2023 Bowl Season Preview

As a general rule, there are two types of college football fans: those who enjoy the traditional bowl games and those who do not. I personally fall squarely into the first category.

I can certainly see the argument that most bowls are unnecessary exhibitions between mediocre teams playing in half empty stadiums in some borderline (to be generous) tourist location. This argument is not factually false.

For a counterargument I can only offer the following: bowl games are fun and at the end of the day sports are meant to be fun.

Do I want to watch the third-place ACC team take on the fourth place Big 12 team on Dec. 28 in Orlando in a bowl game named after my favorite Kellogg's toaster pastry? The answer is clearly "yes," and I will celebrate with not one, but two perfected toasted raspberry Pop-Tarts at half time. Don't judge.

Furthermore, with the expanded 12-team playoff starting next December, it is not clear what the future holds for many of the lower-tier bowl games. I believe that they will mostly continue to serve as "pointless exhibitions" for the non-conference champions that finish the season ranked roughly No. 13 and below.  But the future is murky, which makes particular bowl season even more special. It will be the last of its kind.

So as for me (and my toaster), we plan to enjoy this final ride of the traditional bowl season as we know it. While I am at it, I might as well deploy the same set of analysis that I have provided throughout the season. One last time, I present a batch of Bad Betting Advice for the 2023 college football season.

Bowl Season Picks

Figure 1 below gives an overview summary of all my computer's projections for the bowl season in comparison to the opening Vegas spread. The same data generated by ESPN's Football Power Index is shown in Figure 2.

Figure 1: Comparison of the projected final margins of victory based on my algorithm to the opening Vegas lines for bowl season.

Figure 2: Comparison of the projected final margins of victory based on ESPN's FPI to the opening Vegas lines for bowl season.

For whatever reason, there appears to be significantly more scatter in the data than usual. The computers do not have knowledge of injuries, portal entries, or coaching changes, which likely explains the appearance of the data. If true, the predictions of Vegas may be more likely to be more accurate. But I choose to press on with the same methodology that I have used throughout the season.

In any event, the result of this scatter is a higher than expected number of upsets and picks against the spread. Table 1 below summarizes the upsets picks that can be derived from the figures above.

Table 1: Summary of upset picks for bowl week.

The computers recommend three upset picks where the spreads are over a touchdown including wins by Duke (Birmingham Bowl), Tulane (Military Bowl), and USC (Holiday Bowl). Other notable upset picks include North Carolina over West Virginia (Duke's Mayo Bowl), Oklahoma over Arizona (Alamo Bowl), Maryland over Auburn (Music City Bowl), and Ohio State over Missouri (Cotton Bowl).

A simulation of bowl season using the opening Vegas lines suggest that a total of 13.4 plus-or-minus 2.8 upsets is most likely.

Table 2 give the computers' recommended picks against the opening spread based on the data in Figure 1. 

Table 2: Picks against the opening spread for bowl week. The picks are listed in order of confidence score.

My analysis of ESPN's FPI data results in a total of 11 picks against the spread this bowl season. My computer agrees with four of those picks including Ohio State, Oklahoma, Florida State (Orange Bowl), and Kansas State (Pop-Tarts Bowl) all to cover. The FPI data also predicts that Duke, USC, Syracuse (Boca Raton Bowl), and North Carolina will all cover.

Finally, Table 6 below summarizes my computer's suggested point total bets for the week.

Table 3: Recommended total points bets (over/under) for bowl week. Note the picks labeled as "locks" are at a higher level of confidence.

Despite a terrible performance this year, my computer suggests a total of 10 point-total bets, nine of which are in the category that I have ironically titled "locks." My computer likes the "over" in every game in the table except for Texas versus Washington (64.5) in the Sugar Bowl.

Big Ten Overview

The Spartan football team will be spending the holidays at home (or in the transfer portal) but nine other Big Ten teams will partake in (at least) one more game this season. 

Table 4 below summarizes the details and computer predictions in the bowl games involving Big Ten teams this season. The games are listed from the contest where the Big Ten team is the most heavily favored (Minnesota opened as a 3.5-point favorite over Bowling Green in the Quick Lane Bowl) to the game where the Big Ten team is the biggest underdog (Wisconsin opened as a 10.5-point underdog to LSU in the ReliaQuest Bowl, previously known as the Outback Bowl).

Table 4: Summary of the 2023 Big Ten bowl season action, including my algorithm's projected scores. The teams shaded in green are projected to cover by my algorithm. The teams shaded red are projected to cover by the FPI

As Table 4 shows, it could be a challenging bowl season in Big Ten land. Overall, conference teams are only favored in three of the nine games, with Penn State (which faces Ole Miss (+3.5) in the Peach Bowl) and Michigan (which will face Alabama (+2.5) in the Rose Bowl) joining the Golden Gophers as the only teams favored.

In general, both computers like those three Big Ten teams to both win and cover. My computer in particular projects a double-digit win by both Penn State and Michigan. The only exceptions is that the FPI has Alabama coving the 2.5-point spread in the Rose Bowl semifinal.

