FanPost

2012 Stats--a deeper dive

Ed. Note - this is a user FanPost

Last week, Dr. B posted a statistical analysis--the stats included in that post are the usual, high level, key stats that are utilized to evaluate the success (or lack thereof) of a football team. This post, is similar in nature, but takes it a step further and looks at some additional ratio type stats.

Over the past few years, as a 'hobby statistician', I have dabbled in creating college football prediction and rating models as well as rating models for college basketball. The football models have largely been predicated on determine offensive and defensive efficiencies and translating those into expected points scored and allowed vs. a given opponent.

Here is the general logic...in order to win a football game, you have to score more points than your opponent. Fairly basic, right? So, the key stats, in my opinion, come down to Points scored and points allowed per game. Again, fairly obvious...the holy grail is therefore figuring out which stats correlate and predict scoring points on offense and allowing points on defense. Often, my hypotheses are proven incorrect (for example...through the hundreds of regression models that I've run, I rarely am able to find a way to get turnovers to be a significant contributor to points scored. If you look at the Top 25 scoring offenses in 2011, the range of TO margin was from +21 to -9. The average was 4.4. For the Bottom 25, the range went from +12 to -9 and the average was -2.96. There is some positive correlation for sure, but it's not a strong enough predictor of scoring points to matter.

Anyway....through that process and through the running of hundreds of regressions using a large variety of standard and not so standard independent variables, I have consolidated down to some key stats--most of these are fairly highly correlated (either positively or negatively) with a teams ability to score points. I have used full year data from 2007-2011 to identify these factors...others might not be as strongly correlated, but I think are interesting and feel like they should matter, so I've included.

First, let's look at the YTD stats for 2012. I've include the national average, the 90th percentile for each category, the national leader in Points/G, Clemson and the Clemson's national rank. Right now, based on a very difficult early season schedule, Florida State leads in both categories. Data below...

OFFENSE DEFENSE
Avg. Top 10% FSU Clemson RANK Avg. Top 10% FSU Clemson RANK
Pts/Poss 2.4 3.6 5.0 3.3 22 1.9 0.8 0.1 2.0 74
Negative Plays/G 7.5 4.0 1.5 8.0 68 8.4 12.5 11.0 7.5 46
Pts/Play 0.4 0.6 1.0 0.5 45 0.3 0.2 0.0 0.3 73
TDs/G 3.9 6.0 8.5 4.5 36 2.9 1.0 0.0 2.0 33
Red Zone Opps/G 4.0 6.0 8.5 5.0 28 3.4 1.5 1.5 3.0 40
Red Zone Opps/Poss 32% 46% 68% 42% 28 27% 12% 13% 26% 57
3rd Down Convs/G 6.0 8.5 4.5 10.0 2 5.7 3.0 2.0 3.5 15
Plays/Poss 5.7 6.9 4.8 7.1 11 5.7 4.6 4.3 5.8 73
Poss/G 12.7 14.9 12.5 12.0 73 12.7 10.5 12.0 11.5 25
Scores/Poss 40% 57% 76% 58% 10 32% 15% 4% 43% 95
TDs/Poss 30% 50% 68% 38% 33 23% 8% 0% 17% 44

Despite the perception that our offense generates a lot of possessions, that has not been the case thus far, ranking 73rd in offensive possessions and 25th in defensive possessions. Rather, due to a very strong 3rd down percentage (2nd most 3rd down conversions/game through 2 games), which leads to a high plays/possession number, Chad Morris' crew has generated a ton of snaps without necessarily a ton of possessions.

On the defensive side of the ball...not a lot of good news to be had. Basically 2 areas to be happy about...solid on 3rd downs and holding teams to a lot of Field Goals. While our opponents have scored 100% of the time when getting into the red zone, only 1 out of 6 red zone attempts have resulted in a TD. Again, that's a small sliver of good news...especially considering the long TD plays allowed...only 25% of TDs allowed have been on Red Zone drives.

All that said...a 2 game sample is not great, so, for S&G's let's look at the 2011 Tiger squad across the same dimensions and with the same comparison sets.

