Comparing the final Football Outsiders F/+ projections to the AP poll and conventional wisdom.
28 May 2009
by Bill Connelly
In the last Varsity Numbers column, we took a look at overall "+" rankings. Today we will focus both on 2008 offenses and what may lie ahead in 2009.
As always, we'll start today with a "+" definition:
[T]he "+" concept [is] a primitive attempt at a college-level look at DVOA-type comparisons discussed here. The formulas behind the "+" concept have been strengthened since the writing of that column (instead of looking at game totals and comparing them to opponents' per-game averages, it actually looks at per-play totals, eliminating the general per-game problem of small sample sizes), but the idea remains the same: Figure out a way to factor a team's strength of schedule into statistics to truly find the best offense, defense, and possibly overall team. You read "+" ratings the same way you would read OPS+ or something similar in baseball: 100 is exactly average, under 100 is bad, and over 100 is good.
(If you want to look at this as a DVOA figure of sorts, just subtract 100 and view it as a percentage. So a "+" figure of 122 could also be viewed as +22 percent. A "+" figure of 90 would be -10 percent. The Overall "+" figure below is the same way, only the base is 200 instead of 100.)
First, we look at the overall offensive rankings for 2008.
|2008 Offensive "+" Rankings|
|Rank||Team||Offense S&P+||Rushing (Rank)||Passing (Rank)||Standard Downs (Rank)||Passing Downs (Rank)|
|1||Florida||152.1||152.7 (1)||157.1 (3)||156.8 (1)||147.5 (5)|
|2||Oklahoma||151.7||133.4 (6)||170.5 (1)||128.6 (6)||151.3 (3)|
|3||USC||146.3||132.5 (7)||161.3 (2)||137.4 (2)||151.1 (4)|
|4||Penn State||145.7||152.0 (2)||139.6 (6)||131.5 (5)||159.5 (1)|
|5||Georgia||135.4||117.5 (26)||153.3 (4)||136.5 (3)||140.6 (9)|
|6||BYU||130.3||118.6 (23)||136.2 (8)||121.6 (10)||135.7 (11)|
|7||Texas Tech||128.0||138.9 (4)||121.7 (15)||121.4 (11)||135.7 (7)|
|8||Oregon||127.9||142.5 (3)||109.7 (32)||131.9 (4)||128.0 (15)|
|9||Alabama||126.6||136.0 (5)||117.0 (22)||127.4 (7)||116.4 (27)|
|10||Texas||124.6||115.4 (29)||133.3 (9)||114.4 (24)||152.8 (2)|
|11||Missouri||124.4||116.8 (27)||127.4 (12)||123.8 (8)||124.11 (21)|
|12||Stanford||124.3||125.1 (10)||124.8 (13)||123.7 (9)||135.1 (13)|
|13||Oklahoma State||123.5||123.2 (13)||129.4 (11)||118.5 (14)||136.9 (10)|
|14||Mississippi||122.8||108.9 (45)||140.6 (5)||114.3 (25)||141.9 (8)|
|15||Boise State||120.5||100.2 (69)||139.5 (7)||111.5 (30)||126.2 (20)|
The most interesting name on the list is Stanford, who finished 67th in the standard Total Offense measure. The best explanation for this is that A) they faced two of the top three defenses in the country (TCU and USC) and six of the top 30, and actually put up decent yardage and point totals, and B) while they were not tremendously explosive, they had fantastic success rates, particularly rushing and on passing downs. Efficiency is clearly half of the S&P equation. This says good things about the Cardinal in 2009, as they return eight offensive starters, including quarterback Tavita Pritchard, leading rusher Toby Gerhart, and three senior starting offensive linemen. Now, if only they could figure out how to play a little defense.
Of the two measure that make up S&P, Success Rates are as a whole slightly less correlated to win percentages than Points Per Play -- suggesting that at the college level, explosiveness is slightly more important than efficiency -- but looking at the top 10 for Success Rates+ shows that for most teams, being explosive comes with being efficient. All of the teams on this list were also in the overall top 15, and all but two (Missouri and Stanford) were also in the top 15 for PPP.
|Top Success Rates+|
Here's where Oklahoma and Florida differentiated themselves. Oklahoma (first) and Ohio State (10th) were separated in PPP+ by about 47 points -- the same as the gap between Ohio State and Marshall (99th). While a handful of teams slipped over the 150 mark, the two national title game competitors were clearly head and shoulders above everybody else in the country.
