Thursday, February 14, 2019

2018 Adjusted Pythagorean Record: Big 10

Last week we looked at how Big 10 teams fared in terms of yards per play. This week, we turn our attention to how the season played out in terms of the Adjusted Pythagorean Record, or APR. For an in-depth look at APR, click here. If you didn’t feel like clicking, here is the Reader’s Digest version. APR looks at how well a team scores and prevents touchdowns. Non-offensive touchdowns, field goals, extra points, and safeties are excluded. The ratio of offensive touchdowns to touchdowns allowed is converted into a winning percentage. Pretty simple actually.

Once again, here are the 2018 Big 10 standings.
And here are the APR standings sorted by division with conference rank in offensive touchdowns, touchdowns allowed, and APR in parentheses. This includes conference games only with the championship game excluded.
Finally, the Big 10 teams are sorted by the difference between their actual number of wins and their expected number of wins according to APR.
Northwestern and Ohio State nearly met the game and a half threshold that I use for determining if a team significantly over or under-performed relative to their APR. Those two teams did meet the YPP standard and we went over why last week, so I won't bore you with rehashing that this week.

Whither the Preseason Top-Ten Teams that Finish Unranked?
For the first time since 2012 and just the third time since 2003, Wisconsin finished the 2018 season unranked. This feat was doubly notable since the Badgers began the season in the top-five (they were fourth). A lot of Power Five teams would kill (or more accurately cheat like hell for an 8-5 record and bowl win). However, based on preseason expectations, 2018 qualifies as a disappointment for Wisconsin fans. What then should Wisconsin fans expect in 2019? Are the Badgers likely to rebound? To answer that question, I looked at all preseason top-ten teams (in the AP Poll) that finished unranked since 2005. Besides the three teams that ‘accomplished’ this feat in 2018 (Auburn and Miami joined Wisconsin as unranked finishers despite preseason top-ten rankings), 23 other teams began the season in the top-ten and finished outside the final polls. Those teams along with their follow up (regular season) records are summarized chronologically in the following table.
In the aggregate, those top-ten teams that finish unranked tend to rebound the following season. Of the 23 previous top-ten teams to finish unranked, only seven failed to see their win total improve the following season. In addition, four of those seven teams experienced coaching changes (Florida, Arkansas, Georgia, and Florida State) so that may partially explain why they did not improve. When we break things down further, this is how those 23 teams fared on average in their follow up season.
On average, the 23 teams improved by 1.35 games the following season. Nearly 70% of the teams improved and over half improved by at least two games. Only about a quarter of the teams declined in the follow up season.

Chances are at least two of the three preseason top-ten teams that finished unranked from last season will rebound in 2019, and seeing all three rebound wouldn’t be that shocking.
Were I forced to choose two, I would take Auburn and Wisconsin to rebound. Auburn always seems to zig when the preseason consensus says they will zag and Wisconsin has been one of the most consistent programs in the nation over the past two decades. Miami could certainly rebound as well, but I don’t feel quite as confident in the Hurricanes since their offense looked lost at the end of 2018 and they are going through a coaching change which is one of the common traits in most of the teams that did not improve. While each season is viewed in hindsight as preordained, the reality is the small sample size of college football can produce highly variable results. Pollsters tend to do a solid job of evaluating teams in the preseason, so even if a team disappoints record-wise, their underlying talent is probably closer to where the preseason polls rated them.

Thursday, February 07, 2019

2018 Yards Per Play: Big 10

We now move to our third conference in the offseason recaps. After this week, the recaps will be one quarter done. Here are the Big 10 standings.
So we know what each team achieved, but how did they perform? To answer that, here are the Yards Per Play (YPP), Yards Per Play Allowed (YPA) and Net Yards Per Play (Net) numbers for each Big 10 team. This includes conference play only, with the championship game not included. The teams are sorted by division by Net YPP with conference rank in parentheses.
College football teams play either eight or nine conference games. Consequently, their record in such a small sample may not be indicative of their quality of play. A few fortuitous bounces here or there can be the difference between another ho-hum campaign or a special season. Randomness and other factors outside of our perception play a role in determining the standings. It would be fantastic if college football teams played 100 or even 1000 games. Then we could have a better idea about which teams were really the best. Alas, players would miss too much class time, their bodies would be battered beyond recognition, and I would never leave the couch. As it is, we have to make do with the handful of games teams do play. In those games, we can learn a lot from a team’s YPP. Since 2005, I have collected YPP data for every conference. I use conference games only because teams play such divergent non-conference schedules and the teams within a conference tend to be of similar quality. By running a regression analysis between a team’s Net YPP (the difference between their Yards Per Play and Yards Per Play Allowed) and their conference winning percentage, we can see if Net YPP is a decent predictor of a team’s record. Spoiler alert. It is. For the statistically inclined, the correlation coefficient between a team’s Net YPP in conference play and their conference record is around .66. Since Net YPP is a solid predictor of a team’s conference record, we can use it to identify which teams had a significant disparity between their conference record as predicted by Net YPP and their actual conference record. I used a difference of .200 between predicted and actual winning percentage as the threshold for ‘significant’. Why .200? It is a little arbitrary, but .200 corresponds to a difference of 1.6 games over an eight game conference schedule and 1.8 games over a nine game one. Over or under-performing by more than a game and a half in a small sample seems significant to me. In the 2018 season, which teams in the Big 10 met this threshold? Here are Big 10 teams sorted by performance over what would be expected from their Net YPP numbers.
The two Big 10 Championship Game participants (Northwestern and Ohio State) significantly exceeded their expected record based on YPP (Northwestern by an historic margin) while Rutgers, Nebraska, and Maryland all fell short of their expected record. As is typically the case, results in close games are largely to blame. Northwestern and Ohio State finished a combined 8-1 in one-score conference games while Rutgers, Nebraska, and Maryland finished a combined 1-8.

