#3 on Lindy's Five Essential Websites (Non-Major Media) for 2013
[+] Team Summaries

Friday, November 30, 2012

Championship Week Preview, Part II

For Part I, click here.  To see what NEPA is all about, click here.

Nebraska (10-2) vs. Wisconsin (7-5)

Nebraska passing attack (82.5 NEPA; .261 NEPApp) vs. Wisconsin pass defense (40.4; .102)
Nebraska rushing attack (39.1 NEPA; .070 NEPApp) vs. Wisconsin rush defense (-66.5; -.172)
Wisconsin passing attack (62.5 NEPA; .245 NEPApp) vs. Nebraska pass defense (-45.4; -.129)
Wisconsin rushing attack (12.6 NEPA; .024 NEPApp) vs. Nebraska rush defense (-41.2; -.087)
CFBTN Projection: 29-25, Nebraska

Maybe we should be glad Ohio State cheated.  Nebraska faced both Ohio State and Wisconsin in the regular season, and their loss to OSU was far less captivating (63-38) than their win over the Badgers (30-27).  One might think we could be in for another close one.

+.182 when Nebraska passes
-.051 when Nebraska rushes
+.058 when Wisconsin passes
-.032 when Wisconsin rushes

NEPA certainly favors Nebraska in this one.  As a team, Nebraska has over 200 NEPA this season.  Wisconsin has barely 100.  Really, the game should not be as close as the first matchup turned out.  In fact, the closeness of the first matchup was likely an aberration in itself.  In that game, Nebraska gained 24 1st downs to Wisconsin's 17.  Nebraska outgained UW 440-295.  Turnovers and penalties dug Nebraska into an early hole, down 20-10 at halftime, but they came back to win with a dominant 2nd half.  That 2nd half was closer to the true talent levels of both teams.  I'm hoping this game is close, because close games are more fun.  Unfortunately, I'm predicting a Nebraska blowout.  Think of two second halves from the first matchup.  That would come out to a 40-14 score, and that's not far off from what I'm expecting.  If Wisconsin wins, they'll have the lowest winning % of a Rose Bowl participant in over 80 years.  If NEPA is any indication, that streak will survive.

Preview: Alabama vs. Georgia


Petey Surber
Guest Blogger

For the second straight year, the Georgia Bulldogs will represent the SEC East as underdogs in the SEC Championship Game. Unlike last year, they’re eyeing a spot in the national title game themselves.
Call it a national semi-final game.

Over the years, no conference championship game has grown in popularity, importance, and pageantry the way the SEC championship game has. And this year the stakes couldn’t be any higher. Not only will the winner go on to face undefeated Notre Dame to compete for a national title, but the loser will likely fall out of a BCS Bowl, much to the dismay of Alabama Coach Nick Saban. No surprises here, but everything runs through the SEC once again.

How did they get here?

Friday's trip around the statverse

Brian Fremeau quantifies the luck of the Irish. Turns out they have been about as lucky as any team that finishes the regular season undefeated.

Jack Moore examines the efficacy of going for two to avoid overtime. He focuses on Wisconsin, but his conclusion could elicit a collective grown from Nebraska fans, particularly those more than, say, 28 years old.

ESPN Stats & Info argues that Manziel is the best scrambler in the business

Thursday, November 29, 2012

Championship Week Preview, Part I

I'm providing a NEPA-centric preview for each conference championship game, along with a few other games which will factor heavily into conference titles.  This is Part I, which will deal with Louisville/Rutgers, which will serve as the Big East Championship Game on Thursday, Friday's pair of title games for the MAC and Pac-12, and Saturday's Conference USA championship game.  As usual, I'll be using NEPA, which stands for Net Expected Points Added.  NEPA is explained here and requires no special math skills to understand (try that with QB Rating).  I'm also changing up the format of the statistical matchups.  Now I'm pairing up the units that will actually be facing off with one another.  Enjoy!

Louisville (9-2) @ Rutgers (9-2)

When Louisville Passes: Cardinals Pass O (145.2 NEPA; .388 NEPApp) vs. Knights Pass D (-6.5 NEPA; -.019 NEPApp)
When Louisville Rushes: Cardinals Rush O (5.5 NEPA; .015 NEPApp) vs. Knights Rush D (-88.7 NEPA; -.234 NEPApp)
When Rutgers Passes:  Knights Pass O (49.3 NEPA; .152 NEPApp) vs. Cardinals Pass D (44.2 NEPA; .139 NEPApp)
When Rutgers Rushes: Knights Rush O (-47.6 NEPA; -.130 NEPApp) vs. Cardinals Rush D (-10.4 NEPA; -.027 NEPApp)
CFBTN projection: 23-17, Rutgers

This isn't technically the Big East Championship Game, but it effectively is.  Such an official game does not exist, and officially, if Louisville beats Rutgers, the two will be co-champions with Syracuse.  However, the winner of this game will advance to the BCS, which makes this the Big East Championship Game.  For this week's games, every team I'm previewing has two things in common: they're all good at winning football games (ok, I suppose I'll have to preview Georgia Tech in Part II, but for the most part this is true), and they've all lost at least once this season.  So, one way to see what to expect would be to look at how these teams win, and how these teams lost along the way.

Week 13 NEPA rankings

A quick look at the leaderboards of our NEPA database.

Quarterbacks

  1. Johnny Manziel, Texas A&M (206.0)
    68%, 3419 pass yards, 1181 rushing yards, 43 total TD, 8 INT
    Manziel, in his last chance before the vote, might have clinched the Heisman on Saturday.  Against Missouri, he had his 2nd highest NEPA output of the season (28.7), completing 73% of his passes for 372 and 3 TD, with 67 yards and 2 TD on the ground as well.  Amazingly, with more media attention, he has only gotten better.  And remember, he has missed roughly 12 quarters of play this season.
  2. Jordan Lynch, Northern Illinois (200.3)
    64%, 2750 pass yards, 1611 rushing yards, 39 TD, 4 INT
    With two games to go, Lynch has a very minor chance, but a chance nonetheless, to pass for 3000 yards and run for 2000.  His season has been utterly magnificent. 
  3. Nick Florence, Baylor (178.5)
    61%, 3825 pass yards, 460 rushing yards, 38 TD, 13 INT
    Florence topped 20 NEPA for the 5th time this weekend against Texas Tech.
  4. Tajh Boyd, Clemson (172.2)
    67%, 3550 pass yards, 492 rushing yards, 43 TD, 13 INT
    Boyd slipped in the rivalry loss against South Carolina.  At -2.3, it was his first negative NEPA of the entire season.
  5. Colby Cameron, Louisiana Tech (163.7)
    69%, 4147 pass yards, 177 rushing yards, 35 TD, 5 INT
    Cameron ranks #1 in the country in something very important: Minimum NEPA.  His lowest NEPA of the season is 5.5.  That ends the regular season as the best "worst game" for

Wednesday, November 28, 2012

A closer look at yards per play

The table below charts the percent of plays by the yardage. Ball State leads the nation with the fewest negative yardage plays, followed by Baylor, Penn State, Michigan State and Louisiana Tech. Maryland, Auburn, Memphis and Florida are the most likely to go backwards. Georgia has the most plays over 20 yards, ahead of Texas A&M, Clemson and Oklahoma State. Navy, Air Force, Georgia Tech and Army rank 1 through 4 in plays that go for 3-6 yards - you might notice a pattern. Those are also the four teams that are least likely to go for no gain on a play.

Alabama

Click here for an explanation

Alabama CFBTN 1
Alabama CFBTN 2 Alabama CFBTN 3 Alabama CFBTN 4 Alabama CFBTN 5 Alabama CFBTN 6 Alabama CFBTN 8

Alabama CFBTN 9
Alabama CFBTN 15
Alabama CFBTN 13 Alabama CFBTN 14

Tuesday, November 27, 2012

College Football Conundrums

A continual problem in ranking college football teams is dealing with cyclical win relationships.  No matter the ranking, there are always a number of ranking violations where Team A is ranked ahead of Team B despite the fact Team A also lost to Team B.  For example, we see Alabama ranked ahead of Texas A&M in almost every ranking.  The BCS, AP Poll, and all computer rankings are filled with these ranking violations, conceivably because the loss was viewed as an anomaly or there are circular relationships (X beats Y beats Z beats X) making a violation unavoidable.  While rankings do exist that seek to minimize the number of these violations (see Coleman's MinV), this is not viewed as a serious problem for most computer rankings on the grounds that each match up is probabilistic.  For example, the team panel for Alabama on cfbtn shows Alabama is likely to win the SEC Championship, but not certain, and a Georgia win would not disprove the validity of the approach.  Instead it merely suggests that if the two teams were to play one another over and over, Alabama would most often come out on top.  Last year's pair of games between LSU and Alabama illustrate that we cannot make deterministic predictions about game outcomes given the amount of noise (weather, injuries, emotions, referees, etc.) in college football.  

