only 75%.  That means he was a distance adjusted -6.6% below average.  That is slightly better than the raw numbers, but it continues a disturbing pattern.

Crosby appears to be losing more and more accuracy each season.  In his rookie season, Crosby showed a lot of promise, posting a distance adjusted accuracy that was +5.4% above average.  Then, in 2008, he slipped, posting a distance adjusted accuracy that was -1.1% below average.  For his career now Crosby is a distance adjusted -2.0% below average.

If we take a look at Crosby’s performance from the specific distances, it is clear Crosby gets comparatively worse as he moves outward (remember, this is compared to the average NFL kicker, so an outward decline is not normal).  Crosby is +3% above average on kicks that are 29 yards and in; Crosby is -1.2% below average on kicks that are between 30 and 39 yards; Crosby is -6.3% on kicks that are between 40 and 49 yards, and; Crosby is -6.6% on kicks that are taken from 50 yards and beyond.

How did weather affect his kicking?  Clearly, he’s not kicking in a dome.  But then again, Mike McCarthy will normally factor in conditions when asking Crosby to attempt a field goal.

Overall in 2009, the Packers game conditions last season averaged 48.4 degrees Fahrenheit with an average wind speed of 6.9 mph.  Not terrible.  Meanwhile, Crosby’s average conditions per field goal attempt were slightly better, averaging 51.3 degrees Fahrenheit and a wind speed averaging 5.1 mph.  So he wasn’t asked to kick in the worst of conditions. 

The bottom line is Crosby needs to turn things around.  I think this season will be pivotal for him.  If he continues his descent into a below average kicker, he may need one of those US Post Office address forwarding kits.  But he has shown he is capable of being an above average kicker.  The question is, can he turn things around and get back to being the above average kicker he showed he could be as a rookie?

Spencer Havner: deadly weapon??

Using data from Yahoo Sports and the Football Outsiders, last night I calculated the “Target Value” for each of the Packers downfield receivers in 2009 (I consider screen passes the equivalent of hand-offs.  I will get into that reasoning later.).  You’ll be a little surprised by the results I got.  But first the methodology.

Downfield “Target Value” explained

What I wanted to determine is the number of extra yards each of the Packers downfield receivers produce compared to the number of times they are targeted by Aaron Rodgers.  This I refer to below as the receiver’s “target value”.

To get to that number, I first determined each downfield receiver’s yards after catch and yards per catch.  The yards after catch, I reason, are the receiver’s yards.  The other yards belong to Aaron Rodgers. 

Then I determined the number of times the receivers were “targeted” or thrown at, and the number of receptions they made, or their “success rate”. 

Next I calculated how often the average quarterback is on-target.  It turns out that for Aaron Rodgers, the on-target percentage is somewhere around 70% for downfield passes.  Thus, any “success rate” lower than 70% reduces the receivers total yards produced by the average number of yards downfield Aaron Rodgers pass. (basically it represents the receivers dropped passes)

So for the final “Target Value” you have the receivers yards produced after the catch minus yards cost per dropped pass divided by the number of times Rodgers threw at him.  (I also adjusted by position for each receiver’s ”hit point” or the average point at which he was hit with the pass, and credited that to the receiver.  The reason was, Donald Lee’s “hit point” was 3.6 yards downfield.  Spencer Havner and Jermichael Finley both had hit points that were more like wide receivers, 9.5 yards down field.  That difference, I felt, needed to be reflected.)

With that, here are the “Target Values” for each of the Packers downfield receivers.

1. Spencer Havner, 3rd TE……(+7.21)

Comments:  Havner, the third string tight end, was only targeted 10 times by Aaron Rodgers, but his production was incredible.  His success rate was nearly 100%, his “hit point” was as far downfield as Jermichael Finley’s, and his 7.1 yards produced after the catch were better than any of the Packers other downfield receivers.  Now, perhaps he was a gimmick that team’s did not expect and therefore was left wide open.  We shall see.

