Writers of Pro Football Prospectus 2008

Most Recent FO Features

LongKyl13.jpg

» Word of Muth: Bears Upset 49ers

The Bears got the win and Kyle Long had the block of his dreams, but Ben Muth found plenty of room for improvement in Chicago's offensive line.

26 Nov 2010

VN: What S&P+ Sees

by Bill Connelly

Last week I listed the Top 10 performances of the season according to single-game S&P+. Predictably, the scores in nine of the 10 games were lopsided. One stood out, however: No. 4 on the list was Georgia's 44-31 victory over Kentucky on October 23. It scored a 403.7 on the single-game S&P+ scale; Georgia's offensive performance was just a hair above average, their defensive performance was, according to S&P+, outstanding.

How does a 13-point victory over an average team rank as one of the season's best performances (and help to vault Georgia into last week's Weighted S&P+ Top 10)?

I figured this offered an interesting opportunity for us to look at how S&P+ sees a game. Brian Fremeau has been doing a neat job outlining what FEI sees in a given game; now's the time for us to do the same with S&P+.

What S&P+ Sees: Kentucky vs Georgia

A glance at the box score gives you the reasonably accurate impression that Georgia's win was dictated by special teams and turnovers. Georgia used a kickoff return touchdown to build their early lead, and a plus-four turnover margin for the Bulldogs assured that Kentucky could not come back. For the game, the Wildcats outgained Georgia 423 to 290. Again ... How does this go down as an amazing, defense-powered performance?

You can follow along with the game's play-by-play here. The greatness of Georgia's performance was, in part, dictated by the S&P+ definitions of a "close" game (within 28 points in the first quarter, 24 in the second, 21 in the third, and 16 in the fourth). But we'll get to that in a bit.

Q Score Offense "Close"? Start.
FP
Plays Yds EqPts Succ.
Rate
S&P Result
1 0-0 Kentucky Yes K21 3 2 -0.99 33.3% 0.000 Fumble (5.46 T/O Pts)
1 0-0 Georgia Yes K23 4 23 3.17 75.0% 1.543 TOUCHDOWN
1 G, 7-0 Kentucky Yes K20 15 56 2.60 26.7% 0.440 FIELD GOAL
1 G, 7-3 Georgia Yes N/A 0 0 0.00 N/A N/A KICK RETURN TD
1 G, 14-3 Kentucky Yes K25 1 31 0.00 0.0% 0.000 Fumble (5.53 T/O Pts)
1 G, 14-3 Georgia Yes K33 3 4 0.26 0.0% 0.087 Missed Field Goal
1 G, 14-3 Kentucky Yes K29 4 9 0.24 25.0% 0.310 Turnover on Downs
1-2 G, 14-3 Georgia Yes K38 7 39 4.29 71.4% 1.327 TOUCHDOWN

In the first quarter, Georgia's offense basically just stayed out of the way, as their special teams scored a touchdown and their defense forced 10.8 points' worth of turnovers. Aaron Murray and the offense compiled just 7.72 EqPts of offense in the first quarter and initial plays of the second, but Georgia scored 21 actual points. Meanwhile, Kentucky ran 23 plays and managed just 1.85 EqPts.

Q Score Offense "Close"? Start.
FP
Plays Yds EqPts Succ.
Rate
S&P Result
2 G, 21-3 Kentucky Yes K23 3 8 0.29 33.3% 0.430 Punt
2 G, 21-3 Georgia Yes G22 4 19 0.57 50.0% 0.643 Punt
2 G, 21-3 Kentucky Yes K02 2 -1 -0.03 0.0% -0.013 Fumble (6.22 T/O Pts)
2 G, 21-3 Georgia Yes K05 3 5 2.01 33.3% 1.003 TOUCHDOWN
2 G, 28-3 Kentucky No K29 10 71 6.46 60.0% 1.246 TOUCHDOWN
2 G, 28-10 Georgia Yes G35 3 6 0.18 0.0% 0.180 Punt
2 G, 28-10 Kentucky Yes K08 1 0 N/A N/A N/A Downed / End of Half

The timing of Kentucky's scoring drives in this game were humorous. Another killer fumble set Georgia up for a five-yard scoring drive and took the game out of "close" status, and Kentucky immediately responded with a long touchdown drive. So at the half, Georgia's offense had racked up 10.48 EqPts to Kentucky's 8.57, but they led by 18 and had completely shut Kentucky down when the game was "close."