The ranking violations, however, are born of an interesting potential problem of cyclical interrelationships between teams.  In some situations, there is literally no way to avoid a ranking violation because X defeated Y who defeated Z who defeated X, creating a loop of wins.  The network of inter-relationships within the SEC demonstrates the complexity of this problem to our ranking schemes.  We see LSU defeated Texas A&M who defeated Alabama who defeated LSU.  To add to the problem, Florida defeated both LSU and Texas A&M, but lost to Georgia, who did not play any of the three teams in the cyclical triangle.

Unraveling this web of wins is difficult.  Pollsters use their impressions of performance.  Most computer rankings would use statistics from the individual games (i.e. total yards) to sort it all out.  The Network Rankings I discussed on cfbtn last week would more or less be stuck.  By only looking at wins and losses, we essentially have a 3 way tie between LSU, Texas A&M, and Alabama.

We can solve this problem, regardless of our approach, by incorporating more information, and this is one reason behind the superiority of rigorous methodological approaches to polls that merely gauge impressions.  Rigorous computational approaches allow the incorporation of more detailed and extensive information, providing us with a tool to unravel these ranking conundrums in a more effective manner than Kirk Herbstreit's "look test."  To illustrate, I'll use my own Network Ranking as an example.  To unravel the conundrum I reproduce the above illustration, but this time include all first order relationships teams, or teams that each of the five SEC teams have played.


Quickly, we begin to see why Florida is ranked highest in the Network Ranking and a number of other rigorous approaches.  They seem to have the strongest set of inter-relationships - meaning more wins and more proof of their goodness.  We also notice a few unique ties relevant to individual SEC teams from their non-conference (and non-Sun Belt) opponents.  Florida has Florida State, Georgia has Georgia Tech, LSU has Washington, and Alabama has Michigan.  These important ties might provide us with further information to unravel the conundrum, so by taking another step and including second-order relationships, or teams that opponents of the five SEC teams have also defeated, we may be able to unravel the ranking problem further.



This second order diagram gets very messy, very quickly, but it illustrates the extra-conference linkages of the SEC, with Alabama connected to the Big 10 through Michigan, Florida and Georgia connected to the ACC (Florida more strongly due to Florida States superiority over Georgia Tech in the conference), and LSU's to the PAC12.  In the Network Rankings, calculating these linkages through wins and losses provides their relative rank.  In most other computer rankings, the underlying concept is the same - comparing relative schedules to weight statistical observations of total yards, defense, etc.  Regardless of the approach, we require some form of mathematical rigor to at minimum organize the information and compare evidence in favor of each team, as no matter how sophisticated the expert doing the looking, the "look test" is limited in its ability to compare relative performance by the limited amounts of information the expert polled is capable of considering.

Ranking violations are not an inherent problem for any ranking because we have more than just that one game where Texas A&M defeated Alabama to generate a rank.  We have observations from other games that allow us to determine whether or not Alabama has demonstrated that it should probably be lower ranked than A&M or if that one loss is somewhat anomalous.  Ranking violations are unavoidable, particularly within conferences where teams have similar schedules, but careful use of evidence allows for sorting out the college football conundrums.




Coaching Carousel: California

Out: Jeff Tedford (Fired)
Record: 82-57
Record, previous 139 games: 60-78-1

Tedford's firing comes with a lot of disagreement and a bit of surprise.  Aaron Rodgers has already voiced his disapproval, and Cal hasn't had a more successful coach in my lifetime (or yours, most likely).  Like most of you, I've known two Cals - the Cal before Tedford, the doormat of the old Pac-10, and the Cal with Tedford, a perennial contender who never could quite get over the hump, but still was good for a bowl game nearly every year.

But as we're seeing more and more in college football, even successful tenures can grow stale.  And while California's AD cites graduation rates as a major reason for the firing, graduation rates didn't seem as troubling until the team went 3-9 this year.  So while the AD says it's not all about football performance, I think it's safe to assume that it indeed is.  With our NEPA database going back to 2005, we can examine almost the end of Tedford's tenure, which began in 2002.  Tedford, then a new head coach, went 25-13 over his first 3 years, with '04 being arguably his best - a 10-2 season with a #9 finish in the polls.  Since that 2004 season, what has happened?  First, the offense.



Monday, November 26, 2012

Week 13's Best Offensive Performances

I've said enough about the Aggies recently, so I'll keep it brief. In their last four SEC games (Auburn, Alabama, Mississippi State and Missouri - all teams that have been solid to spectacular on defense this season) only Mississippi State has kept Texas A&M under 20 points in the first quarter (the Aggies scored 14 in that game). With this week a microcosm of the season, Texas A&M has the best offense in the country, followed by Oregon, and then everyone else tries to play football as well. Johnny for Heisman.

Alabama's win over Auburn was arguably the most lopsided game this season. Georgia/Georgia Tech was pretty lopsided as well. USC and Notre Dame both make the top 20 despite scoring only 35 total points because 1) they both are good defensively and 2) they only had about 10 full possessions; Notre Dame scored on half.


Week 13's Best Defensive Performances

It might seem odd that the best defensive performance of the week allowed 26 points, but 1) Florida State has averaged 41.5 points and 477 yards per game this season, 2) Florida turned FSU over five times and 3) FSU's final touchdown was meaningless. The most dominant defensive performance, ignoring the caliber of opponent, came from Utah State, who turned Idaho over on three of their first five possessions and allowed only 210 yards. Idaho did manage three solid drives for 9 points in the middle of the game, but was otherwise hopeless.


Sunday, November 25, 2012

Coaching Carousel: Tennessee

Out: Derek Dooley (Fired)
Record: 15-21
School record, previous 36 games: 23-13
Let's look at Tennessee's NEPA over the Dooley era, but as a frame of reference, we'll include the 3 previous years as well.  So for both graphs, the first three plot points are for Fulmer/Kiffin.  The last 3 will reflect Dooley's performance.:

Let's start with the offense (data indicates NEPA):



My Heisman ballot, Bama/Georgia play for two, and 12-0 in Columbus

If I were in control of the Heisman, here's what the top five would look like.

1) Johnny Manziel - He just had the best season in the history of the SEC, which has not only been the best conference in the country for the last seven years, but it excels defensively.
2) Johnny Football - Who else can strike the pose while stiff-arming a future first round draft pick to the ground before evading 17 tacklers to find the open receiver in the end zone?
3) Scooby Doo Johnny - Because even in a scooby doo costume he's still the best player in the country.
4) Fake Identity Johnny - At least he didn't steal his roommate's laptop
5) Johnny's right knee - a little Rocky, a little Mateen Cleeves, Willis Reed, virus-laden Jordan, etc.

Friday, November 23, 2012

Week 13 Preview

As usual, I'll pick 5 games I'm personally interested in watching, and then I try to convince you to be interested in them too.  Here are the 5 for Rivalry Week!  There's football on Friday, but the best games have been saved for Saturday.

Also, as usual, I'll be discussing NEPA (Net Expected Points Added).