Is Maryland Even Trying?
In early December, Maryland named Alabama offensive coordinator Mike Locksley its newest football coach. Having an Alabama coordinator take the reins of a Power Five program would not normally inspire me to type many words on the subject. But Locksley is not just any Alabama coordinator. In fact, he has been a head coach before.

After a disappointing 4-8 season, Rocky Long resigned as New Mexico’s head coach in November of 2008. In December, the Lobos named Illinois offensive coordinator Mike Locksley head coach. Upon his hire, Locksley was not even forty years old. Couple his relative youth with a solid showing at Illinois despite the handicap of working for Ron Zook and the strength of the New Mexico program (seven straight seasons of at least six wins prior to 2008), and success seemed likely. Oh, how wrong we were. Locksley’s first two New Mexico teams won one game each and his third team began the season 0-4 before he was relieved of his duties. During his time at New Mexico, Locksley was also involved in several scandals. He was the subject of an age and sex discrimination lawsuit, suspended for an altercation with an assistant coach, and a friend of his son (his son was a member of the Lobo football team) was arrested for DWI while driving a car registered to his son. A friend of mind lived in Albuquerque and did some post-graduate studies during the time Locksley was head coach and we texted a bit when he was hired at Maryland. We were both surprised and my friend humorously proclaimed that Locksley had ‘twice as many scandals as wins at New Mexico’. To be fair, that is not entirely accurate, as my count puts the win versus scandal ratio at about even. But I digress. After being fired from New Mexico, Locksley became the offensive coordinator at Maryland under Randy Edsall. After Edsall was fired halfway through the 2015 season, Locksley was named interim head coach and proceeded to lead the team to a 1-5 mark. Thus, through parts of four seasons as a head coach, Locksley has compiled a 3-31 record. How often do coaches with a career winning percentage that low get a second chance? To find out, I looked at all FBS coaching hires since 1984 (what I deem the modern era thanks to the Supreme Court decision in the NCAA v. Board of Regents of the University of Oklahoma case that allowed for more football on television) where the new coach had been a previous FBS head coach. Locksley has the worst winning percentage by far. Here are nine other coaches that round out the bottom ten (note that I did not include ties in winning percentage, I simply ignored them because this isn’t soccer).
When I said Locksley was the worst by far, that was not hyperbole. His career winning percentage is roughly half that of the next lowest coach, Buddy Teevens. There are some interesting names on this list. You have a Pittsburgh Steelers defensive coordinator, a future college football national championship winning head coach, Mack Brown’s less famous (and less successful) brother, a future NFL head coach and sound bite legend, and a Michigan Man to name a few. But what we really want to know is how these coaches did once they were hired. Ask and you shall receive.
Gary Moeller and Gene Chizik were the most successful at their new stops, although Chizik’s success without Cam Newton was severely muted. Bob Toledo and Walt Harris also had decent runs of success with Toledo’s UCLA team winning twenty consecutive games over two seasons. However, the other five coaches combined for just one winning season in 22 years.

So what is a reasonable expectation for Locksley? Well, he won’t be leading a blue blood program like Moeller and he is unlikely to have a generational college player like Cam Newton fall into his lap, so Bob Toledo or Walt Harris is probably the ceiling. Or it would be the ceiling if Locksley was coaching in a different division. As currently structured, Locksley’s Maryland teams will have to play a third of their schedule against Michigan, Michigan State, Ohio State, and Penn State. Couple that quartet with three cross-division games against a group of teams that appear to be trending upward and getting to a bowl game is likely to require a Herculean effort.

This hire isn’t doomed to fail (although I believe it will). Moeller and Chizik proved that in the right situation a failed head coach can enjoy success the second time around. They didn’t quite qualify for the list, but Mack Brown (11-23 at Tulane), Gene Stallings (27-45-1 at Texas A&M), and even our old pal Ed Orgeron (16-27 at Ole Miss and Southern Cal) enjoyed success when given a second chance. Once again though, there are extenuating circumstances. Stallings and Orgeron coached blue-blood programs in their second chances and Mack Brown had just taken Tulane to a bowl game before being hired by North Carolina. However, for a variety of reasons, Locksley and Maryland do not appear to be a good match. For starters, Maryland has not exactly nailed their last few coaching decisions. They forced out one of the best coaches in school history following the 2010 season and in 2015 hired a coach who was at best negligent and at worst a Neanderthal asshole. Plus, Locksley did not just have minimal success in his previous opportunities as a head coach, he had no success. And as detailed earlier, there were a few off the field issues. To his credit, Locksley did coordinate perhaps the best offense in Alabama history this past season, so if he can match that talent level at Maryland, he should do fine.

In all seriousness, the lack of minority head coaches in college football is something that needs to be addressed. However, hiring retreads with a shitty track record as a head coach, regardless of race, does not help qualified coaches of color get their shot at a head coaching job.

