College Bowl Games Prove CFP Committee Made A Bad Call On Alabama & Replay Of Initial Call Needed
The initial call in the conference room by the CFP selection committee was to rank Alabama #6 going into the conference championship weekend in early December. That missed call made it virtually impossible for Alabama to make it into the playoffs. Even with two teams among the top four losing their conference championship games, only one was displaced. That created an opening for #5 Ohio State to make it into the top four in the final playoff ranking. While TCU lost its conference championship game to Kansas State, it remained unchanged at #3 in the final ranking. A replay of the data will show that the #5 slot should have gone to Alabama. Had that correct call been made, as the events unfolded that would have resulted in the Crimson Tide being in the playoffs and Americans getting the enjoyment of watching an extraordinarily talented team.
My November 28 article (“The Statistical Case For Selecting Alabama To Be In The College Football Playoffs”) written before the week 12 rankings demonstrated based on the numbers that putting Alabama #5 was warranted. That analysis was augmented by my December 1 article (“Epilogue To The Statistical Case For Selecting Alabama To Be In The College Football Playoffs”) written after the week 12 rankings came out. It made an even stronger analytical case that Alabama should have been ranked #5, if not even higher, going into the conference championship weekend. Before we go to through a replay analysis to point out the errors and missed data points that led to the bad call, it’s worth highlighting the scores from the bowl games last weekend. They speak for themselves and in volumes in highlighting that Alabama should have been in the top four in the final CFP committee ranking.
In the Sugar Bowl, Alabama shellacked Kansas State 45–20, a spread that was more than three times and 18 points wider than the line on that game. That is the same Kansas State team that in its prior game beat favored TCU 31–28 in the Big 12 conference championship. Based on those two data points from adjacent games, a second derivative analysis would point to Alabama beating TCU by 28 points. Alas, a TCU team that would be no match for Alabama now finds itself going to the national championship game against Georgia. In the Orange Bowl, Tennessee beat Clemson 31–14 despite being a 4-point underdog, translating into a 21 point better performance compared to the spread. LSU pulverized Perdue 63–7 in the Citrus Bowl, an improvement of 42 points on the two touchdowns by which LSU was favored.
I highlight those other two SEC bowl game wins as those are the two teams Alabama lost to by a combined 4 points in the last two seconds of one game and one second of the other game. The articles referenced above go into more detailed analysis that actually makes those narrow losses look even thinner, including a new video clip that shows an illegal block is what actually made the winning score possible for LSU. Beyond that, the striking wins by both teams highlight the strength of schedule factor for Alabama that was not accurately incorporated into the analysis by the selection committee. That LSU, which the numbers show was only the 4th best team in the SEC this season, beat the team that was in this year’s Big 10 championship 63–7 speaks a great deal about the SEC and where it ranks relative to the all the other conferences. To have just one SEC team in the playoffs this year at the same time that there were two Big 10 teams is one more irrational example of the striking lack of a rigorous analytical process by the selection committee.
After the selection committee made their call in the conference room, the games proceeded on the field so under the replay rules it’s too late to overturn the original bad call on Alabama. But it’s still instructive to review what happened to learn from the infractions. Importantly, further insights from a detailed analytical review of the data can result in a better selection process by the committee in the future. Consistent with the way plays are called on the field and to assist in making the points more understandable to the committee, we’ll be using some traditional symbols to highlight the errors that our analysis indicates were made. The symbols will be shown first followed by an explanation of the infraction.
False start related to the numbers. Let’s start by going to the video — or in this case, the picture of the committee in deliberation at the beginning of this article. While open laptops are shown in front of each of the 13 members of the selection committee, no more than 3 members or 23% appear to be actually engaged with their laptops. Furthermore, no laptop screens are visible. Are they really on? And if so, are they only showing Excel worksheets used to crunch the numbers? When the CFP started in 2014, its mission was to have subject matter experts dig deeper into the numbers with the singular goal of matching the four best college football teams in an end of year playoff. As experts, their assessment was supposed to be completely independent and better than polls often affected by regional bias. While the pictures show the selection committee in a conference room with laptops at the ready, there is no indication that rigorous use of those tools occurred. As such, at the start of deliberations there was a false impression that numbers and an analytical process would drive rankings when in fact unsystematic narratives drove the process.
