Week One of the 2022 College Football season and the season of the Michigan State Spartans is now complete. So, how are we all feeling?
On one hand, the Spartans faced some adversity and overcame it to put away a scrappy in-state opponent. Michigan State even covered the spread (although it was -22.5 in some places at kick-off). A win is a win, especially in the first game against a highly motivated opponent.
On the other hand, the Spartans looked sloppy, it was still a one-possession game with under seven minutes to play, and the injuries (some quite serious) are already piling up. What are we to make of all of this?
At this point, I simply choose to reserve judgment for now. Is Michigan State not quite as good as fans had hoped in the off-season? Is Western Michigan a little better than expected? It is hard to say. Would the game have played out more positively for Spartan fans had Payton Thorne not thrown a bad interception in the final minutes of the half? If so, would Thorne have instead attempted a similarly bad throw later in the year (for example) at Washington or at Michigan?
Who knows?
The Spartans have a reservation next weekend to face the Akron Zips. Right now, we only know slightly more about Michigan State than we did a week ago. Next week will be a chance to gather a bit more data about the Spartans and the rest of the college football landscape. As for now, a more reserved approach seemed warranted. Let's all keep our gavels and robes in storage for the time being.
Week One Results
While the jury is still out on the Spartans, we do have the right data from Week One to judge the performance of the computers' predictions.
In my Bad Betting Advice preview of Week One, I presented the picks from my football algorithm as well as the picks from EPSN's FPI metric as compared to the opening spread as found on DraftKings. Now, it is time to adjudicate the results of the week, first by plotting the actual final point differential of each game as a function of the opening Vegas line.
 |
| Figure 1: Recap of the action from Week One comparing the actual game results to the Vegas line (only including games where both teams are in the FBS) |
For the games whose data point falls above the solid diagonal line, the favored team covered the spread (and vice versa). If the data point falls outside of the set of two dotted lines, the final margin was more than one standard deviation (i.e. 14 points) from the opening line. If the data points fall under the vertical red line, the favored team lost straight up.
A glance at Figure 1 suggests that it was a good week to be the favored team. A total of 10 teams including Georgia, SMU, James Madison, Michigan, Tennessee, TCU, and BYU all covered the spread by more than two touchdowns, In addition, the eight total upsets on the week was slightly fewer than the 10 predicted by the weekly simulation.
Table 1 below summarizes the upsets that occurred in Week One.
 |
| Table 1: Summary of the upset from Week One involving two FBS teams |
Based on the opening spread, Old Dominion's (+7.5) win over Virginia Tech was the biggest upset of the weekend. But, an upset of this magnitude is the college football equivalent of jaywalking. In 2021 there were a total of 59 upsets with an opening point spread of +7.5 and above.
The computers both got off to a tough start in 2022 when it comes to upset picks. Of the eight total picks on the board, only one ended up correct: The FPI's pick of Florida to beat Utah.
Table 2 below summarizes the result of the computers' recommended picks against the spread for Week One.
 |
| Table 2: Results of the suggested bets from Week One |
If the computers' performance in upset picks was judged to be a misdemeanor, the performance against the spread was downright felonious. My computer actually was OK in Week One, going one-for-two. The FPI, however, was a disastrous one-for-ten.
As I mentioned in this week's preview, the data from ESPN's computer was less reliable last year than expected, and it is not off to a good start in 2022. If this behavior keeps up, I may need to put the FPI in prognosticator jail for a few weeks to sober up.
Overall, the data for all 46 games is no better than the data for the suggested bets. My algorithm went 15-31 (33 percent) on the week while the FPI was 17-29 (37 percent). This week's Betting Advice was truly Bad.
Updated Big Ten Odds and Expected Wins
Just as it is important to reserve judgment on the Spartans, the same can be said for the rest of the Big Ten and on the college football landscape in general. That said, we now have a full week of actual data and by processing it, it is possible to gain some potential hints as to the direction of the Big Ten season.
Weeks Zero and One were fairly good ones for the Big Ten conference. To date, Big Ten teams are 10-0 in non-conference play. Rutgers was able to earn a road upset at Boston College, and Ohio State protected home turf against No. 5 Notre Dame. It's early but it was also a good start for the Big Ten East, as Penn State avoided to road upset at Purdue and Indiana survived the visit from Illinois this week.
The Iowa Hawkeyes were even smart enough to score two safeties against FCS opponent South Dakota State to pair with a single field goal to disguise that fact that their offense barely gained over 150 yards and never found the end zone in the 7-3 win. I would assume that Iowa will petition the courts and request that a judge seal all records pertaining to the full box score and play-by-play. Let us never speak of this game again.
Tables 3 and 4 below provide an update to the Big Ten odds and win distribution matrix following the results of Week One. As was done in the summer, the data is the output of a Monte Carlo simulation of the full season based on projected point spreads for all remaining games.
 |
| Table 3: Updated Big Ten odds after Week One. The change in rankings and odds from the preseason are shown in parenthesis. The strength of schedule data are noted with the updated national ranking in parenthesis. |
 |
| Table 4: Updated Big Ten win distribution table after the results of Week One |
For the Michigan State Spartans, the verdict after one week of play is that the road looks a little tougher than it did a week ago. The expected win total has dropped by almost to full game down to 7.45 wins and the odds to win both the East Division (seven percent) and the Big Ten (four percent) have both roughly been cut in half, despite the win over teh Broncos on Friday night.