Thursday, January 31, 2019

2018 Adjusted Pythagorean Record: ACC

Last week we looked at how ACC teams fared in terms of yards per play. This week, we turn our attention to how the season played out in terms of the Adjusted Pythagorean Record, or APR. For an in-depth look at APR, click here. If you didn’t feel like clicking, here is the Reader’s Digest version. APR looks at how well a team scores and prevents touchdowns. Non-offensive touchdowns, field goals, extra points, and safeties are excluded. The ratio of offensive touchdowns to touchdowns allowed is converted into a winning percentage. Pretty simple actually.

Once again, here are the 2018 ACC standings.
And here are the APR standings sorted by division with conference rank in offensive touchdowns, touchdowns allowed, and APR in parentheses. This includes conference games only with the championship game excluded.
Finally, the ACC teams are sorted by the difference between their actual number of wins and their expected number of wins according to APR.
I use a game and a half as a line of demarcation to determine if teams drastically over or under perform their APR. By that standard Miami was the only team that saw their actual record significantly diverge from their APR. The Hurricanes undershot their expected record because three of their four conference losses came by one score while three of their four conference wins came by at least 21 points. They wasted a fine defensive performance (second to Clemson in touchdowns allowed) as the offense was unable to consistently punch the ball in the endzone (only Louisville and Florida State scored fewer touchdowns).

How Dominant was Clemson in 2018?
A little less than four weeks ago, Clemson won their second national title in the past three seasons (and third overall). The Tigers dominated (on the scoreboard if not in the box score) an Alabama team that many thought might be one of the best of all-time. Clemson was touted as one of the best teams in the nation all season, but with the Tide sucking most of the oxygen out of the college football ecosystem, I feel like most casual football fans didn’t realize how dominant Clemson was this (I know I didn’t realize it until I was crunching the numbers for bowl season). The Tigers did survive a few tight games in 2018, edging Texas A&M in College Station and rallying to beat their Orange adversaries in Death Valley. However, those games share a common thread: quarterback Trevor Lawrence did not both start and finish them. In games Lawrence both started and finished, the Tigers won by an average of 36 points per game, with no team coming closer than twenty points! Scoring margin is one estimator of team strength, but since I have YPP and APR data going back to 2005, I decided to see where Clemson ranked in those categories compared to the other thirteen national champions in that span. Let’s start with YPP.
Using YPP, Clemson was the most dominant national champion since 2005. In that time frame, only 29 power conference teams (BCS or Power 5) have averaged at least seven yards per play against their conference opponents. Only 14 have allowed fewer than four yards per play to their conference opponents. Clemson is one of only two teams to do both. The other was Florida State in 2012 (the year before they won the national title). So Clemson ranks quite well using YPP. What about APR?
After ranking first in YPP, Clemson drops all the way to…tied for first in APR (notice I used APR as a win percentage to account for potential differences in number of conference games). The Tigers scored six times as many touchdowns as they allowed in conference play, tying them with a dominant national champion that didn’t even win their conference.

The ACC was mediocre at best in 2018, but Clemson thoroughly dominated it, and with their non-conference performance (victories against two SEC bowl teams as well as a solid Sun Belt squad) and subsequent playoff thrashing of two unbeaten heavyweights, the Tigers can make a case they are the best national champion of the new century. A significant portion of the population in the city where I live probably won't like it, but Orange might be the new black (shirt).

Thursday, January 24, 2019

2018 Yards Per Play: ACC

We are now into the third week of our offseason recap and we move to our second conference and home of the current national champion, the ACC.

Here are the ACC standings.
So we know what each team achieved, but how did they perform? To answer that, here are the Yards Per Play (YPP), Yards Per Play Allowed (YPA) and Net Yards Per Play (Net) numbers for each ACC team. This includes conference play only, with the championship game not included. The teams are sorted by division by Net YPP with conference rank in parentheses.
College football teams play either eight or nine conference games. Consequently, their record in such a small sample may not be indicative of their quality of play. A few fortuitous bounces here or there can be the difference between another ho-hum campaign or a special season. Randomness and other factors outside of our perception play a role in determining the standings. It would be fantastic if college football teams played 100 or even 1000 games. Then we could have a better idea about which teams were really the best. Alas, players would miss too much class time, their bodies would be battered beyond recognition, and I would never leave the couch. As it is, we have to make do with the handful of games teams do play. In those games, we can learn a lot from a team’s YPP. Since 2005, I have collected YPP data for every conference. I use conference games only because teams play such divergent non-conference schedules and the teams within a conference tend to be of similar quality. By running a regression analysis between a team’s Net YPP (the difference between their Yards Per Play and Yards Per Play Allowed) and their conference winning percentage, we can see if Net YPP is a decent predictor of a team’s record. Spoiler alert. It is. For the statistically inclined, the correlation coefficient between a team’s Net YPP in conference play and their conference record is around .66. Since Net YPP is a solid predictor of a team’s conference record, we can use it to identify which teams had a significant disparity between their conference record as predicted by Net YPP and their actual conference record. I used a difference of .200 between predicted and actual winning percentage as the threshold for ‘significant’. Why .200? It is a little arbitrary, but .200 corresponds to a difference of 1.6 games over an eight game conference schedule and 1.8 games over a nine game one. Over or under-performing by more than a game and a half in a small sample seems significant to me. In the 2018 season, which teams in the ACC met this threshold? Here are ACC teams sorted by performance over what would be expected from their Net YPP numbers.
The ACC saw a pair of teams significantly over-perform relative to their expected records (Syracuse and Georgia Tech) and one team significantly under-perform (North Carolina). In addition, four other teams (Florida State, Miami, NC State, and Virginia Tech) nearly met the threshold for under or over-performance. Syracuse’s conference record was buoyed by their league-leading in-conference turnover margin of +11 while Georgia Tech was second at +7. The Yellow Jackets also posted a solid, but not extreme 2-1 mark in close games. Meanwhile, North Carolina managed a 1-4 mark in close conference games, losing a pair in overtime (Syracuse and NC State). In addition, six of North Carolina’s seven conference losses came by ten points or fewer.