Notre Dame @ USC

Quarterbacks: Everett Golson (50.9 NEPA; .152 NEPApp) vs. Max Wittek (???)
Running Backs: Cierre Wood (24.1; .229) & Theo Riddick (15.5; .075) vs. Curtis McNeal (16.0; .139) & Silas Redd (6.0; .041)
Defenses: Fighting Irish (-23.0, -.061 vs. Pass / -81.3, -.253 vs. Rush) vs. Trojans (30.8, .077 vs. Pass / -23.5, -.056 vs. Rush)
CFBTN Projection: 23-23

I picked this game because all eyes will be on Notre Dame.  As someone under the age of 30, I've grown up in an era when hating Notre Dame just didn't come naturally.   I was 10 the last time the Irish were ranked #1 in the country.  It's been nearly a quarter century since Catholics vs. Convicts and their '88 title.  So for me, it's actually fun to see Notre Dame back near the top.  They say a strong Notre Dame is good for college football.  I think they're right.

Now, I'm talking about Notre Dame's place in the world because, well, the matchups don't really indicate that this one will be close.  A month ago, this game was billed as a super offense against a super defense.  Notre Dame still has a super defense, but USC will be without QB Matt Barkley (96 NEPA this year) this weekend.  Nothing against Wittek, but USC is going to miss Barkley.   Despite the disappointing year, Barkley has at times been brilliant, brilliant enough to have led USC to a win in this game.  The good news for Wittek is that he'll still have a star receiving corps to throw to, led by Biletnikoff contender Marqise Lee.

CFBTN vs. Vegas, week 13

In recognition of the holidays, I am foregoing my usual Friday post and moving the Saturday post up a day. 

This week, there are five games about which Vegas and the model are more than 4 points apart. That Maryland is again on the list goes to show that the model still hasn't learned that the fifth string quarterback isn't actually a thing. The model was also expecting Matt Barkley to be playing, which explains the varying expectations for USC/Notre Dame. The model has been a Duck doubter all season - it isn't impressed by flashy uniforms - until last week. Oops. And the model likes Iowa State and Georgia because it "thinks" they are good at football.


Thursday, November 22, 2012

Week 12 NEPA rankings

Week 12 saw the fall of #1 and #2, records falling at the hands of Johnny Football, and some pretty incredible MAC-tion.  Oh, and Tavon Austin tried to destroy our computers.  Here's how NEPA sees the top 10 at each position.  I'm presenting this in a new way, a top 10 for each position:

Quarterbacks

  1. Jordan Lynch, NIU (190.1 NEPA; .384 NEPA per play)
    2582 passing yards, 1504 rushing yards, 38 total TD, 4 INT
    MAC football is fun this year.  No, it might not be the greatest conference in the land, and no, these teams probably wouldn't win the SEC or Big Ten.  That's not the point.  The point is that the MAC is fun. Jordan Lynch is a major reason why.  On Wednesday, in what was basically the MAC West title game, Lynch passed for 407 yards and 3 TD - and ran for 162 - in the 31-24 win.  The guy is pretty much single-handedly winning the conference by himself.  According to NEPA, Lynch's performance was over 20 points better than the average player.  As someone who looks at NEPA on a regular basis, I can tell you that anything over 20 is extremely good.  Now, here's the kicker - it was only his 5th best game of the year.  I'm not saying Lynch should definitely win the Heisman, but he definitely should be in the conversation, and it's a shame he's not.
    Fun Fact: Lynch is averaging over 60 yards per punt this season.

Coaching Carousel: Western Michigan

Out: Bill Cubit (Fired)
Record: 51-47
School Record, previous 98 games: 48-50
Western Michigan NEPA under Bill Cubit:



Wednesday, November 21, 2012

Forecasting the 2012 BCS title game

To forecast the 2012 BCS title game participants, I first listed the seven top contenders in order:

1) Notre Dame
2) Alabama
3) Georiga
4) Oregon (Pac-12 champs)
5) Florida
6) Oregon (not Pac-12 champs)
7) Kansas State
8) Florida State

What if we had a tournament?

Using the BCS standings, I simulated standard four and eight team tournaments (1 vs. 4, 2 vs. 3, etc.). In each column is the percent chance that the team would win in that round.

 The four team tournament essentially reflects reality. Alabama and Georgia are playing a semi-final in the SEC championship game. Notre Dame has a play-in game against USC. Florida needs a little help in real life, but a win over Florida State will put them in a good position if Notre Dame or Alabama and Georgia stumble. Again, as in real life, Alabama would have the best odds in a tournament.

In the eight team variant, Alabama, Oregon, Notre Dame, and Kansas State would have the best chances of getting through round 1, and Alabama and Oregon each have about a 50% chance of getting to the title game.




Tuesday, November 20, 2012

The Network Rankings

by J. Patrick Rhamey Jr., PhD

            There are myriad college football rankings, and, as Scott has noted, the statistical choices that go into each are inherently subjective.  I am going to propose one more ranking to the many that already exist, but as the below comparison demonstrates, most existing rankings are currently converging toward what the ranking I describe below already listed last week (Notre Dame 1, Alabama 2), providing a compelling case in favor of my method.  The subjective choice about the relevant data to include is simple: none of it.  To create my rankings I use only wins and losses, ignoring statistics such as margin of victory, total yards, turnovers, etc. for two reasons. First, we are asking an enormous amount from our data if we are generating predictions of probable team performance based upon a limited number of observations (at this point, 10 or 11 games).  Second, statistics used in existing rankings are incomparable across games given the dramatically different contexts.  If two teams both have 300 total yards of offense, but one team played in the snow in Ann Arbor while the other played on a sunny day in the Rose Bowl, we are essentially equating completely incomparable numbers to generate relative rank.  Margin of victory is perhaps one of the more egregious variables, with large numbers frequently a better indicator of poor sportsmanship, personal grudges, or heated historical rivalries than a decisive indicator of excellence.

Fantasy Bumper Crop, Week 13

A fantasy bumper is a player that should perform significantly better than average this week. There are generally two reasons for this: 1) an established player has a good matchup or 2) poor performance or injury to teammate will give the player more opportunities. In the first case, these players are already owned so you want to make sure they are in your lineup. In the second case, they could be a valuable pickup. Since this is a stats-based site and not a fantasy site, and I maintain this site on top of my day job while trying to raise a son and keep my wife from realizing she could do a lot better, these results are based on the best, most recent stats available to me and nothing more. It falls on you to do your own research. You can get fantasy rankings here.

He might not have been able to beat Marshall, but Crawford Jones threw for 316 yards, 5 TDs, and no INTs Saturday. And Tulane is no Baylor, if you know what I mean (yeah, Klein, that one's for you). The rest of the list is made up of the usual suspect opponents - Akron, Miami, NMSU, Idaho, etc. Jalen Whitlow has high expectations against a Dooley-free Tennessee, and Connor Shaw will need to do a number on Clemson.

Monday, November 19, 2012

Week 12's Best Offensive Performances

The EP3, or effective points per possession, measure the average number of "effective" points earned each drive during a game. Effective points consider field position as well as actual points. The EP3+ is the team's average performance over the course of the season adjusted for the opposition. The EP3+D is the defensive score while the EP3+O is for the offense. The Perf-O begins with the single-game EP3 and adjusts for the opponent's EP3+D. The Perf-D begins with the opponent's EP3 and adjusts it for the opponent's EP3+O.

Scoring 42 points on Duke might not seem that impressive, but doing it on 8 full possessions is. Georgia Tech, though, was outshone this week by Baylor, who scored 52 points on 13 possessions, and not against Duke but against the #1 team in the country. Texas A&M finished scoring two plays into the second half, but already had 47 points. Oregon was still better than 100 other teams nationally on offense despite scoring only 14 points and losing. 


Week 12's Best Defensive Performances

The EP3, or effective points per possession, measure the average number of "effective" points earned each drive during a game. Effective points consider field position as well as actual points. The EP3+ is the team's average performance over the course of the season adjusted for the opposition. The EP3+D is the defensive score while the EP3+O is for the offense. The Perf-O begins with the single-game EP3 and adjusts for the opponent's EP3+D. The Perf-D begins with the opponent's EP3 and adjusts it for the opponent's EP3+O.

It goes without saying that Stanford had a very impressive performance on defense Saturday. Holding the Ducks to 14 points through regulation and an overtime is nothing short of spectacular. But the Cardinal were outdone by Vanderbilt, who completely destroyed Tennessee and sent Dooley packing.

It is an indication of just how bad the Wake Forest offense is that Notre Dame gets the shutout but placed eight this week, behind two SEC schools that played FCS opponents. And, in the greatest shocker of all, Baylor, the worst defense in college football (and I'm only exaggerating a little) finds itself in the top 25 after keeping Klein in check.