OFFENSE DEFENSE
Avg. Top 10% Houston Clemson RANK Avg. Top 10% Alabama Clemson RANK
Pts/Poss 2.3 3.1 3.7 2.5 35 2.2 1.5 0.7 2.2 75
Negative Plays/G 7.7 5.6 5.1 8.4 82 7.8 10.1 9.7 6.8 84
Pts/Play 0.4 0.3 0.6 0.4 36 0.4 0.3 0.1 0.4 80
TDs/G 3.6 4.7 6.6 4.1 31 3.4 2.3 0.9 3.7 78
Red Zone Opps/G 3.8 2.7 5.4 4.0 42 3.6 2.7 1.3 3.7 64
Red Zone Opps/Poss 30% 22% 40% 30% 63 29% 21% 11% 28% 54
3rd Down Convs/G 5.8 6.9 7.4 6.7 22 5.8 4.7 3.5 6.1 88
Plays/Poss 5.6 5.2 5.8 5.7 52 5.6 5.2 4.8 5.4 32
Poss/G 12.4 11.3 13.5 13.3 21 12.4 11.4 11.6 13.1 95
Scores/Poss 37% 50% 56% 43% 28 36% 26% 13% 37% 71
TDs/Poss 29% 20% 49% 31% 45 27% 18% 8% 28% 75

The numbers put up by 'Bama's defense are just ridiculous...only 2 teams had a pts/poss below 1--Bama and LSU. Next best in 2011? FSU (1.26), Temple (1.27) and Penn State (1.34).

In fact, since 2007, there have been 8 teams with a points per possession less than one. They are shown below. Not surprisingly, those teams win A LOT of football games.

Year Team Pts/Poss Wins
2008 Boise State 0.98 12
2008 TCU 0.94 11
2008 USC 0.72 12
2009 Alabama 0.93 14
2009 Nebraska 0.84 10
2009 TCU 0.97 12
2011 Alabama 0.70 12
2011 LSU 0.99 13

Houson led the nation in points scored/game, but not in points per possession. Those honors went to Wisconsin at 3.93 points per possession--fairly impressive number. Wisconsin scored on 59.2% of their possesions...more impressive? 91% of the Badgers' scoring drives resulted in TDs...85 TDs to 8 FGs.

Clemson's offense in 2011 ranked in the top 25 in only 2 categories--3rd D conversions and possessions per game...the defense's best ranking was 32nd in plays/possession.

There is my first crack at taking a slightly deeper dive into some stats. I have a ton of data, so if there are suggestions for additional views or ratios, etc, please comment and request. I'm happy to expand or contract. Also, debate on these categories and their potential impact and importance to the end goal (scoring points and keeping the other team from scoring) is also encouraged.

Last point...next week this post will include the 2012 data and will again include FSU regardless of whether or not they are leading the categories as they are our next opponent. This will be standard going forward.

EDIT:

As requested by TheKraken...'07-'11 Top 5 per year coming shortly as I may do that as a separate post.

YEAR TEAM W O PPP D PPP PPP MRG.
2007 Clemson 9 2.65 1.52 1.14
2008 Clemson 7 1.97 1.42 0.55
2009 Clemson 9 2.44 1.62 0.82
2010 Clemson 6 1.99 1.56 0.42
2011 Clemson 10 2.53 2.23 0.30
2011 Alabama 12 3.10 0.70 2.40
2011 LSU 13 3.09 0.99 2.10
2011 Okla State 12 3.42 1.89 1.53
2009 Oregon 10 2.73 1.78 0.95
2010 Oregon 12 3.45 1.34 2.12
2011 Oregon 12 3.43 1.85 1.58
2009 Auburn 8 2.52 2.08 0.44
2010 Auburn 14 3.63 2.09 1.54
2011 Auburn 8 2.09 2.47 -0.39
2008 Oklahoma 12 3.81 1.84 1.96
2009 Oklahoma 8 2.26 1.09 1.17
2010 Oklahoma 12 2.64 1.52 1.13
2011 Oklahoma 10 2.73 1.55 1.18

These opinions are not necessarily those of the Proprietors of Shakin' The Southland.

In This FanPost

Teams