Worst Over First
Speaking of North Carolina, back on September 22nd, the Tar Heels beat the Pitt Panthers in their conference opener to improve to 1-2 on the young season and provide a modicum of hope for their beleaguered fan base. The loss dropped Pitt to 2-2 overall (1-1 in the ACC) and the game seemed destined for the dustbin of history. There was nothing really remarkable about North Carolina’s victory at the time. And then the rest of the season played out. North Carolina lost their last seven conference games, with their only victory post-Pitt coming against an FCS opponent the weekend before Thanksgiving. Meanwhile, Pitt won five conference games in a row and thanks to poor play by the presumptive division favorite, the Panthers had the division locked up the time they traveled to Coral Gables over Thanksgiving weekend. The victory by the Tar Heels meant the worst team in the division (by record) had beaten the division champion. How often has this happened in a BCS or Power Five conference? The answer: Not very, and the ACC has been involved in most of them. The table below lists those occurrences, and since it has only happened eight times including North Carolina's win in 2018, I’ll regale you with some background on each.

November 5th, 2005 
NC State 20 Florida State 15 
Florida State entered this game 7-1 overall and 5-1 in the ACC. NC State was 3-4 (1-4 in the ACC) and fighting for their bowl lives. Befitting a 20-15 final score, the quarterback play was horrid with Marcus Stone, Drew Weatherford, and Xavier Lee combing for 269 passing yards on 62 attempts. NC State lost to Boston College the next week, but won their last two games to finish bowl eligible. They won the prestigious Meineke Car Care Bowl in Charlotte after the regular season. Florida State lost their next two games following their home loss to the Wolfpack, losing the Bowden Bowl to Clemson and the annual rivalry game with Florida (in Urban Meyer’s first season as coach) before upsetting Virginia Tech in the inaugural ACC Championship Game. The Seminoles closed the year losing a tight Orange Bowl to Penn State in a battle of old ass head coaches who would go on to ruin their legacies by supporting Donald Trump and abetting a child molester (its an open question which is worse). While NC State finished tied for last in the Atlantic Division of the ACC, their 3-5 record was only two games worse than Florida State’s.

November 10th, 2007 
Maryland 42 Boston College 35 
In a wild and crazy college football season, Boston College had been ranked second in the country the week before, but an upset home loss to Florida State dropped them to number eight. Maryland entered the game 4-5 (1-4 in the ACC), having lost three in a row. The Terps led by 21 entering the fourth quarter and the final score was not indicative of the quality of the game as Boston College scored with under a minute left to cut the lead to seven. Maryland split their next two games to finish bowl eligible while Boston College won their last two games to clinch a rematch with Virginia Tech in the ACC Championship Game. The Eagles lost to the Hokies and had to settle for a Champs Sports Bowl appearance.

November 1st, 2008  
Clemson 27 Boston College 21 
In a wide-open ACC Atlantic, Boston College entered this game 5-2 overall (2-2 in the ACC). Clemson was a preseason top-ten team, but the Tigers entered on a three-game losing streak. This marked just the second game as head coach for Dabo Swinney (whatever happened to him?) who replaced Tommy Bowden following a loss to Wake Forest. Clemson scored ten points in the final eight minutes to give Dabo his first win as Clemson head coach. Clemson would lose to Florida State the following week, but win their final three regular season games to salvage a bowl bid and get Dabo the full time job. Boston College rebounded from the loss to win their final four regular season games and earn another shot at Virginia Tech. The Eagles would lose to the Hokies once again and have to settle for a Music City Bowl bid. 2008 was a weird year for the ACC with eleven of the twelve conference teams finishing within one game of .500 in league play, so Clemson only finished a game behind the Eagles, but were technically tied for last in the division.

October 3rd, 2009 
Maryland 24 Clemson 21 
The first full season of the Dabo Swinney era saw Clemson lose a pair of tight early season contests to ranked teams (Georgia Tech and TCU). The Tigers headed to Maryland with a 2-2 record (1-1 in the ACC) to face a team that was 1-3. The Terps had been blown out by Cal and Rutgers while also losing to Middle Tennessee State. Their lone win had come against James Madison of the FCS. But the Terps played their best game of the season by far against the Tigers, scoring three straight touchdowns in the second and third quarters to lead by eleven. CJ Spiller ran a kickoff back in the third quarter to cut the lead to three, but Clemson missed a field goal late in the fourth and Maryland held on. Maryland would use the momentum generated by this upset to lose their final seven conference games. Meanwhile, Clemson won six in a row following this defeat to set up a rematch with Georgia Tech in the ACC Championship Game which they also lost.