Sunday, November 18, 2012

Up to date, week 12

CFBTN is now up to date through week 12. NEPAs may need up to an hour to update on the server.

The fraudulent Ducks, Klein throws it away, Johnny Football hauls it in, and SOMEONE will win the national championship

Oregon finally faced a stiff run defense and scored 14 points on 15 possessions. Based on what I've heard from those criticizing me for saying that Kansas State had a better offense, I can only assume Oregon benched its starters before the game even started. Except (former) Heisman candidate Kenjon Barner managed 66 yards on 21 carries, so he must have been playing.

But I can't really brag because Kansas State scored 24 points on the worst defense in the world. (Former) Heisman candidate Collin Klein racked up 2.3 yards per carry and 286 passing yards on 50 attempts. Three of those passes were caught by Baylor defenders. Sam Houston State is significantly better on defense than Baylor. Really. No lie. I'm not even joking. And Manziel put up Klein's numbers, except more rushing yards, more touchdowns and fewer interceptions, in a half and a play.

What does this all mean? First, it means that if Johnny Manziel does not win the Heisman the award needs to be discontinued. His only bad game of the season (Ole Miss) he led the Aggies to a come-from-behind win. His mediocre game (LSU) was against one of the best defenses in the country. Klein just laid an egg against Baylor. I could score 40 on Baylor. And Barner just got 3.1 yards per attempt in a loss that might have cost his team a chance at the national championship, the Pac-12 championship and even the Pac-12 North. The rest of the pack is welcome to make the trip to New York, but the trophy belongs to Johnny.

Second, the MNC game picture is as muddled as ever. Alabama has gone from only mostly dead to living mostly dead. Alabama and Georgia are playing a semifinal in Atlanta. Notre Dame has a play-in game against USC. If Notre Dame loses, Kansas State should take the lead for the second MNC game spot.  But Oregon should slide past Kansas State if they beat Oregon State and UCLA, but Oregon will need Stanford to lose to UCLA or we will have a Stanford/UCLA rematch while Oregon watches. If Kansas State and Notre Dame lose but Stanford does not, we could have a second non-division winner playing for the MNC. If Notre Dame, Kansas State and Oregon all lose, and Florida beats Florida State, we will have another SEC-only MNC.

Third, we learned (again) that assuming currently undefeated teams will stay that way going forward is a big, big mistake. Even if a team will be the better team on the field in every game, they will still probably lose at some point. That goes for Oregon, Alabama, and even Johnny Football.

Friday, November 16, 2012

CFBTN vs. Vegas, week 12

There are three games this week in which Vegas and the model disagree by more than 4 points. The model has a hard time understanding that fifth string quarterback isn't really a thing, and that the Terps have been slow roasted. But the model also sees that FSU came off a bye and looked like horse manure. The model, Vegas, and even Jim Grobe agree that Wake Forest won't score against Notre Dame, but the model has more faith in the Irish offense than anyone with eyeballs and NBC. Finally, the model likes Ole Miss. It sees the Rebs scoring 21, which is weird since only one team has topped 21 this season against LSU, and that'd be Towson. 


Week 12 Preview

Man, week 12 already.  This season seems to have flown by.  Ok, I'm guessing you're not here just to read me wax nostalgic about the slow passage of time.  You're here for the Week 12 preview!  If you're not, you've gotten lost on the internet.  Read it anyway.

This is where I subjectively select the 5 most interesting games for this weekend and write a preview for each.  I'll reference NEPA a lot.  Check it out so you'll see what I'm talking about.  NEPApp is just NEPA per play.

USC @ UCLA

Quarterbacks: Matt Barkley (92.5 NEPA, .248 NEPApp) vs. Brett Hundley (84.7, .193)
Running backs: Curtis McNeal (11.7, .131) & Silas Redd (6.9, .048) vs. Johnathan Franklin (45.9, .202)
Defense: Trojans (16.0, .044 passing / -23.7, -.063 rushing) vs. Bruins (33.4, .089 passing / -58.7, -.158 rushing)
CFBTN projection: 32-30, Trojans

This game serves as the de facto Pac-12 South Championship, with the winner likely headed to Eugene for the conference title game.  USC has the star combo in Matt Barkley and WR Marqise Lee (who leads his position nationally in NEPA).  However, UCLA has the better defense.  A stout Bruin defense does not bode well for the generally underwhelming duo of McNeal and Redd.  I would say USC will slow down Franklin, but they did a particularly poor job of slowing down Kenjon Barner.  Franklin is a different style of runner - to borrow a tired old phrase he's more thunder than lightning - but I'm still not convinced the Trojans will really stop him.  And I haven't mentioned the underrated Hundley.  UCLA's offense doesn't get the national press, and it doesn't have a current or past Heisman darling like USC, but it's very good, possibly better than USC - and going against a worse overall defense.  Still, UCLA's one weakness is a glaring one, especially in this matchup: pass defense.  They're not terrible at it, but "not terrible" won't be good enough against Barkley and Lee.  Lee is such a unique WR - every opponent knows he's getting the ball when Barkley drops back, yet they can do nothing to stop him.  There's no reason to think UCLA will fare better defending Lee than anyone else.  Like Scott's projection, I see this being a close game.  UCLA should force USC to be one dimensional on offense, but they'll struggle to stop that dimension.  This column is more about the preview than the prediction, but heck, I've gotten into the habit of making picks each week, and so does everyone else on the internet, so I might as well.  Elite talent often serves as a tiebreaker for me when a game is too close to call, so I'll pick USC in a close one thanks to the Heisman skills of Marqise Lee.

Ohio plays football, so does the west coast and Louisiana, honoring FCS week, and unnecessary Johnny Football references

The state of Ohio has eight FBS programs, second only to Texas (where Johnny Football lives). Ohio has more rosters to fill than recruiting hotbeds California and Florida. Of the seven teams playing this week, only Miami has a losing record, and even Miami has an 11% chance of becoming bowl eligible. (Akron is terrible, but they have the week off. Akron lost to UMass, but still managed to score 26 on Tennessee; yeah, Dooley should be getting his resume ready.)

Five teams in Ohio had two or fewer losses before games started this week (like Johnny Football's Aggies, who have two losses), but only one of those five (Cincinnati vs. Rutgers) is favored (by the model and in Vegas). Cincinnati is the only team in the state favored against a team from another state.

Thursday, November 15, 2012

Week 11 NEPA leaders

Time for my weekly look at the NEPA (Net Expected Points Added) leaders in the country.  Let's jump right in.

Quarterbacks

Ranking the teams with NEPA

After perusing our team NEPA rankings numerous times over the last week, I wondered what it would look like if I ranked every team according to their cumulative rankings.  I took the offensive NEPA per play, subtracted the defensive NEPA per play (because teams with good defenses have negative NEPAs), and ranked the teams by the totals.  Also, I chose NEPA per play, because it takes away the benefit of playing extra games or overtimes.
Here's what we get:

Wednesday, November 14, 2012

Alabama is only mostly dead

You don't have to look far to find a news article telling you how Texas A&M dashed Alabama's title hopes. In the end that may be true, but for now, Alabama is only mostly dead. 

To get back in the race, Alabama needs at least two of the currently undefeated teams to lose. To figure out the chances of that happening we first estimated probabilities for each team staying undefeated. The model gives Kansas State a 72.3% chance of winning out, Notre Dame a 41.6% chance and Oregon, with the toughest road ahead, a 56.0% chance. 

There's a big difference between mostly dead and
all dead. Mostly dead is slightly alive.
In this scenario, there are eight possible outcomes - they all stay undefeated, they all lose, one loses (x3), and two lose (x3). The probability of all three winning out is .723*.560*.416=.169 or 16.9%. The probability of all three losing is (1-.723)*(1-.560)*(1-.416)=.071 or 7.1%. For one team losing it comes out to 43.3% and the probability of two teams losing is 32.7%. Alabama needs two teams to lose (32.7%) or all three (7.1%), so there is a 39.8% chance that the Tide will again control their own destiny.

