I hope that everyone out there enjoyed the bye week and a stress-free weekend of college football. For those that live in Michigan, I hope that everyone enjoyed the beautiful weather. I was a perfect time to go visit one of the many venues that make up the Southeast Michigan Cider Mill Industrial Complex to perhaps enjoy some doughnuts and everyone's favorite lightly fermented fall beverage.
Michigan State and Michigan were both a bye this week, and so both fan basis had a chance to a full-week head start on trash talking. From what I can tell, there was not a whole lot brewing yet on that front. I am expecting a few things with a little higher proof by kick-off on Saturday. I believe in you, Spartan Nation.
For now, let's jump right into the analysis of games this past weekend.
Week Eight Results
Figure 1 below shows the results of all 52 games involving two FBS teams in Week Seven.
 |
| Figure 1: Results of Week Eight showing the actual point differentials relative to the opening spread. |
Texas Tech, Penn State, Fresno State, Ohio State, Army, and Houston were the teams that significantly beat expectations this week. No team underachieved by more than 14 points relative to the opening spread, yet still won. The team closest to this dubious distinction was Kent.
Table 1 below summarizes that games that ended in an upset in Week Eight in comparison to the picks made by the computers last week.
 |
Table 1: Upsets in Week Eight based on the opening Vegas line compared to the upset projections from last week.
|
I count a total of 18 upsets relative to the opening point spreads, which was slightly higher than predicted. Only two of the upsets came in games where the spread was over 10 points, and only five of the upsets involved Power Five teams.
Based on the spread, the biggest upset of the week was Florida International's win over Charlotte (-14.5). The upsets with the biggest impact were Duke over Miami, Oklahoma State over Texas, and LSU over Ole Miss.
My computer went 3-5 (38 percent) in upset picks this week which helped to pull its year-to-date performance to over 30 percent for the year. The FPI went 2-1 (66 percent) to bring its year-to-date average up to 39 percent.
Table 2 below summarizes the results of the recommended picks against the spread.
 |
| Table 2: Results of the highlighted picks versus the spread in Week Eight. |
It was slim pickings on this front this week with only three picks on the board and only one of those three picks wound up correct. By putting up a donut this week, my algorithm's year-to-date performance just 12-28 (30 percent). The FPI went .500 this week to bring its tally to 24-32 (43 percent) year-to-date.
Table 3 below summarized the results of the curated point total (over / under) picks for the week.
 |
| Table 3: Results of the recommended point total bets ("over / under" bets) in Week Eight. |
The full set of my recommended picks went just 4-7 (36 percent) to bring the year-to-date performance to 74-63 (54 percent). My "lock" picks did a bit better at 4-3 (57 percent) for the week, which is good for 31-19 (62 percent) for the year.
Updated Big Ten Odds
A second glance at Table 1 reveals that none of the Big Ten games this week ended in an upset, including a home win by Wisconsin over Purdue. A second glance at Figure 1 also shows that both Penn State and Ohio State beat the spread by a convincing margin (even if Ohio State did this in surprisingly impressive manner).
The updated odds for the Big Ten race are shown below in Table 4.
 |
| Table 4: Updated Big Ten rankings, expected wins, strengths of schedule and season odds following Week Eight |
Ohio State's impressive (on paper) performance against Iowa has vaulted the Buckeyes once again into first place in the Big Ten East. My simulations gives Ohio State a 55 percent chance to advance to Indianapolis for the Big Ten Championship. Michigan is still very much in the race with odds of 35 percent while Penn State still has a prayer at 10 percent.
In the Big Ten West, Purdue's loss in Madison nudged up the odds for the idle Ilini by a few percentage point up to 87 percent. The Boilermakers (12 percent odds to win the West) are the only other team with odds over one percent to claim the Big Ten West crown.
Michigan State's remaining schedule and expected wins
Table 5 below shows the updated win distribution matrix for the Big Ten following the action of Week Eight.
 |
| Table 5: Updated Big Ten win distribution matrix after Week Eight. |
As expected, Michigan State's expected win total was unchanged over the bye week and remains at 4.71 with just a 16 percent chance to get to 6-6. As for the rest of the Big Ten's bowl prospects, Ohio State, Michigan, Penn State, Maryland, and Illinois have already earned the necessary six wins.