October 16th, 2010 
Texas 20 Nebraska 13 
This game is special because it is the only instance of a last place team from a different division beating an eventual division winner. In 2010, Nebraska began the season in the top-ten (might we see a repeat in 2019?). A 5-0 start against a relatively easy schedule saw them peak at number five. The Cornhuskers had a chance to avenge their 2009 Big 12 Championship Game loss when Texas came to Lincoln. There were already signs this might not be your typical Texas team as the Longhorns were coming off back-to-back losses to UCLA and Oklahoma. However, the Longhorns kept the Nebraska offense out of the endzone all day (their lone TD came on a punt return) and pushed themselves back into the top-25. Following this first (of many) ‘Texas is back’ proclamations, the Longhorns lost five of their last six games, with their lone win in that span coming against Florida Atlantic. Meanwhile, Nebraska won five of their last six regular season games before falling to Oklahoma in the final Big 12 Championship Game (the revived Big 12 Championship Game does not count and is dumb).

October 20th, 2011 
Arizona 48 UCLA 12 
This one shouldn’t really count since UCLA only won the division because Southern Cal was ineligible. In addition, this game is probably more famous for another reason. 

October 13th, 2017 
Syracuse 27 Clemson 24 
The defending national champion Clemson Tigers were 6-0 and ranked second in the country when they made a Friday night visit to the Carrier Dome. Syracuse was 3-3 and looking to qualify for a bowl game in Dino Babers’ second season. The Orange never led by more than seven points, but also never trailed in a game where they outgained the Tigers by more than 100 yards. After holding the Tigers in check, Syracuse lost their last five games, with their defense allowing 162 points in the final three. Clemson won their last six games, with five coming by double-digits, and qualified for the College Football Playoff for the third straight year.

When it comes to shocking division upsets, no conference does it quite like the ACC.

Thursday, January 17, 2019

2018 Adjusted Pythagorean Record: AAC

Last week (or more accurately, a few days ago), we looked at how AAC teams fared in terms of yards per play. This week, we turn our attention to how the season played out in terms of the Adjusted Pythagorean Record, or APR. For an in-depth look at APR, click here. If you didn’t feel like clicking, here is the Reader’s Digest version. APR looks at how well a team scores and prevents touchdowns. Non-offensive touchdowns, field goals, extra points, and safeties are excluded. The ratio of offensive touchdowns to touchdowns allowed is converted into a winning percentage. Pretty simple actually.

Once again, here are the 2018 AAC standings.
And here are the APR standings sorted by division with conference rank in offensive touchdowns, touchdowns allowed, and APR in parentheses. This includes conference games only with the championship game excluded.
Finally, the AAC teams are sorted by the difference between their actual number of wins and their expected number of wins according to APR.
I use a game and a half as a line of demarcation to determine if teams drastically over or under perform their APR. By that standard Tulsa was the only team that saw their record differ significantly from their APR. Tulsa finished with a losing record for the second consecutive season, putting Phillip Montgomery on the hot seat and costing me a few bucks. Tulsa was 1-3 in one-score conference games, including a heartbreaking 25-24 loss against South Florida where they blew a double-digit fourth quarter lead.

A Definitive Historical Examination of the Defense of the 2018 Connecticut Huskies (Peer Reviewed)
Everybody knows Connecticut had a terrible defense last season. But how bad was it? I’m going to make some points of comparison so you can decide for yourself just how bad their defense was.

If you read last week’s post, you know that in conference play, the Huskies collectively allowed AAC opponents to average 8.94 yards per play against them. That number was more than two yards per play below Navy, the second worst defense in the AAC last season. So over, the course of a few hundred defensive plays last season, Connecticut opponents averaged about nine tenths of a first down per play! That was by far the worst in-conference per play defense at least since 2005 (when I began tracking such things).
In addition, have been only four offenses to average at least eight yards per play over the course of a conference season (out of nearly 1700 total team seasons). If we take the collective performance of the AAC teams on Connecticut's schedule, they would boast one of the best offenses of the past fourteen years.
Basically, the 2018 Huskies made opposing quarterbacks look like some amalgamation of Colt Brennan, Case Keenum, Baker Mayfield, and Kyler Murray. However, not only did the Huskies allow a lot of yards, opponents frequently turned those yards into points. AAC opponents scored 55 touchdowns against the Huskies in eight conference games. If you are lacking a calculator, that is nearly seven touchdowns per game or about one and three quarters per quarter. For comparison, the best defense in the AAC (Cincinnati and UCF) allowed 18 touchdowns, or roughly a third of that amount! As with the YPP data, I have been tracking APR data since 2005, and Connecticut was the worst in terms of touchdowns allowed.
Only Bob Davie’s second New Mexico team comes close to catching the Huskies. However, that New Mexico team can take solace in the fact they allowed merely 7.68 yards per play to their Mountain West opponents.

So, yes, we can conclude that Connecticut was historically bad defensively in 2018. There is no question about that. The good news for Connecticut fans (if any) is that things cannot possibly get worse (at least defensively).

Monday, January 14, 2019

2018 Yards Per Play: AAC

College football season always passes by so fast. It seems like just yesterday I was lounging on the sofa enjoying some Week Zero action. And now we have no college football for nearly eight months. Don’t worry though, we’ll get through this together. This post begins our offseason sojourn through each of the ten FBS conferences. We’ll examine each conference in terms of Yards per Play (YPP yeah you know me) and the Adjusted Pythagorean Record (APR). Don’t know that those two acronyms mean? Read on to find out. We start as always, with the American Athletic Conference.