The reason for this small downgrade is two-fold. First, the Spartans' performance against Western Michigan was slightly below expectation, at least in the eye of my computer. Michigan's State's power ranking has dipped down to No. 19.
The second factor is likely the more important one. Almost every single opponent on Michigan State's schedule had a neutral to positive outing in Week One. As a result, the Spartans now own the most difficult schedule in the Big Ten (up from No. 5 in the preseason) and the ninth most difficult schedule in the nation (up from No. 25).
As for the rest of the Big Ten, the race in the East Division still looks like a battle between Ohio State (41 percent) and Michigan (38 percent). The Buckeyes and Wolverines statistically gained some separation from Penn State (10 percent) and Michigan State (seven percent) in Week One.
In the Big Ten West, Wisconsin (40 percent) still has the best odds, but a new dark horse contender has arisen in the form of the Minnesota Golden Gophers (31 percent). Iowa (13 percent) and Purdue (seven percent) are looking more like longshots. The news is perhaps even worse for Iowa, as my algorithm does not even take into account the Hawkeye's poor Week One performance against South Dakota State.
Michigan States's remaining schedule
In my math-based preview of Michigan State, I presented a game-by-game projection of the point spreads and the odds for the Spartans to win each game. Now that Week One is in the books, Figure 2 gives the updated projections for the remainder of Michigan State's schedule.
 |
| Figure 2: Updated odds and spread projections for Michigan State's remaining games, following Week One and based on a 30,000 cycle Monte Carlo simulation |
If we compare this figure to the one generated a month ago, the odds for the Spartans in literally all 11 remaining games have decreased. Based on Table 4 above, the odds are now 50-50 that the Spartans will win eight or more games.
Next weekend's game with Akron and the late season home games against Rutgers and Indiana still all look like likely wins. But, the road games at Washington, Maryland, Illinois, and Penn State as well as the home games against Wisconsin and Minnesota now all look essentially like toss-ups. The Spartans continue to project to be significant underdogs against Ohio State and at Michigan.
It is obviously still extremely early, but if we take these numbers to be accurate, four wins total looks like the floor. If Michigan State can win three of the six toss-up games, that would bring the win total up to seven. The Spartans would need to win more than half of the toss-up games in order to get to eight wins or more. The #math also currently gives the Spartans a 50-50 chance to win at least one of the two games against Ohio State and Michigan.
National Overview
Finally, let's take a quick spin around the country and pass some judgment on the other action in Week One. For reference, my updated odds for each conference can be viewed by clicking on the hyperlink for each conference.
The SEC was found not guilty of stumbling during the opening week as the only loss the conference suffered was LSU's failed comeback attempt versus Florida State. Florida managed the mild upset over visiting Utah, and Ole Miss was the only team that clearly failed to cover (against Troy State). The assumed division favorites (Georgia and Alabama) won their games by a combined score of 104-to-3 and both teams have a more commanding lead based on the updated division odds.
Things were fairly status quo in the Big 12, but this is due in large part to the fact that a full half of the conference faced FCS opponents in Week One. Every Big 12 team won with the exception of West Virginia (who lost to Pittsburgh). The only notable things to put on the record here are that Oklahoma and especially Oklahoma State both failed to cover the spread against UTEP and Central Michigan while TCU won big at Colorado. As a result, my calculations now project Texas (48 percent) and Baylor (31 percent) to have the best odds to reach the Big 12 Championship Game.
It was also a calm week in the ACC this week. Boston College and Virginia Tech both suffered upset losses to Rutgers and Old Dominion, while Florida State upset LSU in dramatic fashion. Pittsburgh and North Carolina narrowly avoided upsets at the hand of West Virginia and Appalachian State. Over all, Clemson () and Miami (56 percent) are both still favored to win their division. I will also note that North Carolina State struggled mightily to beat East Carolina which prompted a 13-percentage point drop in their odds to win the Atlantic Division. My selection of the Wolf Pack to earn a New Year's Six Bowl may have been an error in judgment.
The Pac-12 was found guilty of having a tough start to the season. Both Oregon and Utah (the two preseason conference favorites) had chances to make a Week One splash against SEC opponents. But Oregon got blow out by Georgia and Utah lost at Florida. As a result, the odds of a Pac-12 team making the Playoffs dropped significantly (from 52 percent to just 32 percent). Furthermore, the teams with the best odds to make the Pac-12 Championship game are now Utah (58 percent), USC (30 percent), Oregon State (30 percent), and UCLA (22 percent).
Finally, it was a tough opening weekend for several Group of Five teams with New Year's Six Bowl aspirations. Cincinnati lost to Arkansas, and Houston needed overtime to defeat Texas San Antonio. Boise State got blown out at Oregon State, San Diego State was upset by Arizona, and Appalachian State couldn't get past North Carolina.
Perhaps the biggest Group Five result was Southern Methodist's blow out win over North Texas. As pedestrian as this seems, it did rocket SMU to the top of my Group of Five leaderboard. I suppose that we will need to reserve judgment until next week to see if this new ranking holds.
Against all odds, I have reached that end for today. Until next time, enjoy, and Go State; beat the Zips!
See also: https://www.sbnation.com/e/23102685
Comments
Post a Comment