But then the Tide have to exploit that opportunity. The Tide have a 73.7% chance of winning their next three games. In other words, had they beat the Aggies, they would now have a 73.7% chance of walking unbeaten into the championship game (and a 90% chance or so of getting an invitation to the game). Instead, their chances have been reduced by .398 to .737*.398=29.3%. In other words, the loss cut Alabama's title game chances from around 3/4 to 1/3.  Down, but not out.

Fantasy Bumpers Week 11

A fantasy bumper is a player that should perform significantly better than average this week. There are generally two reasons for this: 1) an established player has a good matchup or 2) poor performance or injury to teammate will give the player more opportunities. In the first case, these players are already owned so you want to make sure they are in your lineup. In the second case, they could be a valuable pickup.  Since this is a stats-based site and not a fantasy site, and I maintain this site on top of my day job while trying to raise a son and keep my wife from realizing she could do a lot better, these results are based on the best, most recent stats available to me and nothing more. It falls on you to do your own research. You can get fantasy rankings here.

Kolton Browning is in a walking boot and North Texas has the country's 96th best defense, a perfect storm of opportunity for Cody Wells. The model also sees Monterrell Washington having a good game for ULM.

With Lamar and Daniels out for South Florida, all eyes will turn to Demetris Murray. The model sees a big game for Murray against a Miami defense that is 121st in rush yards allowed per game.

Collin Klein is good, Baylor is not. How's that for analysis? The same goes for Sankey, Price and Colorado. Louisiana Tech offers the dynamic fantasy combination of a high scoring offense and non-existent defense, so Chuckie Keeton should have a huge day.

Tuesday, November 13, 2012

Making the Case: Johnny Football for Heisman

Manziel rushed for 92 yards and threw for 253 against Alabama. He completed 15 of his first 16 passes and had 200 yards of offense at the half. That’s really good, but just how good is it?

The NEPA (Net Expected Points Added) converts a player’s performance into points based on how many more (or fewer) points the typical team could expect to score on that drive after each play1. Johnny Manziel scored a 15.6 against Alabama. This means he was a little over two touchdowns better than what you could expect from a typical college football player against a typical defense.

But Alabama is far from typical. Manziel's 15.6 NEPA is the best individual performance against Alabama this season. Only LSU’s Mettenberger has managed double digits against the Tide. In fact, Manziel's was the best performance against Alabama by a quarterback since 2008 (Tebow had a NEPA of 16.4 and Brian Johnson a 15.9). That includes Cam Newton's performance in Bryant-Denny Stadium.

Johnny Football is second nationally in NEPA this season behind only Jordan Lynch of Northern Illinois. Lynch is having an incredible season, but Manziel plays in the SEC, not the MAC. Manziel's opponents have allowed only 4.0 points per game to QBs versus 8.1 for Lynch. Johnny Manziel has the largest differential in the nation between his per game NEPA and his opponent's per game NEPA allowed, and it would be larger if Johnny Football was always playing all four quarters.

The player with the highest NEPA while playing in a BCS conference has won the Heisman three of the last four seasons: RGIII, Cam Newton and Sam Bradford. Tebow finished second to Graham Harrell in 2007 and Mark Ingram was second among BCS conference running backs (behind Toby Gerhart; the 2009 BCS NEPA leader was Jerrod Johnson, but the Aggies were 6-7). Manziel will need to hold off Seth Doege and Nick Florence to claim the 2012 NEPA championship like his Heisman predecessors, but like Harrell in 2007, Doege and Florence are not legitimate Heisman candidates.

And Johnny Football's performance is comparable statistically to past Heisman winners. Manziel is only .3 points per game behind Cam Newton’s Heisman-winning pace. He will need a NEPA of 19 against Sam Houston State to pass Newton, so the biggest challenge will be staying in the game long enough to rack up the numbers.

While Klein, Barner and Lee are having outstanding seasons, Manziel has better accumulated stats and the best individual performance on the biggest stage this season. So if Manziel is college football’s most outstanding player statistically, all that could hold him back are contextual factors. He’s a freshman on a two loss team – and Tebow was a sophomore on a four loss team and had a NEPA just a point and a half per game better than Manziel’s. Plus, who else does this on a regular basis?

1 the "Net" in NEPA means that it considers not only the points the offense can expect to score on that drive but the points the opponent can expect to score on their next possession, so field position is important, too.

Polling CFBTN

The BPR ranks teams only on wins and losses, the power ranking is used to predict who would win if two teams played/played again, and the Hybrid ranking combines features of the two and is meant to reflect the logic used in human polling. Check out last week's rankings here.

Notre Dame didn't win much by beating Boston College, allowing Kansas State to again slip ahead into the #1 spot in the BPR. Alabama fell to #6 from #3 and Texas A&M jumped 7 spots to #9, so the SEC now has five teams in the top 10 but none in the top 4. 

Alabama retains is top position in the power rankings but only by the slimmest of margins over Oregon. Florida continues to fall and will soon drop behind a fast rising USC team and out of the top 10. Georgia made a big jump to #6 and Texas moved up five spots to #18.

Finally, with Alabama losing Oregon moves up to #1 in the hybrid rankings. Alabama, though, does stay ahead of Notre Dame. Northern Illinois makes its long awaited debut into the top 25 and Rutgers returns to take the 25th position in the Hybrid rankings.

Monday, November 12, 2012

Week 11's Best Offensive Performances

The EP3, or effective points per possession, measure the average number of "effective" points earned each drive during a game. Effective points consider field position as well as actual points. The EP3+ is the team's average performance over the course of the season adjusted for the opposition. The EP3+D is the defensive score while the EP3+O is for the offense. The Perf-O begins with the single-game EP3 and adjusts for the opponent's EP3+D. The Perf-D begins with the opponent's EP3 and adjusts it for the opponent's EP3+O.

A couple of weeks ago I showed that Oregon only had the nation's second best offense. Apparently they took this personally as they have now had the top offensive performance two weeks in a row and have now surpassed Kansas State as the nation's premiere offense (and, arguably, the best overall team). Last week the Ducks edged out Texas A&M's performance against Auburn to post the best offensive performance of the season. Oregon's work against Cal was not quite of the same caliber this week, but they were again able to edge out the Aggies by the smallest of margins. 

Speaking of Aggies, Johnny Manziel and company finished in the top two for the third week in a row. To put their 1.08 EP3 in perspective, LSU managed only a .04 in their near upset of the Tide. While Oregon has regained its throne as the best offense in the country, Texas A&M has now claimed the #2 spot (and #4 overall behind Alabama and Kansas State).

The Longhorns have returned to form. Arizona posted a massive 3.47 against the worst college football team  ever from a program that has recently produced nationally competitive squads. Syracuse did a number on Louisville, and that number was 2.74, and the other number was 1, as in not 0. 


Week 11's Best Defensive Performances

The EP3, or effective points per possession, measure the average number of "effective" points earned each drive during a game. Effective points consider field position as well as actual points. The EP3+ is the team's average performance over the course of the season adjusted for the opposition. The EP3+D is the defensive score while the EP3+O is for the offense. The Perf-O begins with the single-game EP3 and adjusts for the opponent's EP3+D. The Perf-D begins with the opponent's EP3 and adjusts it for the opponent's EP3+O.

USC rebounds from the Oregon offensive clinic and shuts down Arizona State. Rutgers did the same, but to Army. After a rough start TCU held Klein and family in check, a feat few have accomplished . . . but still lost, because Kansas State returned the favor. Like TCU, Virginia Tech, and Arizona State, Alabama also finished in the top 10 and lost. That Alabama was able to turn in one of the nation's ten best defensive performances despite failing to force a turnover and allowing 20 points in the first quarter gives a hint at just how good this Texas A&M offense has become.

Speaking of the Aggies, Kevin Sumlin's team needed not just a #2 performance on offense (see the next post) but also a top 20 performance on defense to win. Fortunately, the undermanned, often underrated Texas A&M defense was able to deliver such a performance.



Saturday, November 10, 2012

Vegas vs. CFBTN, Week 11

In five games this weekend the model and Vegas disagree by more than five points. The model likes Kansas State and Maryland presumably because it isn't appropriately accounting for injury concerns (injuries are tough for a statistical model). In the other three cases the model doesn't like big road favorites quite as much as Vegas; unless you have some particular reason to care if they win by two or three touchdowns, the inconsistency between Vegas and the model are just not that interesting.