Here are the AAC standings.
So we know what each team achieved, but how did they perform? To answer that, here are the Yards Per Play (YPP), Yards Per Play Allowed (YPA) and Net Yards Per Play (Net) numbers for each AAC team. This includes conference play only, with the championship game not included. The teams are sorted by division by Net YPP with conference rank in parentheses.
College football teams play either eight or nine conference games. Consequently, their record in such a small sample may not be indicative of their quality of play. A few fortuitous bounces here or there can be the difference between another ho-hum campaign or a special season. Randomness and other factors outside of our perception play a role in determining the standings. It would be fantastic if college football teams played 100 or even 1000 games. Then we could have a better idea about which teams were really the best. Alas, players would miss too much class time, their bodies would be battered beyond recognition, and I would never leave the couch. As it is, we have to make do with the handful of games teams do play. In those games, we can learn a lot from a team’s YPP. Since 2005, I have collected YPP data for every conference. I use conference games only because teams play such divergent non-conference schedules and the teams within a conference tend to be of similar quality. By running a regression analysis between a team’s Net YPP (the difference between their Yards Per Play and Yards Per Play Allowed) and their conference winning percentage, we can see if Net YPP is a decent predictor of a team’s record. Spoiler alert. It is. For the statistically inclined, the correlation coefficient between a team’s Net YPP in conference play and their conference record is around .66. Since Net YPP is a solid predictor of a team’s conference record, we can use it to identify which teams had a significant disparity between their conference record as predicted by Net YPP and their actual conference record. I used a difference of .200 between predicted and actual winning percentage as the threshold for ‘significant’. Why .200? It is a little arbitrary, but .200 corresponds to a difference of 1.6 games over an eight game conference schedule and 1.8 games over a nine game one. Over or under-performing by more than a game and a half in a small sample seems significant to me. In the 2018 season, which teams in the American Athletic Conference met this threshold? Here are AAC teams sorted by performance over what would be expected from their Net YPP numbers.
Memphis was the lone AAC team to under-perform relative to their YPP numbers. The Tigers actually bested Central Florida in terms of YPP, although they were not able to do so on the field. How did the Tigers lose three times despite such pristine numbers? As is usually the case, you can blame close games. The Tigers dropped a pair of one-point decisions in conference play, including the aforementioned defeat to Central Florida. In their other one-point loss (to Navy), the Tigers outgained the Midshipmen by nearly four yards per play, but were done in by untimely turnovers and an inability to finish drives. On the other end of the spectrum, despite not registering a single conference win, Connecticut actually exceeded their expected record (by a significant amount) based on YPP. The Huskies were so bad defensively (more on that next week), they actually break the regression analysis and should have won about negative two games instead of zero.

In these weekly posts, this is where I take a deeper dive into the statistical minutia of the conference. However, I noticed over the previous offseasons that I just sort of shifted gears into the next topic. No more. We are professionals here on this nearly fourteen year old blogspot account. And there will be headings!

Bettor Beware!
This may have gone unnoticed by you over the bowl season, but I caught it since the team I root for (Wake Forest) was involved. Four ACC teams played in bowl games where they were betting underdogs to Group of Five teams. Obviously, this probably says a lot about the quality of the ACC in 2018, national champ notwithstanding. Three of those ACC teams were underdogs to teams from the AAC. All three of the ACC teams covered and two won outright (the other ACC team had their game canceled). Is the 2018 postseason an isolated incident or are Group of Five teams a bad bet when they are favored against Power Five teams in bowl games? Let’s investigate!

Beginning with the 2005 season, I looked at all bowl games where a non-BCS or Group of Five team was a betting favorite against a BCS or Power Five team. It should go without saying, but the team labels are time sensitive. In other words, while TCU is currently a member of the Big 12, in 2005, when they played Iowa State in the Houston Bowl, they were members of the Mountain West Conference and are included in the analysis. Similarly, their Rose Bowl classic with Wisconsin is also included here (TCU won both games, but pushed against the Cyclones and failed to cover against Wisconsin).

So how did those ‘mid-majors’ do when matching up as a favorite against ‘major’ conference opponents? Since 2005, mid-majors have been favored 29 times against major conference opponents. Overall, they have done well, winning 18 of those 29 contests straight up. However, their Against the Spread (ATS) record is another issue. Mid-majors are just 11-17-1 ATS. The following table breaks the numbers down chronologically with an obvious arbitrary separation.
Hopefully your small sample alert siren blared when looking at the BCS versus playoff era records, but real or not, mid-majors have struggled to win those games outright since the dawning off the College Football Playoff in 2014. This will bear watching in the coming years.

It does appear that Group of Five teams might not be the ideal choice to back in bowl season when they are favored against Power Five teams. Why might this be so? I have a few theories.