Minnesota and Purdue are virtual shoe-ins to join them while Iowa and Wisconsin's odds are between 70 and 80 percent. In contrast, Rutgers, Indiana, Nebraska, and Northwestern all have longer odds to make a bowl than do the Spartans.
Figure 2 below gives the detailed odds and projected point spreads for all of Michigan State's remaining games. This week I have also included the values calculated prior to the bye week (following Week Seven) as a comparison.
 |
| Figure 2: Updated odds and spread projections for Michigan State’s remaining games, following Week Four and based on a 60,000 cycle Monte Carlo simulation. |
Basically, very little has changed in the last seven days. My projected spreads for the Michigan game and Illinois game improved slightly. But, the odds for the games against Rutgers and Penn State decreased slightly for the Spartans. The biggest change is for the projection of the Indiana game. In this case the spread moved from -3.5 to -6.5.
Note also that my simulation's projection of the point spread for the Michigan game (+21.5) is dead-on of the current line listed on DraftKings. How do you like them apples?
At the end of the day, the Spartans are in the same position as a week ago. Michigan State has two very winnable games at home in the form of Rutgers and Indiana and three very challenging road games. In order to make a bowl game, the Spartans need to find three more wins.
If we assume for a moment that Michigan State beats both Rutgers and Indiana, then only one more road win is needed. My current odds for the Spartans to win each individual game is between seven and 12 percent. But, the way that the math works out, the odds of Michigan State winning at least one of those three games is 25 percent.
National Overview
As we do every week, let review the action that you might have missed while mainlining some cider and donuts this weekend.
Georgia and Kentucky had the week off and Tennessee easily defeated some dude named Martin this week, so the race is essentially unchanged in the SEC East. The defending champion Bulldogs (82 percent) still hold a commanding lead over Tennessee (17 percent) based on the results of my simulation.
In the SEC West, however, there were a couple of upsets they may add a little pumpkin spice to the race. First Texas A&M dropped a game at South Carolina and as a result the Aggies are not only out of the division race, they are below .500 for the year. More importantly, LSU ran away from Ole Miss late. As a result, Alabama (who easily took care of Mississippi State), LSU, and Ole Miss now all have a single loss in conference play. The Crimson Tide still own the best odds to win the West (64 percent), but Alabama still has road games remaining at LSU (19 percent) and Ole Miss (17 percent).
In Big 12 action, Oklahoma State rallied late to top upset Texas while TCU protected their home turf against Kansas State. As a result, the race in the Big 12 is still wide open. TCU is now alone in first place in the standing and my simulation agrees that the Horned Frogs have the best odds (52 percent) to advance to the Big 12 Championship Game. My math has Texas (49 percent) in second place, despite their 3-2 record in conference play. Kansas State (45 percent) and Oklahoma State (28 percent) are both a game out of first place and still very much in the race.
In the ACC, Clemson essentially has both the Atlantic Division and the entire conference locked up following their comeback win over Syracuse. My computer gives the Tigers a 99 percent chance to win the division and a 75 percent chance to win the conference. In fact, my computer gives Clemson a 46 percent chance to run the table at 12-0. Idle North Carolina retains the best odds to win the Coastal Division (75 percent) and the opportunity to lose to face Clemson in the ACC Championship game.
Out west in Pac-12 action, USC (77 percent) and Utah (55 percent) have the best odds to advance to the Pac-12 Championship Game, but both teams had a bye in Week Eight. Oregon (41 percent) beat UCLA (17 percent) in the only significant conference action. With the win by the Ducks, the Pac-12 is trending towards a three-team race for the top two spots.
Finally, in Group of Five action approximately 24 hours after I predicted that South Alabama would claim a spot in the New Year's Six, the Jaguar went out and lost to Troy. In additional Central Florida took a bad upset loss at East Carolina. As a result, the pool of candidates for the NY6 spot continues to get more and more shallow.
Right now 7-1 Tulane (34 percent) and 6-1 Cincinnati (13 percent) sit at the top of my Group of Five Leaderboard, but neither team has an impressive resume and my calculations suggest that neither team is likely to run the table. There is a solid chance that the Group of Five representative in the New Year's Six will have at least two losses and may not even have a win over a Power Five team.
Against all odds I have reached the end for today. Until next time, have fun with rivalry week. Go State, beat the Skunk Bears!
Comments
Post a Comment