2. Jermichael Finley, 1st TE……(+5.60)

Comments:  Jermichael can stretch the field.  What a weapon.  He catches everything at wide receiver “hit points”, and then advances down the field.  His success rate was remarkable, and his yards after catch were the highest on the team for any “high target” receiver.  No wonder he’s garnering such attention.

3. Jordy Nelson, 4th WR……(+3.69)

Comments:  Here’s another surprise.  The Packers number four receiver on their depth chart had the team’s highest “Target Value” among all of their receivers.  Why?  Because he gets downfield and he catches nearly everything.  His yards after catch (4.0) are not as impressive, and are in fact the lowest of any downfield receiver.   But there is certainly value in the act of completion.  And he seems very capable of getting open.  I wish I had “player minutes” or at least “plays per game” statistics, but my sense is that Nelson played less than number three receiver James Jones and yet was targeted an equal percentage of the time, implying he got open more often.

4. Donald Driver, 2nd WR……(+2.95)

Double D, still getting it done.  His success rate and his YAC were both pretty respectable.  Plus, he was targeted at a nearly equal rate as number one reciever Greg Jennings, implying he can still get open.

5.  Greg Jennings, 1st WR……(+2.76)

Comments:  I think Jennings had a bit of a down year in 2009.  I haven’t yet gone back and looked at past seasons, but I will.  Jennings success rate in 2009 was the second lowest on the team.  Granted, he is covered by the opponents number one cornerback, but that’s part of being a number one receiver.  Jennings Yards After Catch were just average.  At the moment, the Vikings hold a major advantage over the Packers at number one receiver.  Sidney Rice, at least in 2009, was much more productive than Greg Jennings.

6. James Jones, 3rd WR……(+2.33)

Coach McCarthy said Jones had the best off-season of any Packers player.  Does that mean he started catching a reasonable number of passes?  Jones is a drop machine and he doesn’t make up for it either with plus “hit points” or with YAC.  Nor does it appear that Jones gets open at any eye-popping rate.  If he doesn’t show massive improvement in the preseason, I think the Packers would be foolish not to move Nelson up to number three receiver.

7. Donald Lee, 2nd TE……(+0.70)

Donald Lee is basically a blocking tight end.  He’s good at hauling in passes that are on target, but his “Hit Point” last season was only 3.6 yards down field, much less than the Packers other two tight ends.  You can probably extrapolate from that the usage of Lee.  When Rodgers couldn’t find anyone, he settled on Lee as an escape valve.  Lee’s YAC was nothing to write home about either.  But good blocking definitely has value, just not ”Target Value”.        

The value of Ryan Grant versus Adrian Peterson

Ryan Grant may be getting the shaft from Packer fans.  Yes, he had a poor 2008, after his long holdout.  But arguably he was as valuable to the Packers as Adrian Peterson was to the Vikings in 2009.

Each man averaged 4.4 yards per carry.  That is above average, but only by 0.2 yards.  Nevertheless, every yard above average tends to produce a win above average.  And, bare in mind, these numbers are not adjusted for strength of defense.

So, on Grant’s 282 carries he produced for the Green Bay Packers +56.2 yards above average, and he caught the ball 25 times, producing +9.5 yards above average.  Peterson, on the hand carried the ball 313 times and gained +62.6 yards above average, plus he caught the ball 43 times and produced +113.5 yards above average.   So the main difference in value from scrimmage between the two was, ironically, produced by Peterson’s receiving yards, not his rushing yards.

But now we get to the rub, ball safety.   I went through the statistics on Yahoo.com’s NFL running backs page and I calculated that among running backs with at least 100 rushing attempts, the fumbles lost per play (rushing attempts + receptions) was only .008%.

So, in Ryan Grant’s 307 touches last season the average back would have lost 2.56 fumbles.  Grant lost only one.  Grant, therefore gained the Packers the value of his 1.56 non-fumbles.  Each fumble costs a team approximately 45 net yards, so Grant’s ball security added +68.9 yards to the Packers net total, for a personal total net yards above average of +128.6 yards.  Translating those extra yards into wins above average, and Grant  produced +0.30 wins for the 2009 Green Bay Packers above what the average running back would have produced.

