Writers of Pro Football Prospectus 2008

Most Recent FO Features

SmithSte01.jpg

» The Week In Quotes: August 29, 2014

This week: Josh Shaw lies, Steve Smith intimidates, Le'Veon Bell relaxes, Matt Simms dances, and Clint Trickett kisses and tells.

22 Sep 2010

FEI: Three And Out

by Brian Fremeau

After a few tumultuous weeks to start the season, last weekend played out mostly as expected. Of course, there was plenty of drama on the way. There were thrilling overtime victories from Michigan State and Auburn and another dozen games decided by a single score. Texas A&M and Michigan survived what would have been shocking upset bids from lowly Florida International and Massachusetts. And the head-to-head matchups between FEI top 30 foes (Georgia-Arkansas, Iowa-Arizona, Georgia Tech-North Carolina, Texas-Texas Tech) were all decided on the final possessions.

There is a new No. 1 team in FEI and a few shake-ups throughout the ratings. But the shift for most teams had less to do with specific results from last weekend and more to do with the continued devaluation of projected data. Many of the changes are subtle -- the ACC teams that are clearly overrated are starting to drift down the rankings -- but there have been a few major FEI moves.

The teams that have outperformed their initial projections most notably are South Carolina, Stanford, Nevada, and Michigan. That group is currently ranked an average of 25 spots better than initially projected, and their current FEI ratings are an average of .62 standard deviations above their projected value. (In the case of the Wolverines, remember that FEI disregards FCS results, and Michigan's dominant win over Connecticut and road victory at Notre Dame were both unexpectedly efficient performances). On the flip side, BYU, Connecticut, and Wake Forest are the biggest underachievers to date relative to their projected value.

And at the top, Alabama takes over the No. 1 position following a crushing win over Duke. The Crimson Tide have impressed each week and appear to have overcome personnel losses from last season's championship squad without missing a beat. That said, Alabama's schedule isn't getting any easier -- according to FEI, they're about to embark on a difficult stretch. The Crimson Tide will face five current FEI top 20 teams in their final nine games, starting with a trip to Arkansas this weekend. Based on projected win expectations, Alabama is only expected to go 6-2 against FBS opponents down the stretch (7-2 with the inclusion of their game against FCS Georgia State). As described in last week's ESPN Insider piece, a multiple-loss national championship game participant is a very real possibility in 2010.

Three and Out

I've received several requests to provide a bit more insight into the raw drive efficiency data that goes into the formula for FEI. Since we're still a few weeks away from debuting this year's Offensive and Defensive FEI data, I thought it might be interesting to take different slices each week of the raw numbers themselves. I'm calling this section "Three and Out," a set of three tables each week ranking the top teams, offensively and defensively, according to various drive splits.

This will certainly be a recurring series in this space, if not a weekly one. The idea isn't necessarily to unearth new groundbreaking drive metrics (though it might). And due to sample size, especially early in the season, we may only be able to draw general conclusions. But I'm openly soliciting feedback here on drive splits that may be of interest to you. The most popular splits may become permanent weekly data tables, or I may simply use this space to dig up quirks and peculiarities in team performances.

Let's get to it. All data provided in these tables includes only FBS vs. FBS non-garbage possessions. None of these numbers are adjusted for the strength of the opposition faced.

Three and Out Percentage
Offensive Leaders Defensive Leaders
Team Three
and Out
Off
Poss
Pct. Team Three
and Out
Def
Poss
Pct.
Arizona State 0 8 .000 Texas 21 35 .600
Boise State 1 21 .048 South Carolina 9 16 .563
Idaho 1 20 .050 Stanford 9 16 .563
Oregon 1 18 .056 Alabama 13 26 .500
UTEP 1 18 .056 Oklahoma State 18 38 .474
Kentucky 2 29 .069 Boston College 5 11 .455
Alabama 2 25 .080 Idaho 8 18 .444
Pittsburgh 1 12 .083 Western Michigan 12 27 .444
TCU 2 18 .111 Arkansas 10 23 .435
South Carolina 2 18 .111 Boise State 9 21 .429

What better way to start the "Three and Out" series than with three-and-outs? Officially, these are possessions that lasted three plays or fewer that did not result in a touchdown and concluded fewer than 10 yards from starting field position. Generally speaking, teams that limit three-and-outs also have strong offensive efficiency, but there is not a direct relationship between the two metrics.

