After talking up how Rodgers would do well to emulate Tom Brady’s previous behavior by working overtime to build up loyalty within the Packers players and organization, Seth Wickersham strikes back with this, in an article about livening up NFL minicamps:

4. Scrap official team minicamps for more creative ones. You could have the "Hands Off Camp," featuring John Abraham, Jevon Kearse, and Donovan McNabb. Or the "You Wish You Could Pull This" camp, featuring NFL players with celebrity girlfriends. Or the "We Thought Jeff Tedford Would Help Our NFL Careers" camp with Trent Dilfer, Joey Harrington, Akili Smith, Kyle Boller, and a few others.

He didn’t explicitly name Rodgers, but I’m sure he would be included in the "few others" category.  (Incidentally, that comment wounds me deeply on two levels.  Come on, Eagles!  Keep those guys healthy!)  But as for the Tedford Effect, does it really exist?  From Projecting Rookie Quarterbacks, by David Lewin in Pro Football Prospectus 2006:

At first, I built a regression model to project NFL success based on career DPAR per game.  I tried a number of different variables in the regression, although due to the small sample size of college games, I was somewhat limited.  In the end, two factors stood out far above the all others: completion percentage and games started.  Statistics like attempts, completions, and yards were relevant, but are highly collinear with games started and not quite as statistically significant, so they were left out.  Height proved not to be statistically significant, and neither did draft position (an issue discussed further below)…

So what do these numbers mean?  Beginning with games started, these finds probably have two meanings.  One, more college experience usually leads to better NFL performance.  This makes sense, as anyone who has played quarterback could tell you. 

The second implication is, perhaps, more interesting.  The correlation of a high number of games started with greater NFL success indicates that the more games played in college, the better the job scouts do at evaluating players.  While this is not always the case — Cade McNown is the most prominent exception — it is a clear trend.  It makes sense that the more film available on a player, the better that scouts known his talent level.  There are quarterbacks who started a lot of college games and posted good numbers, but scouts saw enough of them to know they lacked NFL talent.  Chris Rix, who started 38 games for Florida State, in a good example; scouts saw him play a lot in college and knew he wasn’t very good; consequently he went undrafted.  Imagine instead if Rix had come out after leading Florida State to a 10-3 record in his junior year.  With only two seasons as a starter at the college level, scouts might have seen him as "projectable," a proven winner with athleticism and a strong arm.  Kliff Kingsbury is another example; he was a wildly successful four-year starter in Texas Tech’s spread offense, but scouts saw enough of him to know he was a sixth-round project instead of a first-round prospect…

The second college statistic that points to NFL success is completion percentage, which is fairly straightforward.  Accuracy, though somewhat dependent on the system and the quality of each player’s receivers and opponents, carries over from college to the NFL.  Completion percentage is more relevant than touchdown percentage and interception percentage because it is based on significantly more observations (there are many more completions than touchdowns or interceptions) and is less dependent on the situation.

Essentially, higher DPAR is good, the more games started the better, and you can’t teach accuracy at the next level.  The bulk of Lewin’s analysis studies the 27 first-round quarterbacks taken within the 10 years prior to his study (1997-2006), which includes former pupils of our Lord and Savior, Akili Smith, David Carr, Joey Harrington, Kyle Boller, and Aaron Rodgers.

After studying the statistics and coming to the conclusion that games started and completion percentage were the most important numbers that should be used in evaluating quarterbacks, Lewin says of each:

Akili Smith
Age-adjusted DPAR/G projection: 0.42
Actual age-adjusted DPAR/G: -0.40

Smith is a perfect example of why teams need to consider college statistics, not just scouting reports.  Smith started only 19 games and completed only 56.6% of his passes in his two years at Oregon.  Despite having the talent that scouts drool over, Smith was one of the biggest busts in recent memory.