On Peterson’s total number of plays, the average back would have lost 2.49 fumbles.  Peterson lost 6.  So Peterson’s fumble problems cost the Vikings -141.5 yards. Peterson, then, only produced +34.6 yards above average for the 2009 Minnesota Vikings.  That translates into only +0.08 wins above average.

Lets assume that Peterson did not have a fumble problem.  Then what would the comparison be?  Well, if we don’t include fumbles in the equation, then Ryan Grant produces only +65.7 for +0.16 wins above average, while Adrian Peterson adds a much improved +0.43 wins above average to the Vikings.

But fumbles lost must be included.  They are brutally damaging to a team, and they are the reason why Ryan Grant had more value for the Packers than Adrian Peterson did for the Vikings.

EDIT:  This version is radically different from my original post.  In the original I used fumbles per rushing attempt, assuming that the type of attempt did not matter.  I got some puzzling results that made both Grant and Peterson appear to be extremely surehanded.  Grant is, but I was almost certain Peterson was not.  After posting it, I checked the far superior Yahoo.com NFL statistics and I found that when you consider only running backs, fumbles per rushing attempt basically disappear, which I thought had to be the case.  Most fumbles appear to occur on sacks or quarterback scrambles. 

These are the kind of mine fields you have to watch out for with the awful statistical records kept in the sport of American football.

All-Time Ranking: Green Bay Packers starting quarterbacks since 1932 (by adjusted QBER)

In relative terms, the most effective Quarterback in Green Bay Packer history wasn’t from Mississippi or Alabama or even California.  He was a Wisconsinite, believe it or not, and in fact he was born right in Green Bay.

He stood 5’11 and completed barely 40% of his passes.  Nevertheless, and even though he somehow made only one All-NFL team, he was the dominant quarterback of the 1930s — the King of the Forward Pass.  His name was Arnie Herber.  He led the Packers to their first two World’s Championships of the playoff era with his brilliant performances in 1936 and 1939, then he left the Packers to fight in World War II.

Ranking the Packers Starting Quarterbacks

I ranked every starting quarterback in Green Bay Packers history using a metric I call “Quarterback Effectiveness Rating” or QBER.

QBER correlates strongly with both Professor David Berri’s QB Score and also with Football Outsiders “Defense Adjusted Value over Average”, two of the best football win metrics.  That means even though QBER is a performance metric, it still correlates very strongly with a quarterback’s ability to produce wins for his team.  And that, of course, is the name of the game in sports.

QBER, in a basic sense, is a measurement of the number of forward yards a quarterback produces for his team for every non-productive play (incompletions and sacks).  Forward yards are equal to passing yards plus running yards minus yards lost by sack minus interceptions * 50 yards (the accepted value of yards each interception costs a quarterback’s team).

QBER is basically dependent upon three statistics:  Completion %, Interception Rate, and Yards per Completion.  If the Quarterback is above average in all three areas, he will have a high QBER.

To avoid historical issues involving liberalized passing rules, I adjusted each quarterback’s statistics in the three areas above to 2010 values by adjusting the quarterback according to the prevailing averages when he played.  The average QBER is thus set to the 2010 average of 10.18.  Above that, and the quarterback was above average in effectiveness and win production.  Below 10.18 and he was below average.

1. Arnie Herber (1932-1940)

Adjusted QBER: 18.44

I remember seeing him at the Packer Hall-0f-Fame and not knowing who the hell he was.  He was the Packers most effective quarterback ever, that’s who he was.   During his short but brilliant career, he was above average in all three major categories, and his incredible 16.1 yards per completion in 1936 still stands as the highest mark ever, and his 33.09 QBER that season not only led the Packers to the World Championship, it is the second highest single season QBER I have found to date (the highest being Sammy Baugh’s ridiculous 1945 season — check out those stats some time).  To put it in perspective, Tom Brady’s QBER from his tremendous 2007 season was 21.77.  Dan Marino’s legendary rookie season in 1984 was 22.52.

