Ponder: An Elite Quarterback

Earlier this season, I argued that Christian Ponder had the potential to be an elite level quarterback based on his first year performance.  I compared his 1'st year numbers to the 1'st year numbers of two widely accepted elite quarterbacks (Colt McCoy '06 and Sam Bradford '07).  Ponder played like a warrior in the first nine games of our season.  He made Seminole Nation so very proud during an otherwise dark chapter in Seminole history.  His dazzling season at the healm of Jimbo Fisher's high octane offense was tragically cut short as he delivered a bone-jarring tackle (with broken ribs) to Clemson's DeAndre McDaniel and suffered a season ending injury.  Ponder was considered by many to be a dark horse in the race to the Heisman.  Based on data from his 2009 performance, I will prove that Christian Ponder is an elite quarterback just as deserving of the title "Best Collegiate Football Player" as the other Heisman candidates.

In this analysis, I compared Ponder's '09 performance to the performance of the 3 Heisman finalist QBs:

1. Timmy Tearsbow
2. Colt McCoy
3. Kellen Moore

Out of curiosity, I included EJ Manuel in there as well (don't read too much into his results, as 3 observations provide little statistical confidence).

The metric I will use to gauge the QB's performance is the ESPN calculated passer rating.  Below is a box plot of these 5 QB's passer ratings:

Here's a little stats 101 review (copied from my previous post) of what this box plot is showing:  This box plot is a visual representation of the passer ratings these QBs posted during the 2009 season.  The top of the box is the third quartile (Q3, meaning 75% of the observations are less than or equal to this value).  The middle line is the median (meaning that 50% of the observations are less than or equal to this value.)  The bottom of the box is the first quartile (Q1, meaning that 25% of the observations are less than or equal to this value).  The whiskers extend to the upper and lower limits.

Much more inside.  Continue reading!

Upper limit = Q3 + 1.5(Q3-Q1)

Lower limit = Q1 -1.5(Q3-Q1)

Ponder:

Q1: 125.03

Median:  137.05

Q3: 179.675

Whiskers to:  116.59, 201.89

Manuel:

Q1: 104.83

Median:  112.68

Q3: 173.9

Whiskers to: 104.83, 173.9

Tebow:

Q1: 128.1

Median:  166.64

Q3: 186.4

Whiskers to: 84.86, 216.5

McCoy:

Q1: 125.475

Median:  156.84

Q3: 181.17

Whiskers to: 81.82, 191.38

  Moore:

Q1: 149.64

Median:  165.79

Q3: 188.12

Whiskers to: 130.59, 211.93 

On the surface, it appears that Ponder can hang in there with the elite quarterbacks in the country.  His median passer efficiency is slightly lower than the other guys, but his whiskers don't extend as low as Tebow or McCoy.  Generally speaking, his passer efficiencies are comparable to Tebow &  McCoy.   Based entirley on this, it looks like Moore is the best QB in the group.  If I was an expert analyst at ESPN, I'd probably call it quits here and declare "Moore the best QB in country" (maybe this is a good time to state that the views and opinions expressed by the author are his own and do not represent the views and opinions of Tomahawk Nation). 

The above analysis doesn't take into account the quality of opponent and doesn't put performance into proper perspective.  It's like saying a kid who got an A in algebra is as smart as a kid who got an A in calculus; both got A's in math.  To put this into proper perspective, let's look at the box plot (below) of the FEI  adjusted defensive efficiencies that these QBs faced throughout the season.  Lower indices represent better defenses.

Ponder:

Q1: -0.4935

Median:  -0.238

Q3: 0.1325

Whiskers to: -0.66, 0.507

Manuel:

Q1: -0.377

Median: -0.044

Q3: 0.019

Whiskers to:  -0.377, 0.019

Tebow:

Q1: -0.43225

Median: -0.146

Q3: 0.2037

Whiskers to: -0.677, 0.43

McCoy:

Q1: -0.3535

Median: -0.0805

Q3: 0.07525

Whiskers to: -0.544, 0.186

Moore:

Q1: 0.1125

Median: 0.34

Q3: 0.487

Whiskers to: -0.368, 0.646

 With the exception of Moore, all of the QBs played pretty tough defenses.  Ponder's opponents had defenses with a lower median FEI defensive efficiency than any of the other QBs.  Now the ESPN analyst's claim that "Moore is the best QB in the nation" is starting to look a little shaky.  Moore had the best passer ratings, but he posted those ratings against cupcake teams.  Ponder posted elite passer ratings against a grueling defensive schedule.

Let's dig a little deeper into this by exploring the effect that a good defense has on a QB's performance.  Regression analysis can be used to quantify the effect an independent variable has on a dependent variable.  In this analysis, the independent variable (predictor) is the defensive FEI.  The dependent (response) variable is the passer rating.  The following plot shows Ordinary Least Squares (OLS) regression equations that model passer ratings as a function of defensive efficiency.  Think of it like this:

y = mx + b

ESPN Passer Rating = m*(Defensive FEI Index) + b

In the above plot, the red-square points represent individual games from Ponder's 2009 season.  In each game, he played a team with a specific Def. FEI index (X-axis) and posted a particular passer rating (Y-axis).  The red trend line is the regression model overlaid on the plot.  The color key shows which points / trend lines correspond to which QB.

Here are the models if you are curious:

Ponder (Red):

Passer Rating = 154.3 + 29.27 * (Defensive FEI Index)

Manuel (Blue):

Passer Rating = 136.8 + 47.6 * (Defensive FEI Index)

Tebow (Black):

Passer Rating = 162.9 + 59.47 * (Defensive FEI Index)

McCoy (Green):

Passer Rating = 159.3+ 105.2 * (Defensive FEI Index)

Moore (Orange):

Passer Rating = 163.1+ 19.87 * (Defensive FEI Index)

Key Points:

• There is still a lot of variability in the data that is not explained by the models.  The R-squared values for these models hover around 10-30%.  This means that defensive FEI indices only explain about 10-30% of the variability in passer ratings.  Therefore, these models should not be used to precisely / accurately predict passer ratings.  They are meant to be used for illustrative purposes to show the general effect a good defense has on the play of a quarterback.
• Tebow and McCoy (the only two previous Heisman candidates) were the most sensitive to good defenses.  They had the steepest slopes in their models.  Here is numerical proof that Tebow is not invincible and is just as much if not more susceptible to good defense than Ponder.
• Moore played a bunch of cupcakes.  I'm not saying he's bad, but I think to be seriously considered for the Heisman, you need to prove yourself against better defenses than he did.
• McCoy may be in for a bit of a shock when he plays Alabama.  The lowest Def. FEI rank he's seen all season is -0.3535.  Alabama's D has an index of -0.677 (the best in the country).  If you look at his trend line, it is evident that he posted passer ratings around 80 against comparable, elite defenses this year, but again I wouldn't put too much faith in using these models for prediction purposes.
• EJ Manuel has great potential, and I believe that this experience will go a long way in his maturation.  I wouldn't draw any meaninful conclusions from his blue line above due to poor statistical confidence, but it does show you how he performed relative to elite talent in the three games he played in.
• Based on Ponder's trend line, he performed the best of all of the QBs against defenses with defensive FEI indices of -0.368 or below.  Let that sink in.  Ponder had the best play against the most talented defenses among all of these QBs.  Here is quantitative proof that Ponder is an elite level quarterback performing as well if not better than the 2009 Heisman candidates for the QB position.  Congratulations Christian Ponder on a Heisman-worthy 2009 campaign!