Three other Big Ten teams enter bowl season as slight underdogs. Maryland (which faces Auburn (-2.5) in the Music City Bowl), Ohio State (which plays Missouri (-2.5) in the Cotton Bowl), and Rutgers (which squares off with Miami (-2.5) in the Pinstripe Bowl) are all picked to lose by less than a field goal.

However, the computers are generally more optimistic than the folks in Vegas. Both machines pick Ohio State to win by close to or over 10 points. My computer also likes Maryland in the upset. The computes are not so confident about the Scarlet Knights. Both computers have Miami covering.

The final trio of Big Ten West teams are all picked to lose by at least a touchdown. Those teams are Northwestern (which faces Utah (-7) in the Las Vegas Bowl), Iowa (which faces Tennessee (-8.5) in the Citrus Bowl), and Wisconsin (which faces LSU (-10.5) in the ReliaQuest Bowl).

The computers both like Utah and LSU to cover in the Las Vegas and ReliaQuest Bowls but they also both predict the Iowa will cover the spread in the Citrus Bowl.

For those fans out there wondering about Michigan's odds to win a national championship. My computer has been high on the Wolverines (and the Big Ten) all year. My projected odds give Michigan a 62% chance to be the team left standing to raise the crystal pigskin on Jan. 8. Texas has the second best odds at 34% with Alabama (9%) and Washington (6%) viewed as longshot.

A peak at the Vegas futures market paints a difference picture. The official money lines do favor Michigan (+195 or 34%), but Alabama (+205 or 33%) and Texas (+260 or 28%) are not far behind. In both analyses, Washington (+700 or 12.5%) is considered to be the longshot.

National Overview

Table 5 below provides a snapshot of the participation of each conference and the distribution of match-ups for the 2023 bowl season. I have colored-coded each cell to represent match-ups between two Power Five schools (in green), two Group of Five teams (in blue), and cross-over games (in yellow).

Table 5: Matrix of Bowl game participation and conference match-ups. Cells in green represent games involving two Power Five teams. Cells in blue represent games involving two Group of Five teams and cells in yellow represent a Power Five/Group of Five cross-over.

As the table shows, 19 of the 41 bowl games involve two Power Five teams. Six of those 19 games are SEC/Big Ten matchups while three of the games are Big 12/ACC matchups and three more are Big 12/Pac 12 matchups. 

The ACC is the only Power Five conference which has a bowl game scheduled against at least one team from each of the other four Power Five leagues. The Big Ten does not face a Big 12 team in this year's bowl season while the SEC and Pac 12 also do not face each other.

A total of nine bowl games feature one Power Five team playing a team from the Group of Five. The most frequent example of this matchup is the three bowl games between the ACC and the American Athletic Conference (AAC). The SEC is the only Power Five Conference which does not have a member playing a Group of Five team.

The remaining 13 bowls feature a pair of Group of Five teams. All but two of these games involve a Sun Belt team going up against a member of Conference USA (two games), the Mountain West (three games) or the MAC (four games). Mountain West members also have bowl games scheduled against a Conference USA and a MAC opponent.

Finally, Table 6 below provides an estimate of how each conference is expected to fare in bowl season.

Table 6: Predicted conference records and expected wins based on the opening Las Vegas point spreads and the predictions of my algorithm and ESPN's FPI.

The table lists the expected win/loss record for each conference based on the straight-up prediction of the Vegas lines and the predictions of both computer systems. In addition, the number of expected wins (and expected win percentage) derived from the point spreads is shown in the third column.

Notre Dame is a solid 6.5-point favorite over Oregon State the Sun Bowl, which give the FBS independents the highest expected win percentage. Beyond that, this analysis suggests that the Big 12 (expected win percentage of 60%) and the SEC (59%) are most likely to have the best bowl seasons. 

My computer predicts that the Big 12 might even exceed this expectations with a projected 8-1 record (if Oklahoma can upset Arizona). Interestingly, both computers predict that the SEC may underachieve be a game or two compared to the Vegas predictions.

The Vegas lines suggest that the Pac 12 will most likely go .500 during bowl season. The FPI agrees with this prediction, while my computer predicts that the Pac-12 will finish at just 2-6.

The ACC (46%) and Big Ten (43%) are expected to bring up the rear in Power Five bowl performance based on the Vegas spreads. That said, the computers generally expect those leagues to overachieve slightly. The exception is that my computer is not as confident in the ACC.

In the Group of Five, the Vegas lines suggest that it will be a great bowl season for the Sun Belt (54%) but both computers seem highly skeptical. In fact, both machines will finish at 3-9 despite the fact that Sun Belt teams are favored in eight of the 12 bowl games.

In contract, my computer is expecting that the AAC and the MAC will lead to the way in the Group of Five with records of 4-2.

For reference Table 7 below summarizes details and computer predictions for the remaining bowl games not shown above in Table 4.

Table 7: Summary of the 2023 bowl season action, including my algorithm's projected scores. The teams shaded in green are projected to cover by my algorithm. The teams shaded red are projected to cover by the FPI.

That is all the advice that I have to give this bowl season. Enjoy the games.