|Top Offenses in PPP+|
Now let's move to rushing and passing. One thing to keep in mind here is that this is a per-play measure. That's why you see Texas Tech at fourth in the rushing list or Ohio State at 18th on the passing list. The way Texas Tech, for example, uses the run within their offense is extremely effective -- defenders are still going to be playing the pass at all times, even possibly on third-and-one -- but if they were to lean more heavily on the run, the numbers would probably suffer. Regardless, this list is only intended to look at who was most effective when they actually chose to run the ball. We will explore running games more in-depth with a measure called Points Over Expected (POE) later in the summer.
|Best Rushing S&P+|
Clearly Florida's running game will be outstanding again in 2009 with Tim Tebow, Chris Rainey, and Jeffrey Demps, but do not underestimate the loss of Percy Harvin. He was by far the most effective rushing receiver in college football last year, and while Florida is not hurting for overall weapons, he was their most explosive.
|Best Passing S&P+|
Simply because they did not compete for the national title, Georgia's season was considered a bit disappointing. The blame for that, however, cannot be pinned on Matthew Stafford, A.J. Green, and the Bulldog passing attack. Meanwhile, you would expect teams like Oklahoma, BYU, Texas, Missouri (11th), Tulsa (14th), and Texas Tech (15th) to be on the list, but the S&P+ measure rewards teams for using the pass effectively instead of just putting up big totals. Therefore you see teams like USC (22nd in overall passing yardage) and Penn State (37th) making their way into the Top 10. If you are a Jets (and therefore Mark Sanchez) fan, this should be pretty encouraging.
The concept of 3rdPV was discussed back in December. It takes the EqPts and PPP ideas and applies them to downs and distances. It is a much more static number than just PPP, with higher highs and lower lows. Therefore this will tell you who made more timely plays and leveraged opposing defenses well.
|Best 3rd PV teams|
Texas Tech made the list by staying out of uncomfortable Passing Down situations, while Texas made the list by overachieving on Passing Downs. Baylor is an interesting addition to the list -- they most likely made it due to young quarterback Robert Griffin's ability to scramble for first downs.
Be on the lookout for Griffin in 2009, by the way. He still needs more weapons around him, but he is as exciting to watch as anybody in the country.
Back in November, we looked at the ratio between success on Standard Downs and Passing Downs and wondered if a ratio that favored Passing Downs success too much would lead to failure down the line. At the time, Texas' offense had actually performed better on Passing Downs than Standard Downs, and in their lone loss of 2008 (at Texas Tech), they got murdered on Passing Downs. It just didn't seem like the kind of success a team could rely on much. (The concept of disproportionality allowed me to raise a large red flag about the Penn State-Iowa game as well, which was nice.)
(In previous VN columns, Standard Downs have been referred to as "Non-Passing Downs," but we're changing the terminology to something a little more, well, standard.)
Now that all of the "+" numbers have been calculated, let's look at the SD-PD ratio in a different way. Let's look at the teams with the biggest difference between their Passing Downs S&P+ and Standard Downs S&P+.
|Teams With Disproportionately High PD S&P+|
|Rank||Team||PD S&P+||SD S&P+||Ratio||Overall S&P+ Rank|
|8||New Mexico State||98.42||80.92||1.216||98|
Some teams were disproportionate simply because their performance on Standard Downs was so bad. But some -- like Texas, Mississippi, Oregon State and Penn State -- were good on Standard Downs and very good on Passing Downs. Teams like that could be in for a fall.
To see how legitimate this category is as a predictor, let's look at teams with the most disproportional 2007 performance and how their disproportionality translated to 2008 success. The "% Change" column below reflects the change in their overall Close-Game S&P+ from 2007 to 2008.
|Teams With Disproportionately High PD S&P+, 2007|
The conclusions here are mixed. Teams that returned star quarterbacks and major receiving threats from 2007 to 2008 (teams like Georgia, Oklahoma, and Texas Tech) improved in 2008. Meanwhile, teams with offensive deficiencies (Washington State and Washington, for example) or teams having to replace key offensive cogs (Kentucky, Wisconsin) regressed, some significantly.
Looking again at the 2008 list, Texas might be OK. They return Colt McCoy and his top weapon, Jordan Shipley, plus major big-play weapons in Malcolm Williams, Dan Buckner, and others. They still don't have a go-to running back, and while they could regress a bit, the damage might not be too bad. Meanwhile, Ole Miss returns Jevan Snead but loses big-play threat Mike Wallace, and Penn State has to replace almost their entire receiving corps. Look for a potential drop from the Rebels and a potential large drop from the Nittany Lions.
7 comments, Last at 03 Jun 2009, 4:58pm by bird jam