Friday, November 9, 2012

Week 11 Preview

As usual, I'm picking out what I subjectively decide are the 5 most interesting games of the week, and I see what NEPA says to expect from those games.

Louisville @ Syracuse (Noon, ABC/ESPN)

QBs: Teddy Bridgewater (116.7 NEPA, .383 NEPApp) vs. Ryan Nassib (70.4, .166)
RBs: Senorise Perry (28.8, .187) & Jeremy Wright (23.3, .148) vs. Prince-Tyson Gulley (0.6, .005) & Jerome Smith (-1.0, -.006)
Defense: Cardinals (40.1, .151 passing / -22.7, -.075 rushing) vs. Orange (49.7, .187 passing / -37.0, -.110 rushing)
CFBTN projection: 31-31

Normally, an undefeated playing a 4-5 conference opponent wouldn't strike me as an important game.  However, the unbeaten Cardinals are favored by a skinny 2 point margin, so there's clearly some thought that Syracuse could win this game.  This could have BCS implications, as the Big East is still pretty wide open, with four teams still realistically in contention for the conference (these two, Rutgers, and Cincy).  

More fun with the transitive property, the SEC is good at football, some other conferences are not

The problem with using the transitive property (e.g., if A>B & B>C then A>C) in college football is that a team's performance can vary a lot from game to game. Texas can lose to West Virginia, almost lose to Kansas, and then own Texas Tech after Tech destroys West Virginia. But, with 124 teams and 12 to 14 game schedules, the transitive property is essential.

Freakishly Symmetric
To make the most of it, then, I've tracked every connection between each college football team and Alabama, using up to 6 different games to make the connection. I added up the point differentials of the connecting games and averaged them together (weighting cases for the error introduced when you use additional steps to make the connection).

For example, Texas A&M lost to LSU by 5. LSU lost to Alabama by 4. So Texas A&M would get a -9 for that connection. Then, Texas A&M beat Mississippi State by 25 after Alabama beat them by 31, so Texas A&M gets a -6. But also Alabama played Western Kentucky, who played Kentucky, who played Arkansas, who played Ole Miss who played Texas A&M.

The result is what I call the Transitive Gap (TG). Alabama comes out on top, beating its opponents on average by the exact same average that its opponents get beaten by the Tide. Coming in at #2 is Texas A&M, meaning that no one else comes as close to beating the same teams (or teams that played teams) as badly as Alabama does than the Aggies. Texas A&M is followed by Kansas State, Oklahoma and Oregon. (In case you're wondering, the results do depend somewhat on the team used for comparison - Alabama in this case. That being said, Alabama is consistently on top with Oregon and Texas A&M usually in the top 3. Kansas State, Oklahoma, South Carolina and Georgia are often in the top 5.)

Coaching Carousel: Joker Phillips

Reason left: Fired (November 4)
Record: 12-22
SRS: -0.77
SRS previous 3 years: 6.34
(SRS: Sports Reference's Simple Rating System)

It's easy to make the basketball school references here, but that underscores just how disappointing Joker Phillips has been at Kentucky.  Rich Brooks did a wonderful job in Lexington, turning Kentucky into a viable SEC team.  They weren't an SEC contender, but they weren't... you know... Kentucky.  Now they're Kentucky again.  And Kentucky doesn't like being Kentucky.

Brooks took the team to a bowl in each of his final 4 seasons, and Phillips continued this in his rookie year.  It's the longest bowl streak in school history.  Since that first year, the Wildcats have progressively gotten worse.  The most surprising thing was how they got worse - on offense.  Phillips was offensive coordinator for Brooks' potent offenses when the team was at its best.  When Phillips took the reins, it was natural to worry over some regression, but that regression was expected to happen on the defensive side.  Phillips has been unable to find a QB to run the offense.  This year's committee of Patrick Towles, Morgan Newton, and  Jalen Whitlow have been particularly bad, combining for -56.4 NEPA.  It's easily one of the least productive passing games in the country this year.  And finally, you can't really give Phillips a pass for going up against more talented defenses - he was facing more talented defenses when his offense was considered powerful and dangerous.  Now the only danger the offense poses is to itself.  No, seriously:



Still, 34 games does seem like a bit of a quick trigger when you consider that:
1. Phillips was hand-picked for the job by his predecessor, arguably the best coach in UK history.
2. He's a UK alum unlikely to bolt for a better job.
3. He did reach a bowl in 33% of his seasons, which is pretty good for UK.

But those 3 things weren't enough to overcome the obvious: the team just keeps getting worse and worse.  This move says that Kentucky wants to win, and pretty badly.  Just how desperate are they?  I ask because a notable former SEC coach is apparently interested... 
Brent Blackwell compiles the EPA rankings for cfbtn.com.  Follow Brent on Twitter by mashing the pretty button below.

Thursday, November 8, 2012

Week 10 NEPA rankings

You might have noticed some changes to the EPA/NEPA page.  It's more navigable than before, with more options.  Scott has added pages for individual receivers, team offense, and team defense.  This should allow more time for analysis.  Very exciting stuff.

I'll also note that we've decided to sort the leaders by Net Expected Points Added.  It's more contextual, and after 10 weeks of looking at both I feel it is a better measure of real value to the team.  It will likely have a minimal effect on the ordering of the rankings, but it's worth pointing out.

Ok, on to the leaderboards:

Thursday's Trip around the Statverse

Bill Connelly continues to produce. And I might not like the conclusion, but his preview of A&M and Alabama is definitely worth the effort.

Brian Fremeau discusses Fornelli's blind taste test of college football teams. And don't fail to notice who comes out on top in the FEI.

ESPN Stats & Info begin with the assumption that Oregon's offense is the nation's best unit, an assumption my  and others' advanced stats will force you to at least reconsider. It's pretty much downhill from there. They also look at conference rankings, putting too much emphasis on human polls - possibility for bias, limited in scope - and generally ignore the computer polls that, like my rankings, put the Big 12 on top.

A numerical breakdown of Kansas State/TCU from PURPLEKOOLAID. This is why I'm beginning to fall in love with Kansas State (the community, not the team; I hate Bill Snyder): Kansas State loses the eyeball test so their fans are learning to really appreciate the numerical side of college football. I like numbers, I like smart fans, and I'm drinking the Purple Kool Aid.

And as always, please let me know if you stumble on (or even wrote) some fancy shmancy work using stats in college football.

Wednesday, November 7, 2012

Combinatorics in College Football

I started with the assumption that there are currently four college football teams that have legitimate BCS title game resumes. Next, I ranked those teams based on where they might finish in the BCS if they stay unbeaten:

1) Alabama
2) Oregon
3) Kansas State
4) Notre Dame

(Sorry Louisville.) This list is open to debate, of course, but that isn't the focus of this analysis and it gives us something to work with.

Then I took the schedules of the four teams, attached to each game a probability based on the model, and simulated the rest of the season 1,000 times.

team (regular season %undefeated, post-champ game %undefeated)
1) Alabama (84%, 67%)
2) Oregon (61%, 47%)
3) Kansas State (62%, 62%)
4) Notre Dame (47%, 47%)

The results are illustrated in the table below. Red means the team lost at least once, green that the team did not. The number at the top of each column is the number of times that outcome occurred divided by 10 (or the percentage).


So, for example, in 84 or 8.4% of simulations all four teams stayed undefeated (all four teams are green in that first column). Three percent of the time they all lost. In 20.5% of simulations three teams lost, in 33.8% two teams lost at least once, and in 28.9% of simulations just one team bit the bullet.

That means we have  a one in three chance of avoiding the controversy that comes when a team with a similar resume and equal number of losses is left out (that's assuming, of course, that a one-loss SEC team doesn't refuse to concede to an undefeated Notre Dame).

The most probable individual scenario is that an undefeated Kansas State and undefeated Alabama play for the title after Oregon and Notre Dame lose (13.7%). If Oregon does really have the leg up on Kansas State, the most likely result is Oregon playing Alabama (30.1%). Oregon would also get in, I assume, if Kansas State, Notre Dame and Oregon all lose, and would play the winner of Alabama/Georgia (10%; sorry Louisville).