2. Aaron Rodgers (2008-date)

QBER: 15.33

Ted Thompson’s bold, courageous move to install Rodgers as his starting quarterback has paid off handsomely.  Believe it or not, Rodgers has already proven himself a more effective quarterback than Brett Favre ever was (in Green Bay).  Favre had only one season in Green Bay where his adjusted QBER was higher than Rodgers career QBER so far (that was Favre’s 1997: 15.57).  And as good as he’s been, if Rodgers didn’t get sacked so much, his numbers would be monstrous.  But I believe he consciously trades interceptions for sacks — a wise decision, actually.

3. Bart Starr (1956-1972)

Adjusted QBER: 15.19

Because he did not have a strong arm, people just refuse to recognize the greatness of Bart Starr.  After Jim Taylor ran out of gas, Bart Starr basically carried the Packers to their final three World Championships.  His 1966 season adjusted QBER stands as the second best in Packers history (19.61).  He carried a well above average completion rate, a well above average interception rate, and a just slightly below average yards per completion. Because of that he had a more effective career than Johnny Unitas (13.51), Fran Tarkenton (14.18), Brett Favre (12.12), Bob Griese (12.63), and Joe Namath (11.44), to name just a few.

4. Brett Favre (1991-2007)

Adjusted QBER: 12.12

Favre’s 17.89 last season in Minnesota was by far the highest mark of his career.  When Favre was in Green Bay he was an above average pass completer, and he was average when it came to yards per completion, but he was reckless with the football.  His interceptions killed Packer drives and killed his effectiveness numbers.  Its fitting that the gunslinger’s final pass in a Packers uniform was a hideous interception that cost them a trip to the Super Bowl and possibly a 13th World Championship.  Adding insult to injury, end zone views of the play showed several Packer recievers wide… wide….open.  Favre chose instead to throw a dangerous out pattern that the defender easily jumped in front of for the dagger.

5. Don Majkowski (1986-1992)

Adjusted QBER: 11.34

The Majik Man.  The fact that he finishes fifth on the list of all-time effective quarterbacks for the oldest franchise in the NFL explains why Brett Favre was so deified.  Majikowski was basically a mediocre quarterback who had one pretty good season in 1989.  After that injuries and a contract hold out marked the end of his days in Green Bay (I still remember a sign at Old County Stadium during his 1990 holdout — “Is the Majik gone?” — the fan could not  have been more prescient.)

6. Tobin Rote (1950-56)

Adjusted QBER: 10.93

This one shocked me.  I thought those were the dark days of Packer football, but Rote was above average.  He actually led the Detroit Lions to a World Championship after the Packers dismissed him, and he was later inducted into the Hall of Fame.  He wasn’t much of a passer during his Packer days, but he was a terrific runner.  Kind of the Vince Young of his day.

7. Zeke Bratkowski (1966-71)

Adjusted QBER: 10.52

Zeke only actually started 13 games during his 6 year stint with the Packers, but he’s rightfully a lot better known than many of the other quarterbacks on this list.  I never realized he was already 33 when he came to the Packers, but he performed very well for a backup.  Better, in fact, than every starting Packer quarterback after he left in 1971 up until Don Majkowski.  Long after their playing and coaching days ended, he and Bart Starr both lived on the same street in De Pere, Wisconsin and I delivered papers to their houses.  True story:  Bart Starr came out one cold, cold December morning to shake my hand and give me a Christmas check (“Hi son, my name is Bart Starr“).  Class act.  But I never once saw Zeke, even though I had to collect from his house.  His wife always took care of the payments.  He was nowhere to be seen.  Oh, apparently he’s also the middle man in the old Packer phrase (and Polish joke) from the 1960s… lets see if I get this right… “Skoronski (the center) to Bratkowski (the QB) to Grabowski (the RB)”.

8. Lynn Dickey (1976-1984)

Adjusted QBER: 9.44

Now we come to the below average guys.  Lynn Dickey always struck me as decent, but he was really nothing more than that.  He was completely immobile, and threw too many picks.  But it was him or David Whitehurst, so it was like a Faustian choice.