  1. Motivation. This works for both teams. In the typical G5/P5 clash, the P5 team is usually expected to win and despite the best efforts of their coaching staff, may take the G5 team lightly. Conversely, the G5 team probably has the proverbial chip on their shoulder and is ready to face a team that is higher than them in the college football hierarchy. Some of the individual players from the G5 team may even have extra motivation if both teams are from the same region and the G5 players were not recruited by the P5 team. These roles are subverted in the bowl games where the G5 team is favored. In such cases, the G5 team has usually had a good, but not great season. Perhaps they are disappointed they did not qualify for a New Year’s Six Bowl and are taking their P5 opponent lightly. Meanwhile the P5 team has likely had a disappointing season and may even need a victory in the bowl game to salvage a winning record. Knowing they are underdogs to an ‘inferior’ opponent could provide them with extra motivation in the preparation for and actual playing of the bowl game. 
  2. Loss of quality head coaches. When G5 teams have good seasons, they often lose their coaches. Those coaches often leave between the end of the regular season and the playing of the bowl game. When those coaches leave, an interim is usually appointed to coach the bowl game before the permanent coach and his staff arrive. I never played organized football, but it would seem quite possible this interim coach, despite his qualifications, may have all the respect that a substitute teacher engenders in a high school classroom. Through no fault of the interim coach, the preparation that went into games during the regular season is probably lacking during bowl season. Despite perhaps being the better team, the lack of an effectual (permanent) leader at the top could harm these G5 teams in their bowl games. I didn’t do a deep dive into interim coaches at G5 programs, but anecdotally, the Temple Owls have been favored against P5 teams in two of the last three seasons with an interim coach in charge. The Owls lost both of those games straight up despite being solid to heavy favorites. 
  3. Talent level. Cincinnati was better than Virginia Tech in 2018. That is not a controversial statement. Over the course of twelve games, their actual performance along with randomness inherent in any college football season (in-game turnovers, weather, player dismissals, injuries, etc.) rightly made them a betting favorite against Virginia Tech. However, talent-wise these teams were much closer and Virginia Tech may have had more raw talent. Recruiting rankings are not an infallible proclamation of team strength, but they do a good job of estimating a team’s baseline talent level and in the aggregate, P5 teams have more talent than G5 teams. Sure, some G5 teams will out-recruit P5 teams, but in a random bowl matchup, the P5 team will usually have more talent than the G5 team. 
Those are just a few thoughts I had as to why G5 teams have struggled when favored in bowl games against P5 teams. If you have any thoughts, please drop them in the Comments section.

Finally, I apologize for the delay in this post, as I was under the weather last week. However, your patience will be rewarded as there will be two posts this week. The APR AAC post will be up on Thursday and barring any health challenges, posts will resume on a weekly basis every Thursday. Thanks for reading.

Sunday, January 06, 2019

Strangers in the Field Part IV: How'd We Do?

As I have done the past few years, I went to Vegas over the summer and made some wagers. Here is how they turned out.

Over/Under Win Totals

Ball State over 4 wins -105 ($40 to win $38.10)
This ended up being a push. Ball State was still a pretty bad team in 2018, but the MAC was a pretty bad conference and the Cardinals were probably fortunate to win four games, although they could have cleared the bar had they beaten a bad Western Kentucky in the non-conference.

Central Florida under 9 wins +115 ($30 to win $34.50)
I will never ever bet on UCF to go under their win total. Despite a new coach and some big defensive losses, the Knights were only really challenged once in conference play (at Memphis).

Coastal Carolina over 4 wins -105 ($40 to win $38.10)
The Chanticleers somehow beat the Conference USA champion (UAB) and the Sun Belt’s West division winner (Louisiana-Lafayette) to cash here. They were still a pretty bad team and I was quite fortunate to not add this one to the loser’s pile.

Florida International under 5 wins -110 ($30 to win $27.30)
Not even close. FIU easily surpassed this low win total despite major attrition from last season’s team that was fortunate to qualify for a bowl.

Georgia State under 4.5 wins even ($30 to win $30)
This was one of my few good calls in the win total department. The Panthers came perilously close to losing to Kenesaw State in their opener and never threatened to go over this number.

Georgia Tech over 6 wins even ($30 to win $30)
Hope seemed lost after the Yellow Jackets stumbled out of the gates with losses to South Florida and Pitt. At 1-3 with Georgia waiting at the end of their schedule, the Yellow Jackets won out with the exception of Duke to send Paul Johnson off with seven wins. However, Nate Woody was not able to fix the defense as I predicted he would when making this play.

Indiana over 5 wins -115 ($20 to win $17.40)
The Hoosiers managed a push and even had a shot to go over their win total in the finale against Purdue. Looking back, this was probably not a good play as the Hoosiers were fortunate to beat an improved Virginia team to manage said push.

Iowa over 7.5 wins -105 ($40 to win $38.10)
The Hawkeyes were much better than their 8-4 regular season record, but after three close losses in a row (Penn State, Purdue, and Northwestern), they needed to beat Nebraska in the season finale to get to eight wins. A field goal at the gun pushed them over the total.

North Carolina over 5.5 wins -135 ($20 to win $14.80)
This was an awful bet, but I was bailed out thanks to their canceled game against Central Florida. This bet was made at a different casino (Golden Nugget) than the Central Florida bet (Southpoint) so it ended up being a push instead of a loss.

Old Dominion over 5.5 wins even ($150 to win $150)
This was my biggest bet and it was basically dead in the water after the opener. The Monarchs were not good at all (99th in Bill Connelly’s S&P+ ratings), but the bet was somewhat justified by their easy schedule. They lost to teams ranked 111th (East Carolina), 112th (Charlotte), 118th (liberty), and 129th (Rice) in S&P+. If they won half of those games, this bet cashes.