Q Score Offense "Close"? Start.
FP
Plays Yds EqPts Succ.
Rate
S&P Result
3 G, 28-10 Georgia Yes G30 7 70 5.55 85.7% 1.650 TOUCHDOWN
3 G, 34-10 Kentucky No K38 7 34 0.68 42.9% 0.526 Interception (4.68 T/O Pts)
3 G, 34-10 Georgia No G20 3 1 0.01 0.0% 0.003 Punt
3 G, 34-10 Kentucky No K34 10 66 5.31 40.0% 0.931 TOUCHDOWN
3-4 G, 34-17 Georgia Yes/No* K46 4 46 4.71 100.0% 2.178 TOUCHDOWN

* The first play of the drive took place in the third quarter and was therefore under "close" circumstances. Once the fourth quarter began, however, the 17-point margin was no longer "close."

A 34-yard pass from Aaron Murray to A.J. Green helped set up a short Washaun Ealey touchdown run on the first drive of the third quarter. From that point on, the game was not "close" for the rest of the quarter. Kentucky lost almost another five points to turnovers -- for the game, they would lose the turnover points battle by 21.92.

Q Score Offense "Close"? Start.
FP
Plays Yds EqPts Succ.
Rate
S&P Result
4 G, 41-17 Kentucky No K21 8 79 5.81 50.0% 1.226 TOUCHDOWN
4 G, 41-25 Georgia No G20 12 67 3.59 58.3% 0.882 FIELD GOAL
4 G, 44-25 Kentucky No K41 6 59 4.40 50.0% 1.233 TOUCHDOWN
4 G, 44-31 Georgia Yes N/A N/A N/A N/A N/A N/A UK Onside Kick
Successful
4 G, 44-31 Kentucky Yes K34 4 3 0.06 0.0% 0.015 Turnover on Downs
4 G, 44-31 Georgia Yes K37 1 0 N/A N/A N/A Downed / End of Game

Kentucky rack up 138 yards on 14 plays in the fourth quarter while down big. The Wildcats did well enough to get the game back into "close" status after 24 minutes of what S&P+ sees as garbage time. And as soon as they got the game "close" again, they went four-and-out and the game ended. Therefore, the S&P+ formulas only see a single scoring drive (the first-quarter field goal drive) and a large mess of turnovers and punts.

Under the "close" definition, this game was a statistical blowout. But to anybody who watched the game, this was not necessarily the case. Does that mean the S&P+ definitions need to be changed? I looked into the issue this week. First things first: The current definitions of "close" lead to correlations of 0.80 between S&P+ and win percentage. Getting rid of the "close" designation and counting all plays bumps that correlation down to 0.73. Clearly some garbage time designation is needed. If we expand the definitions to games within 28 points in the first three quarters and 21 in the fourth, the correlation falls to 0.76. When it comes to correlations between S&P+ and success, cutting things off as early as possible leads to better results. I expanded the definitions this season, and it resulted in little change to the correlations. I like where the definitions fall, even if it resulted in the oddest, least impressive "amazing" performance you've ever seen in Athens on October 23.

F/+ Rankings

Full rankings here.