But if you are to take one thing away from analysis it should be this: the outcome of this season is by no means a foregone conclusion, and it's a bit silly to start worrying now about who will finish on top of the BCS until teams have had a few more chances to lose . . . because some of them will.

Idaho starts the coaching carousel

Obscured by attention given to bigger games involving good teams with bright futures, the first mid-season coaching change of 2012 took place October 21, with Idaho firing Robb Akey after a 70-28 loss to Louisiana Tech, moving the Vandals to a disappointing 1-7.  Was it the right decision?  I doubt that can be answered in the time I'm willing to spend on this, but despite his 20-50 career record, I think there's a case to be made that Akey was one of the best coaches in Idaho's underwhelming FBS tenure.  After a 3-21 start, Idaho went 8-5 in 2009, winning the 2nd bowl game in school history, and he followed that up with a 6-7 year in 2010.  Since Idaho moved to the highest level of football in 1996, they've engineered just 5 seasons with at least 6 wins, and Akey is responsible for 2 (Chris Tormey managed the other 3 against ridiculously easy schedules).  Sports Reference's Simple Rating System isn't a bad rough and dirty measure of team strength, and see how Akey compares with Idaho history:
Akey: -12.61
Previous 5 years: -15.22
Modern era (1996-): -12.01

Using those measures, Akey was better than his immediate predecessors and basically as good as the average Idaho coach since Idaho returned to FBS football.

Akey did improve Idaho's recruiting.  Per Rivals' recruiting rankings, in the 5 years prior to Akey's arrival, Idaho pulled in only 2 players ranked 3 stars or better. In Akey's 6 years, they've gotten 22.  Now, that's not to say that he's turned Idaho into a program with viable talent.  He hasn't, but he did improve the recruiting base.

Ironically, I think Akey's recruiting success is one reason for the timing of his firing.  Akey's recruiting is still better than what it used to be, but it is starting to stall.  Only 2 recruits are committed for 2013, and impact recruits are getting hard to come by again.  Perhaps the powers that be want to thank Akey for the good work he's done - and despite the 20-50 record, I think it's safe to say the work was comparatively good - and move on to someone who can do something with the talent onhand.  For now they turn to OC Jason Gesser, only 33 and still in his first year of collegiate coordinating.  He'll finish out the year while they search for a successor, as I doubt Gesser is the man for the job.  In fact, there's probably evidence that Akey's hiring of Gesser might have been responsible for his firing:  QB Dominique Blackman ranks in the bottom five in our QB EPA rankings, a whopping -60 points below average this season at the time of the firing.  

There's no doubt the next hire will be a key hire for Idaho, bigger than the Akey decision was in 2007.  Akey seemed to have considerable job security when the year began, so this seems a bold move for the Vandals.  There's talent onhand, but the momentum in Idaho seems to be teetering on the brink of collapse;  that is to say, though they're unlikely to miss Akey, the wrong hire could quickly send the program back to the Tom Cable era.  It's a bold move - the right hire could make Idaho a potential contender wherever they wind up, but the wrong one could really set this program back a decade.  With Idaho's football and even FBS status in jeopardy, it'll be interesting to see what kind of candidates decide to interview for this unenviable position.

Weighing in on Kickers

The NEPA is a statistic we have invented that allows us to measure the point contribution of individual players on individual plays above what the average player would do in that situation. The trouble with the NEPA, and advanced stats in football in general, is that we can't remove the influence of teammates - e.g., the offensive line. Even on this site we rave about skill position players while generally ignoring the other players that make it possible, even though I was an offensive lineman in my day.

But let's assume, for the sake of simplicity, that half of a running back's production is the product of the supporting cast while the other half reflects his own ability. There are major problems with this assumption, but it helps to demonstrate my next point. If we do that, there are 15 running backs with a personal NEPA over 15.95 (total NEPA of 31.9).

What's magical about 15.95? It's what a running back would need to be more valuable to his team than Tulane's Cairo Santos, a kicker. To be fair, Santos isn't just any kicker; he's the best kicker in college football. He's almost a full half point better per kick than everyone else.

Tuesday, November 6, 2012

Polling CFBTN

The BPR ranks teams only on wins and losses, the power ranking is used to predict who would win if two teams played/played again, and the Hybrid ranking combines features of the two and is meant to reflect the logic used in human polling. Check out last week's rankings here.

Notre Dame retains its lead over Kansas State in the BPR, while Alabama and Oregon close in on those two with more impressive wins Saturday. With Texas A&M and the SEC title game still on the schedule, Alabama will catch both Notre Dame and Kansas State, but Oregon would need help to get in to the title game if the BPR were making the decisions. 

Georgia moves ahead of LSU (after the loss) and Louisville (despite winning). Clemson at 15 and Florida State at 18 is a solid reminder that the ACC is not good at football. Only Texas Tech and Michigan squeeze into the top 25 with three losses.

Alabama still dominates the power rankings, as they have done all season, even after a close shave from LSU, who moves up one spot despite the loss. Oregon makes a strong move in to the 2nd spot Texas A&M continues its ascent up the power rankings to #4 from #8 a week ago. The Aggies are 4 points closer to Alabama this week than they were a week ago. 

Florida State and Texas A&M jump both Florida and Notre Dame after they struggle for wins against inferior competition, even though Florida State did not even play. South Carolina at #10 does not reflect the Lattimore injury. With LSU at #11 the SEC has six of the top 11 spots. 

Again, Alabama has a strong lead nationally in the hybrid rankings. Oregon jumps Kansas State but has only the smallest of leads over the Wildcats. 


This Week's Fantasy Bumpers

A fantasy bumper is a player that should perform significantly better than average this week. There are generally two reasons for this: 1) an established player has a good matchup or 2) poor performance or injury to teammate will give the player more opportunities. In the first case, these players are already owned so you want to make sure they are in your lineup. In the second case, they could be a valuable pickup. Since this is a stats-based site and not a fantasy site, and I maintain this site on top of my day job while trying to raise a son and keep my wife from realizing she could do a lot better, these results are based on the best, most recent stats available to me and nothing more. It falls on you to do your own research. You can get fantasy rankings here.

A few teams consistently find their opponents on the fantasy bumpers list: West Virginia, Idaho, Colorado, etc. BYU, for example, is clearly no offensive powerhouse this season; Lavell Edwards would be spinning in his grave if he were in one. But the Cougars are projected to score 45 against Idaho, the worst team in the FBS, and Nelson should be the primary beneficiary. Matt Scott will look to complete the roller coaster cycle - beat USC, get destroyed by UCLA - and take out their frustrations on the Buffs of Colorado. And Wes Lunt should find more passing lanes against WVU than against KSU if he is able to play.

Monday, November 5, 2012

Week 10's Best Offensive Performances

The EP3, or effective points per possession, measure the average number of "effective" points earned each drive during a game. Effective points consider field position as well as actual points. The EP3+ is the team's average performance over the course of the season adjusted for the opposition. The EP3+D is the defensive score while the EP3+O is for the offense. The Perf-O begins with the single-game EP3 and adjusts for the opponent's EP3+D. The Perf-D begins with the opponent's EP3 and adjusts it for the opponent's EP3+O.

Johnny Football was two-upped this weekend. Not only was the Aggie's 690 yard performance against Mississippi State only good for 2nd in the Perf-O rankings, Oregon had the audacity to displace Texas A&M's win against Auburn as the best single-game offensive performance of the season. (I have rejigged the Perf formula this week, so this week's results are comparable to posted past results.) I think Oregon heard I was saying some disparaging things about them last week and got upset. USC, while losers on the day can be proud to know that, offensively, they were better than all but two teams on Saturday. Unfortunately, one of those two teams was their opponent. 

LSU comes in 23th despite a measly 0.04 EP3. Because an EP3 of 0 is average, anything above that against Alabama is pretty dang good. Pitt (vs. Notre Dame), Oklahoma State (vs. Kansas State), and San Diego State (vs. Boise State) also managed to make the top 25 with an EP3 below 0.5. San Diego State was the only of the four to pull off the win. Then, there's Alabama. The Tide had one of the lowest winning EP3s of the season, but the model concluded that both the Alabama offense and LSU defense were playing at a high level.