In 2009, the Florida Gators led the nation offensively, only going three-and-out on 13.7 percent of their possessions. TCU led the nation defensively, forcing three and outs on 49.3 percent of its opponent's possessions.

Percentage of Available Yards
Offensive Leaders Defensive Leaders
Team Avail
Yards
Yards Pct. Team Avail
Yards
Yards Pct.
Indiana 510 471 .924 Boston College 739 122 .165
Nevada 1222 948 .776 Oregon 1611 356 .221
Alabama 1685 1275 .757 Texas 2629 593 .226
Stanford 915 691 .755 Stanford 1201 290 .241
Oregon 1125 818 .727 Ohio State 2177 538 .247
Navy 1416 986 .696 Texas A&M 1866 479 .257
Kentucky 1840 1270 .690 Florida International 2201 566 .257
UTEP 1208 828 .685 Rutgers 1045 277 .265
TCU 1299 871 .671 Arizona 1647 468 .284
Boise State 1468 945 .644 LSU 2564 749 .292

Available yards* are calculated based on total starting field position over the course of every non-garbage possession. A sequence of three drives begun at a team's own 10-yard line, its own 30-yard line, and the opponent's 45-yard line would have a total of 205 available yards (90+70+45=205). The table above represents the total field position gained, not necessarily the box score yardage recorded (i.e. a team's penalties count against its yardage).

In 2009, Houston's offense led the nation, accumulating 65 percent of its available yards on the season. Defensively, TCU led the nation, allowing only 24.1 percent of available yards.

*The available yards metric was first tracked and introduced to me by my friend Jay Barry of the now defunct Blue-Gray Sky Notre Dame blog. Other variations on the same theme have appeared on other college team blogs since.

Percentage of Explosive Drives
Offensive Leaders Defensive Leaders
Team Expl.
Drives
Off
Poss
Pct. Team Expl.
Drives
Def
Poss
Pct.
Alabama 12 25 .480 Ohio State 0 30 .000
Nebraska 11 29 .379 Mississippi State 0 27 .000
Oregon 6 18 .333 Boston College 0 11 .000
Iowa 7 22 .318 Central Florida 0 25 .000
Kentucky 9 29 .310 Vanderbilt 1 37 .027
Arkansas 7 23 .304 Texas A&M 1 31 .032
Oklahoma State 11 37 .297 Utah 1 30 .033
Nevada 5 17 .294 Alabama 1 26 .038
Central Michigan 6 21 .286 Middle Tennessee 1 24 .042
Michigan State 10 37 .270 Fresno State 1 24 .042

Explosive drives here represent any drive in which the average gained per play was greater than 10 yards. So a drive that covered 70 yards in 6 plays would be considered "explosive" (70/6 = 11.7 yards per play), and one that covered 70 yards in 9 plays would not (70/9 = 7.8 yards per play).

Arkansas ranked first offensively in 2009 by this metric, producing an explosive drive on 23.3 percent of its possessions. Defensively, Nebraska led the nation allowing an explosive drive on only 2.5 percent of its opponents' possessions.

Interested in seeing more "Three and Out" tables in future FEI columns? Please suggest any other drive splits that might interest you by adding a comment below or sending me an e-mail or a tweet, and I'll try to include as many suggestions as possible.

FEI Week 3 Top 25

The principles of the Fremeau Efficiency Index (FEI) can be found here. FEI rewards playing well against good teams, win or lose, and punishes losing to poor teams more harshly than it rewards defeating poor teams. FEI is drive-based, not play-by-play based, and it is specifically engineered to measure the college game.

FEI is the opponent-adjusted value of Game Efficiency (GE), a measurement of the success rate of a team scoring and preventing opponent scoring throughout the non-garbage-time possessions of a game. FEI represents a team's efficiency value over average. Strength of Schedule (SOS) is calculated as the likelihood that an "elite team" (two standard deviations above average) would win every game on the given team's schedule to date. SOS listed here includes future games scheduled.

Mean Wins (FBS MW) represent the average total games a team with the given FEI rating should expect to win against its complete schedule of FBS opponents. Remaining Mean Wins (FBS RMW) represent the average expected team wins for games scheduled but not yet played.

Only games between FBS teams are considered in the FEI calculations. Since limited data is available in the early part of the season, preseason projections are factored into the current ratings. The weight given to projected data will be reduced each week until Week 7, when it will be eliminated entirely. Offensive and defensive FEI ratings will also debut in Week 7. The FEI ratings published here are a function of the results of games played through September 18.