San Jose State over 2.5 wins -125 ($60 to win $48)
Of all my losers, this is probably the bet I would be most likely to make again. San Jose State had an extreme negative turnover margin in 2017 and still managed to win one conference game. In 2018, I figured they could double that conference win total and beat UC-Davis from the FCS. They opened the year losing to UC-Davis and then managed just one win all season.

Tulane over 5.5 wins +115 ($40 to win $46)
It took a gamble on a two-point conversion, but the Green Wave managed to go over their win total despite an option offense that continued to struggle in their third season under Willie Fritz.

Tulsa over 4 wins -130 ($40 to win $30.80)
I expected a bounce back and Tulsa did not deliver. Their offense continued to stink and they eventually turned the reigns over to an inaccurate freshman quarterback. However, had the Golden Hurricane not choked away a lead against South Florida, they would have pushed.

UCLA over 5.5 wins even ($30 to win $30)
Not even close. Cincinnati and Fresno State were better than expected meaning the Bruins would have needed six conference wins to eclipse their season win total. I didn’t lay a lot on this, but I should have stayed away altogether.

Western Michigan over 6 wins -135 ($40 to win $29.65)
This one looked like it was going to cash easily, but the Broncos lost their starting quarterback after a 6-2 start and lost three straight before upsetting Northern Illinois in the regular season finale.

Games of the Year

September 15th Boise State +4.5 Oklahoma State -110 ($25 to win $22.75)
I bought at a good price as this line opened at Oklahoma State -3 and trickled down before kickoff. However, the Cowboys blew the Broncos away kicking off what would be a very weird year for Mike Gundy’s team.

October 6th Florida State +6 Miami -110 ($25 to win $22.75)
I bought at a horrible price as this line was nearly a touchdown higher by kickoff. Florida State played perhaps their best game of the season through three quarters and this looked to be an easy cover. Then turnovers and a lack of offense nearly swung this back to the Hurricanes. Miami won, but took a knee inside the ten to run out the clock and preserve this undeserved cover.

November 10th Tennessee -1 Kentucky -110 ($25 to win $22.75)
This handicap was both good and bad. I thought Tennessee would be better than they were in 2018 and I had no inkling that Kentucky would be as good as they were, particularly on defense. However, I was able to identify a spot for a letdown as the Wildcats were coming off a date with Georgia (that ended up being for the SEC East title). Kentucky played their worst game of the season and Tennessee won rather easily despite being a touchdown underdog.

November 24th South Florida +4 Central Florida -110 ($25 to win $22.75)
As I mentioned earlier, I am never going against Central Florida again. The Knights were better than I thought and while I figured South Florida was due for some regression, I figured they would have their act together by the end of the season. This ticket never came close to cashing.

Conference Champion Bets

Georgia Southern to win the Sun Belt +1000 ($10 to win $100)
This was a pretty good bet as Georgia Southern upset Appalachian State at home to take control of their division race. However, a road loss to Louisiana-Monroe and a home loss to Troy prevented them from taking the division. Had they won either of those games, they likely would have qualified for the championship game and been favored to win it.

Louisiana Tech to win Conference USA +700 ($10 to win $70)
The Bulldogs upset North Texas early in the season and figured to be the frontrunner following that game. However, UAB handled them the very next week on their way to easily winning the division and (not as easily) the conference title.

Miscellaneous Bets

Atlanta Braves to win World Series +2500 ($10 to win $250)
The Braves won the NL East without much trouble as the Phillies faded, but never seriously threatened the Dodgers in the Division Series.

Oakland Athletics to win World Series +3000 ($10 to win $300)
The A’s may have been the third best team in all of baseball, but they were unable to beat the Yankees in a winner take all Wild Card Game.

Pittsburgh Pirates to win World Series +6000 ($10 to win $600) The Pirates made a brief run right before I made this bet, but faded shortly after.

Reckless Parlay 1:
$10 to win $110

Game 1: August 31st Utah State +25 Michigan State
Winner.

Game 2: September 1st Miami (Ohio) -2 Marshall
Nope.

Game 3: September 1st Coastal Carolina +30 South Carolina
Nope.

Game 4: September 1st Louisville +25.5 Alabama @ Orlando
LOL. Nope.
One out of four.

Reckless Parlay 2:
$10 to win $60

Game 1: September 9th Cleveland +6 Pittsburgh
Winner.

Game 2: September 9th Cincinnati +3 Indianapolis
Winner.

Game 3: September 9th Washington +1 Arizona
Winner.
Winner. Despite the presence of Hue Jackson, the Browns came through for me.

Money Wagered: $810
Money Won: $657.35
Return on Investment: -18.85%

This was by far my worst summer of betting. Despite my poor predictions regarding win totals and line moves, had Old Dominion come through for me, I would have actually made a profit. Its back to the drawing board. Hopefully we can get back to profitability next year. Sorry if you tailed. :)

Tomorrow marks a rather somber day for me as it is the final college football game until August. However, Statistically Speaking will help you get through the long offseason by starting the YPP and APR conference reviews on Thursday. For those that aren't  regular readers, we'll review each FBS conference through the lens of Yards per Play (YPP) and the Adjusted Pythagorean Record (APR) with one post per week. As always, we'll go alphabetically starting with the AAC. That will get us close to Memorial Day and then I'll have some more sporadic posts over the summer until the football season begins anew. As always, thanks for reading and feel free to drop a comment should you feel the urge. See you on Thursday.