Week 10's Best Defensive Performances

The EP3, or effective points per possession, measure the average number of "effective" points earned each drive during a game. Effective points consider field position as well as actual points. The EP3+ is the team's average performance over the course of the season adjusted for the opposition. The EP3+D is the defensive score while the EP3+O is for the offense. The Perf-O begins with the single-game EP3 and adjusts for the opponent's EP3+D. The Perf-D begins with the opponent's EP3 and adjusts it for the opponent's EP3+O.

UCLA managed one of the better performances of the season in their 66-10 win over Arizona. Not only did the Bruins have the week's 3rd best Perf-D, they were also 5th in Perf-O. LSU was in the top 25 in Perf-O and -D but still managed to lose. LSU played the best defense in the nation on Saturday, but Alabama was also playing top 20 offense. Missouri also finished in the top 10 in a losing effort.





Monday's Trip around the Statverse

Please let me know if you stumble on any great reading in the world of advanced statistics in sports, especially college football.

Kelly Dwyer rips Doug Collins for being too lazy to use advanced stats in the Ball Don't Lie blog.

Jeremy Mills at ESPN Stats & Info takes an look at college football and the BCS rankings. The folks at Stats & Info also take a closer look at two guys (Martin, Luck) that were playing college football at this time last year and are tearing it up at the next level now.

Andrew Karl has an article in the Journal of Quantitative Analysis in Sports on the potential impact of modeling assumptions on team rankings in college football.

This is a year old but new to me: Bill Connelly gets some love from Vox Magazine for his work on Rock M Nation.

Sunday, November 4, 2012

The SEC West gets Johnny Footballed, a win is a win is a loss, mere mortals, and the battle for Louisiana's soul

218, 290, 134, 352, 361
716, 481, 410, 671, 693

Texas A&M

"If we didn't think it could work, we wouldn't run it." So said Kevin Sumlin about the Aggie's up-tempo, not boring offense in the SEC. In five games against SEC West opponents, Texas A&M has been held under 670 yards twice, against Ole Miss and LSU. In those two games they still managed over 400 yards despite losing 6 and 5 possessions to turnovers. Only LSU held A&M under 200 yards rushing. The Aggies have averaged 3.19 points per possession in possessions that were not cut short by the end of the half; they have averaged 4 points per possession when not stopped by turnovers.

What does it all mean? When the Aggies avoid turnovers they are as good as anyone in the country (e.g., Alabama). Good offenses avoid turnovers and good defenses create turnovers, but the timing of turnovers is much more random than other events in football (e.g., long pass plays). In other words, this next weekend the football gods will be playing dice with the college football universe.

Friday, November 2, 2012

Week 10 EPA Preview

Like last week, in this space I'll take a look at what I subjectively decide are the 5 most interesting games of the weekend.  Specifically, I'll look at what EPA tells us to expect.

Texas A&M vs. Mississippi State (12:00 ET, ESPN)

CFBTN projection: 36-29, Aggies
Quarterbacks: Johnny Manziel (94.67 EPA, .234 EPApp) vs. Tyler Russell (37.31 EPA, .142 EPApp)
Running Backs: Christine Michael (7.51 EPA, .127 EPApp) & Ben Malena (3.37 EPA, .038 EPApp) vs. LaDarius Perkins (9.48 EPA, .057 EPApp) & Josh Robinson (-1.07 EPA, -.036 EPApp)

Beware the Pothole in Week 10

My house was recently hit by a big storm called a hurricane. You might have heard about it, it was on the news - not my house, the hurricane. A hurricane is a specific type of tropical cyclone. It is also the name for a very mediocre football team, which is bad news for the folks in Blacksburg. Hurricanes are dangerous because 1) they come with a lot of water and 2) they blow stuff around and generally make a mess. At our house, the hurricane knocked out one fence plank, one table leg, and pulverized a lawn chair. It made a mess, but in the end we were very lucky. We wish the best to all of those that were less fortunate.

You must now be asking what this has to do with Notre Dame? College football could be facing its own storm, a figurative storm that could really make a figurative mess of things. There are currently four teams that are clear-cut title contenders: the aforementioned, and too oft-mentioned, Notre Dame, Oregon, Kansas State, and Alabama. They each play a game this weekend. In each case there is a solid reason to believe they could lose. A loss or two would be very disruptive to the current landscape of 2012 college football season. I have nothing against any of these teams, but that would be awesome!

Thursday, November 1, 2012

Kansas State's offense is REALLY good

I listen each week to The Solid Verbal's preview podcast. They do a great job covering college football as very well-informed fans, which is rare. But even when drawing on that wealth of knowledge they have accumulated over the years, sometimes they are very wrong.

I apologize in advance because I have a hard time differentiating between the co-hosts, but this week one of the two, either Ty Hildenbrandt or Dan Rubenstein, picked Oklahoma State to beat Kansas State on the grounds that Kansas State will eventually lose because "despite being very solid on defense, I still have questions about their offense". [Correction: Ty Hildenbrandt has informed me that 1) these are his words and 2) he did not pick OSU to pull off the upset but that he was generally concerned KSU was too dependent on Klein.] He then added that they could be beat if a team figured out a way to take Collin Klein out of the game. Unless you're willing to take a crowbar to his knees, it's about time people recognized that you can't just take Collin Klein out of the offense, and with Klein, Kansas State has the best offense in the country. And it's not that close.

Quarterback EPAs, Week 9: Collin Klein and a bunch of mortals

In my week 9 preview last week, I wrote about 5 games and what EPA told us might happen in each.  I didn't predict each game precisely, but I said Braxton Miller would be the difference against Penn St, and he was.  I said Mississippi State likely wouldn't get the miracle effort needed to beat Bama, and they didn't.  I said Georgia had a puncher's chance against Florida due to their running backs and Florida's tendency to shoot themselves in the foot on offense, and that's precisely why Georgia won.  I said Notre Dame, if their defense could keep things close, could keep things mixed up enough on offense to win, and that happened.  I brag about what I got right because I must point out that I also said something very, very foolish:

 It could be a lean day for Collin Klein.

My bad.  Collin Klein didn't only not have a lean day, he feasted like a king, all the way to 19 EPA against one of the nation's best defenses - the same defense that rendered Geno Smith helpless.  He takes a 10 point lead in our EPA standings and firmly takes hold of the Heisman race with 109.19 EPA.

#2 is now Jordan Lynch of Northern Illinois (99.81 EPA), who has over 1900 passing yards, 1100 rushing yards, 32 TD, and only 3 interceptions.  Little known fact - Jordan Lynch is the only player in the nation controlled by a kid on his Playstation.

#3 is the best QB in the SEC, Johnny Manziel (94.67 EPA).  As a freshman in the SEC, Johnny Football has over 2200 pass yards, almost 800 rushing yards, 29 TD and only 6 turnovers.  Also, he's a freshman.  So, good luck with that for the next 2 or 3 years, SEC West.

The top 10 also includes names from past updates, like Seth Doege, Cody Fajardo, Nick Florence, Tahj Boyd, Rakeem Cato, and Matt Barkley.  Yes, Barkley is back in the top 10, despite the loss to Arizona, which was anyone's fault but his own.  Heisman contender Braxton Miller ranks 19th; he's very good, but I'm not sure he's done enough at this point to deserve the chance to clap for Collin Klein in person.

At the bottom?  Jordan Webb (-70.99) gives Colorado the wonderful distinction of having the least productive QB as well as the least productive running back.  Kansas'  Dayne Crist is next from the bottom, and I can only hope Charlie Weis will try to silence me for mentioning it.  Idaho's Dominique Blackman has been so terrible he might have gotten his coach fired, and Auburn's Kiehl Frazier seems hell-bent on the same result.  Now, these guys only get to compete for worst QB in American because Missouri has split time between James Franklin (-44.51) and Corbin Berkstresser (-48.41).  Not even Webb is within 20 points of their collective awfulness.

The strangest names to see near the bottom?  Hawaii's Sean Schroeder and SMU's Garrett Gilbert.  I expect more from Norm Chow and June Jones quarterbacks.