FEI ratings for all 120 FBS teams are now listed in the stats page section of FootballOutsiders.com. Click here for current ratings; the pull-down menu in the stats section directs you to 2007 through 2009 ratings.

Rank Team FBS
W-L
FEI Last
Wk
GE GE
Rk
SOS SOS
Rk
FBS
MW
FBS
RMW
1 Alabama 3-0 .268 2 .521 4 .128 7 9.0 6.2
2 Florida 3-0 .259 1 .195 20 .163 17 9.1 6.3
3 Oregon 2-0 .217 3 .568 2 .303 52 9.1 7.4
4 Ohio State 3-0 .216 4 .375 8 .332 54 10.2 7.5
5 Boise State 2-0 .201 8 .304 10 .537 98 10.9 9.4
6 Texas 3-0 .201 6 .234 15 .372 61 10.3 7.7
7 Virginia Tech 1-1 .187 5 .098 34 .184 24 8.0 6.6
8 LSU 3-0 .185 9 .231 16 .100 4 7.3 5.0
9 South Carolina 2-0 .178 13 .387 6 .095 3 7.2 5.5
10 USC 3-0 .176 12 .177 23 .262 45 10.0 7.4
11 Georgia Tech 1-1 .175 10 .017 49 .171 21 7.7 6.4
12 Auburn 3-0 .173 21 .097 35 .142 12 7.4 5.0
Rank Team FBS
W-L
FEI Last
Wk
GE GE
Rk
SOS SOS
Rk
FBS
MW
FBS
RMW
13 Oklahoma 3-0 .169 7 .216 18 .333 55 9.3 6.7
14 TCU 2-0 .154 20 .300 12 .687 114 9.8 8.1
15 Michigan 2-0 .152 18 .221 17 .219 32 7.6 6.2
16 Clemson 1-1 .151 11 .124 30 .161 16 6.8 5.5
17 Iowa 1-1 .146 14 .195 19 .279 48 7.6 6.1
18 Stanford 2-0 .145 31 .612 1 .239 38 7.4 5.8
19 Miami 0-1 .145 16 -.132 81 .117 5 6.2 6.0
20 Arkansas 2-0 .140 30 .152 26 .146 13 6.9 5.4
21 Penn State 1-1 .138 23 -.132 82 .147 14 7.3 6.2
22 North Carolina 0-2 .137 22 -.074 70 .175 22 6.4 5.6
23 West Virginia 2-0 .133 17 .093 36 .378 63 8.0 6.4
24 Boston College 1-0 .131 19 .187 21 .367 60 7.4 6.5
25 Nebraska 3-0 .131 29 .384 7 .483 86 8.3 5.6

Posted by: Brian Fremeau on 22 Sep 2010

15 comments, Last at 23 Sep 2010, 8:42pm by TN

Comments

1
by Kal :: Wed, 09/22/2010 - 3:23pm

On the explosive plays/available yards - does that include special team returns?

2
by Brian Fremeau :: Wed, 09/22/2010 - 3:29pm

No, these charts pertain only to offensive possessions. Special teams and defensive returns are not included here.

3
by Revenge of the NURBS (not verified) :: Wed, 09/22/2010 - 4:35pm

Does an "explosive" drive need to end in a TD? Or would a 42 yard pass, followed by 3 runs for 3 yards each, and then a punt (or even a FG), qualify as "explosive"?

4
by Brian Fremeau :: Wed, 09/22/2010 - 4:47pm

The table here includes any drive that averaged at least ten yards of field position gained per play, regardless of the outcome of that drive. A 42 yard gain, two 0-yard rushes and an interception would qualify as a 4-play "explosive" drive under this definition. (The outcome and distribution of success are obviously relevant to the efficiency of the offense. This table is more of a record of each team's frequency of drives that feature one or more explosive plays).

Good question, too. If you're interested in seeing only those explosive drives that result in touchdowns (or scores of any kind), that's the kind of thing I'm happy to provide as requested.

8
by Portmanteur (not verified) :: Thu, 09/23/2010 - 11:05am

What about some sort of "late game efficiency" table, measuring either explosiveness or available yards percentage during drives in the fourth quarter where the team was down by a score. I'm probably getting too narrow-minded here, especially since our eyes should already be pretty good at telling us which teams are capable of coming back in the final minutes.

Maybe a table showing the number of drives that resulted in points while the team was down by less than a score. Or number of game-tying or go-ahead drives when down by a score, 4th quarter or otherwise.