David Carr
Age-adjusted DPAR/G projection: 3.62
Actual age-adjusted DPAR/G: 1.28

If you normalize sacks to the league average, Carr’s numbers actually look decent…Carr posted impressive numbers at Fresno State, especially his senior year.  He threw for 46 touchdowns and 4,839 yards, completing 64.5% of his passes.  While Carr was probably unworthy of the number one overall selection, he has a chance to become an above-average quarterback over the next two or three years.

Joey Harrington
Age-adjusted DPAR/G projection: 1.28
Actual age-adjusted DPAR/G: 1.66

Harrington fooled scouts coming out of college in large part because of the hype from his Heisman campaign, funded by Nike CEO and Oregon booster Phil Knight.  His senior numbers were good, but definitely not Heisman caliber.  Scouts liked Harrington’s leadership ability, good arm strength, and size.  His stats however, tell a different story: fewer than 30 games started and a completion rate of 55%…Once Daunte Culpepper has fully healed, Harrington will probably spend the rest of his career on the bench, Cade McNown-style.

Kyle Boller
Age-adjusted DPAR/G projection: 0.55
Actual age-adjusted DPAR/G: 2.58

Boller rocketed into the first-round on the strength of his postseason workouts, which is almost always a bad sign.  In fact, one of the most impressive things Boller did at these workouts was kneel at the 50-yard line and throw the ball through the uprights.  Scouts thought this demonstrated a spectacular arm.  Anyone who has watched Boller play (ed. note - me, Ragnarok, and TwistNHook) knows he has good but not great arm strength…Boller also ran a 4.6 in the 40-yard dash, which everyone knows is pretty much meaningless for a quarterback.  By this time, scouts were in love.  Boller could do no wrong.  This was news to Kyle, as he had done a lot of wrong (48 interceptions) in college at the University of California.

Boller was a four-year starter at Cal.  Normally this is good, because that means there is a lot of film to watch and scouts know pretty well if you are a first-round talent or not.  In this case, the process somehow failed.  If Billick has bothered to watch Boller’s college game film, he would have noticed that the object of his desire completed only 47.8% of his passes in college.  In the modern era there has never been a successful NFL quarterback with a college completion percentage so low.  When Boller’s low completion percentage and overall abysmal college performance were pointed out, scouts would respond that Boller made progress in his senior season, when he threw 28 touchdown passes.  It is true that Boller was much improved as a senior.  However he still completed only 53% of his passes.

Aaron Rodgers
Age-adjusted DPAR/G projection: 3.33

Rodgers is unlikely to get much of a chance in 2006, but when he does he should become a solid but unspectacular NFL player, along the lines of Brian Griese.  Like Griese, he is in the unenviable situation of following a legend.  No one can really live up to Favre’s legacy, and Rodgers is definitely not going to be able to.  I expect the Packers to draft another quarterback as soon as Rodgers gets enough playing time to show the Packers what he is: a decidedly average player.

And herein lies the rub.

At Cal, Rodgers was coached by Jeff Tedford.  This puts him in an unfortunate group of recent first-round quarterbacks that includes Joey Harrington, Akili Smith, David Carr, and Kyle Boller.  Originally I thought that I might need a Jeff Tedford variable that reduces a player’s projection.  After running the numbers, it turns out that there was no trend with these guys over- or underperforming their projections.  Tedford quarterbacks just somehow look better to scouts than their numbers justify.

There you have it.  There’s no magic here; indeed, it would seem that our glorious leader doesn’t actually inflate his quarterbacks’ numbers at all.  Their performance at the NFL level is right at what their college numbers would say it is, so it’s hard to understand why scouts might on the one hand decry the Tedford Effect that causes them to take his proteges higher than their NFL performance deserves, and then propogate it by drafting them higher than their college track record deserves.  So the next time you hear from someone that Jeff Tedford quarterbacks are all first-round busts, tell them that they really shouldn’t have gone in the first round to begin with, and should be viewed accordingly.  Blame the scouts in this instance, not our coach.