F/+ Top 25 (After 12 Weeks)
RK TEAM RECORD F/+ LAST
WK
CHG S&P+ RK FEI RK OFF.
F/+
RK DEF.
F/+
RK
1 Boise State 10-0 +31.5% 1 +0 278.5 1 0.237 7 +14.0% 6 +17.5% 1
2 Auburn 11-0 +28.0% 2 +0 250.9 4 0.306 1 +24.8% 1 +3.3% 44
3 Alabama 9-2 +26.0% 3 +0 250.6 5 0.266 2 +16.7% 3 +9.2% 19
4 Ohio State 10-1 +25.8% 5 +1 256.0 3 0.236 9 +12.3% 10 +13.4% 4
5 TCU 11-0 +25.3% 4 -1 258.1 2 0.215 11 +8.9% 22 +16.3% 2
6 Stanford 10-1 +22.8% 6 +0 239.6 8 0.259 4 +15.9% 4 +6.9% 24
7 South Carolina 8-3 +22.3% 14 +7 247.9 6 0.206 13 +11.9% 11 +10.4% 14
8 Arkansas 9-2 +22.1% 7 -1 239.7 7 0.243 6 +15.9% 5 +6.2% 27
9 Virginia Tech 9-2 +21.3% 9 +0 232.5 12 0.264 3 +12.8% 8 +8.5% 23
10 Wisconsin 10-1 +19.0% 15 +5 229.8 15 0.231 10 +13.2% 7 +5.8% 30
F/+ Top 25 (After 12 Weeks)
RK TEAM RECORD F/+ LAST
WK
CHG S&P+ RK FEI RK OFF.
F/+
RK DEF.
F/+
RK
11 Oklahoma 9-2 +18.4% 18 +7 233.3 11 0.202 15 +8.4% 23 +10.0% 16
12 Miami 7-4 +18.4% 8 -4 235.1 9 0.192 17 +5.0% 35 +13.3% 6
13 Missouri 9-2 +18.2% 12 -1 231.6 13 0.205 14 +7.6% 27 +10.6% 13
14 Iowa 7-4 +18.1% 10 -4 234.4 10 0.191 18 +6.4% 30 +11.8% 8
15 Oregon 10-0 +18.1% 11 -4 221.8 27 0.253 5 +9.3% 21 +8.8% 22
16 LSU 10-1 +17.8% 17 +1 223.9 25 0.236 8 +6.2% 31 +11.6% 9
17 Nebraska 9-2 +16.5% 16 -1 224.4 22 0.208 12 +5.7% 33 +10.8% 11
18 Oklahoma State 10-1 +15.9% 13 -5 229.0 17 0.173 19 +11.9% 12 +4.0% 39
19 Texas A&M 8-3 +14.5% 26 +7 229.2 16 0.145 25 +1.2% 53 +13.4% 5
20 Pittsburgh 6-4 +14.3% 20 +0 223.4 26 0.169 20 +9.6% 20 +4.7% 34
F/+ Top 25 (After 12 Weeks)
RK TEAM RECORD F/+ LAST
WK
CHG S&P+ RK FEI RK OFF.
F/+
RK DEF.
F/+
RK
21 Florida State 8-3 +14.0% 22 +1 224.0 24 0.161 21 +11.5% 13 +2.6% 49
22 Michigan State 10-1 +13.9% 19 -3 224.3 23 0.156 22 +7.3% 28 +6.6% 26
23 Oregon State 5-5 +12.5% 23 +0 227.8 18 0.110 31 +9.6% 19 +2.9% 47
24 West Virginia 7-3 +12.1% 31 +7 217.7 32 0.154 23 -3.8% 81 +15.9% 3
25 Florida 7-4 +11.9% 28 +3 224.4 21 0.116 29 +3.0% 42 +8.9% 21

Considering many of the top teams had byes last week, it should be no surprise that there were almost no changes among the F/+ Top 6. With an incredibly dominant performance over Troy, South Carolina jumped back into the Top 10. (Remember: "Championship teams are generally defined by their ability to dominate inferior opponents, not their ability to win close games.") Wisconsin, Oklahoma, Texas A&M, and West Virginia all surged ahead in the rankings as well. Nobody near the top fell significantly. The biggest tumble among Top 25 teams was Oklahoma State, which apparently did not earn nearly enough style points in its blowout win over Kansas. USC also fell from 21st to 29th following its forgettable loss to Oregon State.

"What The ...?" Team of the Week

West Virginia is receiving enough votes in the AP Poll to place them 27th (they are 30th in the USA Today coaches poll), so the fact that they are 24th in F/+ should not come as much of a surprise. But the degree to which we have forgotten about the Mountaineers is impressive. As one of approximately 26 early-season victims of a tight loss to LSU, the Mountaineers moved to 5-1 and 25th with a 20-6 win over South Florida. But then they were upset at home against Syracuse and lost in overtime on a Friday night at Connecticut. They have since won two in a row against Cincinnati and Louisville. If they can knock off Pittsburgh (also surprisingly highly rated in F/+ thanks to competitive losses and a tough strength of schedule) and beat Rutgers, they'll have a decent chance at sneaking into a Big East title as long as Connecticut (which, amazingly, would control their own destiny in that scenario) loses to either Cincinnati or South Florida.

To the extent that the Mountaineers have been "powered" by anything -- they're still only 3-2 in the putrid Big East -- it has been defense. Noel Devine is still Noel Devine (despite his 4.5 yards per carry), and Geno Smith is completing 64 percent of his passes, but West Virginia ranks just 81st in Off. F/+. They have not faced too many defensive heavy hitters, and they have still scored more than 31 points just twice (and once was against UNLV). The defense, meanwhile, ranks third in Defensive F/+ and has not allowed more than 21 points in a game all season. Even in the Big East, that is damn impressive.