Intentional grounding of a relevant fact. The reference to schedules by the committee has been anecdotal with a focus on selected games. At no time is there any indication of the sort of systematic evaluation of all teams and games based on the credible data now available that quantifies strength of schedule. That data could be put to work on the member’s laptops to make an array of adjustments to various other metrics, similar to how diving scores are adjusted based on a degree of difficulty factor. For instance, the TeamRankings strength of schedule factors are updated each week and at any given time measure the entire schedule, so they are comparable to other cumulative measures. The factor for Alabama before the conference championship weekend was 13.7, ranking it #1 among all 131 FBS college football teams. That factor was 1.21 times the same metric for Georgia, 1.29 times for Michigan, 1.34 times for TCU, 2.63 times for USC and 1.18 times for Ohio State. If there are better specific metrics on strength of schedule than the ones issued by TeamRankings, those could be used. But the key is to incorporate such specific available data to further refine the metrics to develop a systematic apples to apples comparisons of the best teams. In not doing that, a particularly relevant fact was ignored by the committee.
Roughing the analytical data. Even before incorporating precise strength of schedule data into the analysis, a more detailed review of the numbers would have supported ranking Alabama higher than the committee elected. After the first twelve games, Alabama ranked #4 among all teams in the CFP ranking in terms total season points scored versus given up at 2.26 times (TCU was #8 at just 1.69 times). Alabama ranked #5 in terms of the difference in points scored versus given up at an average of 22.8 per game (TCU was #7 at 16.8). Does a team with 2 losses for a total of 4 points really have a worse record than a team that lost 1 end of season game it was favored in by 22 points? What does the actual record look like based on the score at the end of each quarter instead of just at the end of each game? Isn’t the average difference in points scored versus points given up a more precise measure of actual performance over an entire season than just the won-loss record? These are all examples of analytically digger deeper into the data, all of which puts Alabama in a more favorable light than its 10–2 record. Instead, the selection committee substituted narrative for analysis with the overriding narrative that a 2 loss team had never been selected to be in the playoffs. In giving that prominence, they played roughshod with the actual data.
Personal foul on the viewing public. Ever since 1969 when Alabama played Mississippi in the first nationally televised prime time college football game in a 33–32 shootout with outstanding performances by Alabama QB Scott Hunter and Mississippi QB Archie Manning, Alabama has delivered great college football to living rooms across the country. ABC’s Roone Arledge would say that game quickly catapulted college football into a spectator sport with broader national interest. No team has been in more post-season games since then than Alabama and no team has delivered more post-season viewing excitement to the public than Alabama.
Related to actual CFP games, no team comes close to matching Alabama’s 9–4 record and its 9 wins is more than the cumulative wins of Georgia, Michigan, TCU and Ohio State combined. Performance when it matters most is the mark of champions and no college team has demonstrated that more than Alabama, either in the CFP era or in the decades going back to that monumental 1969 game. In not putting Alabama at least at #5 in the ranking prior to the championship games, the selection committee denied the American public what would have been a much more exciting playoff.
The promise of a new CFP system was that subject matter experts would rigorously analyze all the data to come up with the four best college football teams to square off in an end of season playoff. Instead of taking that approach, this year the selection committee has taken a less intensive approach that resulted in their bad calls concerning Alabama. Analysis was trumped by narrative, rarely a good situation related to any decision-making process. That the CFP ranking after 12 games was exactly the same through the first six positions as the AP and USA Today Coaches ranking, the two primary polls, and almost the same through the entire rankings suggests no new approach was followed. That multicollinearity isn’t a good sign. If the CFP was meant to duplicate the polls, its reason for existing is questionable. The CFP isn’t supposed to be like the polls, yet that is precisely what it looked like at the end of this year. The CFP comes out Tuesday night well after the polls have been released. How much is it influenced by those polls? Is there a conscious desire to make the CFP look like the polls?