5
by Jeff Fogle :: Wed, 09/22/2010 - 4:48pm

Could an "explosive" drive lead to a concussion that would make a person think the Texas/Texas Tech game was "decided on the final possession?"(lol). Some confusion there I think BF...Texas scored with 9:26 left to go up 24-14. Both teams had the ball a few times after that...

6
by Brian Fremeau :: Wed, 09/22/2010 - 5:06pm

A sloppy way of grouping that game with the others, I agree. I meant to imply it wasn't a blowout and that the outcome wasn't ultimately decided until the game's last few possessions.

7
by jgrenci@zoomint... :: Wed, 09/22/2010 - 9:30pm

Hello Brian, I have a question regarding three and out

you state

'I've received several requests to provide a bit more insight into the raw drive efficiency data that goes into the formula for FEI.'

from this, I gather that three and out values are part of the formula for FEI ratings, but on the page where you describe the FEI formula, I see no mentioin of it.

is it used in the formula?
thanks

The Wizard

12
by Brian Fremeau :: Thu, 09/23/2010 - 3:04pm

Actually, three and outs are not an explicit component of the FEI ratings. My intention for providing "more insight into the drive efficiency data" in these tables and others in future FEI columns is to parse the raw drive data down in other unique ways.

Collectively, three-and-outs, percentage of available yards, and many other perspectives on raw drive data are all contributing factors to an individual team's Game Efficiency, Offensive/Defensive Efficiency, etc. But they aren't factored individually into FEI.

9
by Joseph :: Thu, 09/23/2010 - 11:15am

Brian, in your opinion, based on your rankings, is the SEC champ pretty much assured of a spot in the BCS championship game? Or is it too early to tell?

11
by mm (not verified) :: Thu, 09/23/2010 - 2:08pm

Alabama and Florida are the 2 best teams according to his ratings, but they're only projected to get to 10 wins. That means the SEC champ could easily be the best team in the country, but have too many losses to get in the top 2 BCS spots. The current rankings do agree with the SEC's self-impression: there are a lot of excellent teams (5 of the top 12) that will beat up on each other before the end of the year.

Of course, these rankings can change as the year goes on.

13
by Brian Fremeau :: Thu, 09/23/2010 - 3:06pm

What mm said. It will be very interesting if we are debating the relative merits of a two-loss SEC champ and an undefeated Boise State and/or TCU at season's end. Not just FEI, but I think the polls will struggle with it as well.

10
by TomahawkNation (not verified) :: Thu, 09/23/2010 - 12:46pm

Can you include the worst as well from the BCS conferences?

14
by Brian Fremeau :: Thu, 09/23/2010 - 3:53pm

I debated whether to include 5 or 10 worst in these tables. Here's a bottom five in each category:

Worst BCS Conf Team Offenses - 3 and Outs
No. 116 Washington State (.500)
t-No. 111 Wake Forest (.455)
t-No. 111 South Florida (.455)
No. 109 North Carolina (.440)
No. 108 Rutgers (.438)

Worst BCS Conf Team Defenses - 3 and Outs
No. 118 Tennessee (.100)
t-No. 114 South Florida (.111)
t-No. 114 Arizona State (.111)
No. 112 Indiana (.125)
No. 111 Wake Forest (.130)

Worst BCS Conf Team Offenses - Available Yards
No. 118 Rutgers (.179)
No. 109 Cincinnati (.273)
No. 108 Washington State (.284)
No. 103 Tennessee (.324)
No. 102 Boston College (.327)

Worst BCS Conf Team Defenses - Available Yards
No. 116 Duke (.645)
No. 115 Wake Forest (.642)
No. 112 Maryland (.633)
No. 102 South Florida (.554)
No. 101 Oregon State (.553)

Worst BCS Conf Team Offenses - Explosive Drives
t-No. 109 Miami (.000)
t-No. 109 Minnesota (.000)
t-No. 109 Arizona State (.000)
t-No. 109 Indiana (.000)
t-No. 106 Colorado (.032)

Worst BCS Conf Team Defenses - Explosive Drives
No. 120 Duke (.348)
No. 118 Wake Forest (.304)
No. 114 Washington State (.250)
No. 113 Cincinnati (.240)
No. 107 Maryland (.227)

15
by TN (not verified) :: Thu, 09/23/2010 - 8:42pm

Brian you are a machine! Thanks!