AP + F/+ = BCS?

Here's your weekly look at what the BCS standings would say if made up of 60 percent AP poll, 40 percent F/+ rankings:

1. Auburn (11-0)
2. Boise State (10-0)
3. TCU (11-0)
4. Ohio State (10-1)
5. Oregon (10-0)
6. Stanford (10-1)
7. Alabama (9-2)
8. Wisconsin (10-1)
9. LSU (10-1)
10. Arkansas (9-2)

Upset Watch

Kentucky over Tennessee. Spread: Kentucky +2.5 | F/+ Projection: Kentucky by 6.2. There is not yet any weight for recent games in the S&P+ or F/+ formulas, so Tennessee's three-game winning streak is tempered by its poor early play. Kentucky sewed up bowl eligibility last week and has put together enough respectable performances to rank 36th in overall F/+ despite the five losses. If Tennessee is going to knock off the Wildcats for the 26th consecutive season and get to six wins and bowl eligibility, they're going to have to earn it.

Washington over California. Spread: California -7 | F/+ Projection: California by 0.7. Yeah, good luck predicting what these two teams are going to do.

Auburn over Alabama. Spread: Alabama -4 | F/+ Projection: Alabama by 1.3. It's odd saying that the undefeated, second-ranked team in the country could pull off an upset win, but such is life for Auburn heading to Tuscaloosa this weekend. The line was at 4.5 for a while, but it is holding steady at four now.

Maryland over N.C. State. Spread: N.C. State -2.5 | F/+ Projection: N.C. State -2.8. Maryland lost a turnover-aided heartbreaker to Florida State last week, and now a Maryland win would help Florida State clinch the ACC Atlantic title. I still honestly have no idea how the Terrapins are as good as they are -- they do not rank in the Top 25 in any of FO's major categories other than Field Position Advantage -- but they will be a tough out for the Wolfpack in College Park.

The Playlist

In honor of yesterday's big dinner ...

"America Eats Its Young," by Funkadelic
"Bird Food," by Ornette Coleman
"Eat The Rich," by Aerosmith
"Eat to Live," by Talib Kweli
"Eat Your Hear Up," by The Blow
"The Food," by Common
"Gotta Eat," by Lupe Fiasco
"Jive Turkey," by the Ohio Players
"Please Don't Talk About Murder While I'm Eating," by Ben Harper
"Zombie Eaters," by Faith No More

Closing Thoughts

Seriously, how is Black Friday still a thing? Sale shopping is why the Internet was created.

OK, sale shopping and football statistics are why the Internet was created.

Posted by: Bill Connelly on 26 Nov 2010

16 comments, Last at 28 Nov 2010, 10:30pm by zlionsfan

Comments

1
by Joseph :: Fri, 11/26/2010 - 2:25pm

Hey Bill--regarding "garbage time"--check the correlations for this scenario which happens in the Kentucky/Georgia game:
Any game which has "multiple sections" of garbage time, only the "last" one counts. In other words, it's not "garbage time" until there is no more close time. For a GREAT example, think of Bills-Oilers about 20 years ago, with Frank Reich leading them back from 35-3. That would fit your definition of "garbage time", yet the Oilers' D deserves EVERY BIT of the positive points (since D points SHOULD be negative) they would get for allowing BUF to come back, and BUF's O deserves ALL of their points for making the comeback.

My feeling is this--there are only a handful of games each year that may have multiple garbage times--but if Team A's D is bad enough to let Team B's offense back into it, then they shouldn't be rewarded for it.

7
by DSMok1 (not verified) :: Sat, 11/27/2010 - 11:45pm

I absolutely agree: garbage time should be counted backwards from the end of the game.

11
by Dennis :: Sun, 11/28/2010 - 12:27pm

I also agree. A defensive should be penalized for a letting a team come back to make a game "close" after it was in "garbage time".

2
by FireOmarTomlin :: Fri, 11/26/2010 - 5:05pm

Think Auburn DEF needs a readjusted ranking

LMAO.

fark the SEC.

---------------------
Men are more ready to repay an injury than a benefit because gratitude is a burden and revenge a pleasure.