Alas, the rankings of the selection committee are final and not appealable. For only the second time in the nine-year history of the CFP, the Crimson Tide in their classic uniforms did not take the field in the playoff games. Last weekend’s Sugar Bowl game not only underscored the bad call by the committee but gave college football fans across the country a taste of what they missed by Alabama not being in the playoffs this year. Who wouldn’t want to watch a college team with the best offensive player in the country (QB Bryce Young) and the best defensive player in the country (Linebacker Will Andersen)? Underscoring their extraordinary talents, they will likely be the #1 and #2 picks in the upcoming NFL draft. That Alabama is consistently at or very near the top in terms of NFL draft picks should come as no surprise as under any objective analysis it has the best head coach and coaching staff of any college in the country. The analysis backing up that claim is in my January 16, 2022 article entitled “The Statistical Case For Nick Saban Being The Greatest Coach Of All Time.”
While the missed call by the selection committee on Alabama cannot be unwound, something can and should be learned from the mistakes that were made as that insight can be used to improve the process. Among the takeaways this year is that real analysis, inclusion of precise strength of schedule data and rigorously looking beyond win-loss all move the process to the overriding goal of having the best and most exciting playoff games for college football fans across the country. Often the data points to consistent patterns that can result in the establishment of policies that further improve the process. Among the cardinal rules that a broad analysis of the facts suggest should be formally adopted and made part of CFP selection committee policy going forward are the three below.
1. The winner of the SEC championship game will automatically be seeded #1 in the final CFP playoff ranking, a permanent seat similar to the U.S. at the United Nations Security Council.
2. As long as Nick Saban is the head coach of Alabama’s team, it should be guaranteed to be among the teams selected for the college football playoffs.
3. At no time should any college football conference have more teams in the playoff than the SEC conference and it is expected that typically the SEC will field more teams than any other conference.
In tomorrow night’s CFP final between Georgia and TCU, the line has the bulldogs favored by 13 points over the Horned Frogs (factually they are horned lizards as horned frogs don’t exist). Georgia will blow past that spread for reasons detailed elsewhere. Indeed, Alabama is the only college football team that would have been able to compete with Georgia in the playoffs. But if Georgia fails to beat TCU by more than 28 points, a case can be made based on the earlier referenced second derivative analysis that Alabama would have beaten Georgia.
So, while Alabama fans are disappointed that the Crimson Tide wasn’t in the post-season playoffs this year, we’ve already turned our attention to next year. Like a new row of sharks’ teeth rolling into place, new talent is coming onboard. Once again Alabama has the top recruiting class in the country. But this time it appears to be its widest over-performance relative to other teams ever. Based on 247Sports rankings, for next years team Alabama will be getting 9 of the top 48 and 14 of the top 77 high school recruits in the country. We don’t yet know who among them will be future first round picks in the NFL draft, but there will most certainly be multiple players in that category. In the 13 seasons Nick Saban has coached Alabama, it has had 39 first round draft picks for an average of 3 per year. Given their high school rankings, it sounds like this incoming class of talent will beat that average and result in more than 3 first round NFL draft picks in three or four years.
Looking forward to the Crimson Tide showing what it can do next season!
Jock McCown is a native of Alabama who now lives in Pound Ridge, which is an hour north of New York City. He is an internationally recognized expert on container shipping and maritime commerce most known for his regular reports analyzing the numbers of the container conveyance systems that drive the world economy. When not engaged in serious matters related to his vocational field, as a biking enthusiast he will frequently be practicing his favorite transport avocation pedaling among the bucolic hills of northern Westchester County.