9. John Hadl (1974-75)

Adjusted QBER: 8.44

Never has so much been given up for so little.  Didn’t the Packers trade a couple of number ones for this aging stiff?  He was in and out of town pretty quickly.  Most of his better, but still overrated days, were with the AFL’s Chargers (the AFL’s passing abilities ere in general wildly overrated… the league was much worse in every category than were the NFL passers).

10. Jerry Tagge (1972-73)

Adjusted QBER: 8.43

Tagge played on the legendary 1970 and 1971 national championship teams at the University of Nebraska, and he was a hometown boy.  I think he went to Green Bay West.  I remember the guy that ran the batting cages and putt-putt golf on Military Avenue in Green Bay had a picture of Tagge on his wall.  He wasn’t much of a pro quarterback, however.  I guess he had alcohol problems.  Tough break.  Pretty decent college quarterback, though.  Two time Orange Bowl MVP.  Actually I found his stats here.  Sometimes its just circumstances that overwhelm a guy.

11. Randy Wright (1984-1988)

Adjusted QBER: 8.43

I always liked Randy “Wrong”.  He was an outstanding quarterback at Wisconsin — a transfer from the University of Notre Dame, as I recall.  And he executed the famous “bounce pass” play to Al Toon.  Your author was in the stands for that game.  It was my first Badger football game, and I never forgot how amazed I was at that play.  It led me to Madison years later for my degree.  A perfect bounce pass with a football off the astroturf.  Oh, and he quarterbacked the Packers during the dark days of the Forrest Gregg Era.  He was as effective as former All-American Jerry Tagge.

12. Babe Parilli (1952-53; 1957-58)

Adjusted QBER: 8.34

The “Babe” of football, indeed.  Babe Parilli really wasn’t too effective for your Green Bay Packers, starting his career at City Stadium, without much success.  Believe or not, though, he went on to make first team All-Pro in the AFL with the Boston Patriots, which tells you all you need to know about the myth of the AFL’s passing superiority.

13. David Whitehurst (1977-1983)

Adjusted QBER: 8.23

On the playgrounds in the early 1980s, playing Nerf Football, no one wanted to be “David Whitehurst”.  His adjusted QBER kind of explains why.  According to pro-football-reference his first name is actually “Charles”.  As Johnny Carson would say, “I…I did not know that”.  Whitehurst tasted some limited success in 1978 when the Packers went 8-7-1, but missed the playoffs.  I believe his son is considered an interesting up-and-coming quarterback right now.  Are you beginning to see why Favre was so revered?  From Bart Starr to him, an absolute barren wasteland.  Well, at least we didn’t ever have a JaMarcus Russell in the mix.  The Raiders somehow let him start games with a QBER of 2.12.  Kind of puts the Packers gloomy era in perspective.  But, I digress, on to our final Packer starting quarterback.

14. Scott Hunter (1971-73)

Adjusted QBER: 7.93

He quarterbacked the Packers to their only division championship of the 1970s, and I can honestly tell you I’ve never heard of him.  Normally if a guy was a Packers starter, even a brutal one, his name somehow comes up in front of the television on a Packer Sunday (“He’s another [.....]“).  Packer starters are quite well known in Wisconsin households, and for good reason.  As my rankings show, there really haven’t been many of them.

Welcome to City Stadium

This is the introductory post for this Green Bay Packers football blog. 

For nearly a generation and a half, prior to the construction of legendary Lambeau Field, City Stadium on the shores of Green Bay’s East River was the home of the Green Bay Packers.  The Stadium still stands, minus the bowl seating it had during its Packer days.  It serves as the home of the Green Bay East High School Red Devils. 

I am a Green Bay Packer fan, so while this blog will be about the National Football League generally, it will be about the Green Bay Packers specifically.  I will often write opinion based pieces, but what I like to write are pieces that analyze sports through statistical analytics.  I have not developed any football metrics of any great sophistication, so I will be relying heavily upon the excellent work of Brian Burke at AdvancedNFLStats and the Football Outsiders.