3
by Tampa Bay Mike (not verified) :: Sat, 11/27/2010 - 1:49pm

Thank goodness the Boise State nonsense is over. Here's to hoping that TCU doesn't get into the title game over a more deserving one loss team, if it comes to that.

4
by FireOmarTomlin :: Sat, 11/27/2010 - 3:25pm

And what will you say if TCU trucks the one-loss B10 team they play in the RoseBowl?

Oh wait, it will be like when Utah trucked Bama "They had nothing to play for, blah blah blah".
As if pride, draft status, recruiting etc don't matter. LMAO

----------------
Men are more ready to repay an injury than a benefit because gratitude is a burden and revenge a pleasure.

12
by Tampa Bay Mike (not verified) :: Sun, 11/28/2010 - 1:37pm

I'll say good for them, now play a tougher schedule next year i.e. accept the road games and see if you can earn a better spot. It's that simple. Your one example for Utah doesn't affect anything, anymore than BSU getting crushed by Georgia does.

6
by Bill Connelly :: Sat, 11/27/2010 - 7:07pm

Cannot disagree more. TCU is a phenomenal team.

10
by Flounder :: Sun, 11/28/2010 - 10:38am

Wisconsin/TCU should be a super interesting game.

13
by Tampa Bay Mike (not verified) :: Sun, 11/28/2010 - 1:42pm

TCU and 20 other teams would be phenomenal with their schedule, who cares. Fortunately the voters aren't going to allow this.

14
by Eddo :: Sun, 11/28/2010 - 1:57pm

Then why are only two teams phenomenal with TCU's schedule (them and Utah, recently)? I mean, I can see why people don't want to see a team like Hawaii, who comes out of nowhere to have an undefeated season, play for the title; but TCU (and Boise State) have a track record of going undefeated, beating AQ-conference teams, and winning bowl games.

And I hate to break it to you, but the Mountain West is at least as good as the Big East, possibly the ACC. Would you be opposed to an undefeated West Virginia or Georgia Tech team playing for the national title?

15
by Tampa Bay Mike (not verified) :: Sun, 11/28/2010 - 3:07pm

You counter your own schedule strength argument by mentioning Utah. I'm against any team playing a weak schedule and getting consideration for the NC. Now, since you broke it to me, I must concede that your assertion can only be true. But don't forget to reference last year's rankings of ACC teams (or was it two years ago?) for the difficulties in determining conference strength. I'm not defending the Big East or ACC here, they both suck as does the MW. It would be nice to have the NCAA mandate more connectivity between the conferences, but even then the little guys will still have to suck it up and take the road only dates when they play larger schools if they want a national title shot. It's just economics.

16
by zlionsfan :: Sun, 11/28/2010 - 10:30pm

There is zero incentive for a non-AQ team to play a more difficult schedule, as Boise State demonstrated. Even if they were able to determine in advance the quality of their opponents, if they play anything less than a full slate of AQ schools, their wins will be dismissed as flukes and their losses will be used as evidence that they don't belong. If they play only power teams on the road and beat them all, they'll be excluded because of their weaker conference schedule, and they'll also have to absorb a significant hit to their budget to play those teams.

The current system is intentionally skewed to invite only existing members to the little dance, and the people who set it up know this. It may occasionally provide us with a reasonable 1-2 matchup, but as often as not, there's a deserving AQ school that doesn't get in (usually because it's not possible in a two-team system), not to mention the non-AQ schools. Even if it were expanded to a mini-playoff, it's not going to draw the same excitement the other sports get because people like to see underdogs. Even if TCU or Boise State eventually plays for a MNC, they aren't actually "underdogs" in the sense that Butler's men's team was because we have so few chances to gauge how good they really are.

5
by James Stephenson (not verified) :: Sat, 11/27/2010 - 4:50pm

FYI, that Ken-UGA game was in Kentucky.

James

8
by John (not verified) :: Sun, 11/28/2010 - 12:37am

Miami fired its head coach today, and it struck me that I wasn't entirely sure whether it was the Heat, the Dolphins, or the Hurricanes - none of them would have particularly surprised me.

9
by Jeff Fogle :: Sun, 11/28/2010 - 1:56am

BC, appreciate you taking the time to put together the lengthy explanation of the Kentucky/Georgia game. Would have thanked you earlier but it's been a busy holiday weekend. Still think the S+P had a bad read here. Trying to think of how to explain it. Almost like not seeing the forest for the trees because a study of leaves is included. Will make for a good discussion at another time...