clock menu more-arrow no yes

Filed under:

Against All Odds: Week Eight Recap (Job Security)

New, 1 comment

While it was a no doubt a tough loss on Saturday, it is far too early to burn it all down. That said, is anyone’s job safe? Also, what kind of job did I do with my betting advice and what are MSU’s current odds?

NCAA Football: Rutgers at Michigan State Raj Mehta-USA TODAY Sports

On October 14th, Mel Tucker tweeted out a video with the message “Ball security is job security.”

However, following MSU’s seven-turn over, excruciating 38-to-27 loss to Rutgers, I think that it is fair to ask the question, “exactly whose job is still safe?”

Let’s start with the wide received position. Jaden Reed clearly looks like MSU’s best receiver, yet he fumbled twice. Jalen Nailor is the clear No. 2... yet he muffed a punt. Running backs? True freshman Jordan Simmons had the best day overall (except for a fumble). Quarterback? Honestly, if you would have told me that Rocky Lombardi was going to throw for over 300 yards and three touchdowns with close to 70 percent accuracy, I would have been pretty happy... as long as you left out the fumble and two interceptions.

In fairness to Rocky, though, it is not clear that any of his three turnovers were actually his fault. The first interception may have been on Nailor, the second interception was in desperation, and the fumble I would attribute to the offensive line. As for the o-line, I was always taught to saw that if you don’t have anything nice to say, then it is best not to say anything at all.... so let’s talk about the defense.

The defense, while not outstanding, was OK. They struggled on a few drives, but in general played well when they weren’t put in terrible positions by a fumble, interception, or a turnover on downs. Tackling and pursuit angles could certainly been better, but I am willing to give them a pass, for now. If nothing else, Rutgers actually looked competent on offense, which no one really expected. At the end of the day, however, the defense did not get the job done either.

How about the coaching staff? Honestly, with so many turnovers, it is hard to say. That said, I was certainly confused by the decision to kick the extra point (and not go for two) when MSU was down 15 points and later 18 points, yet when MSU was down eight points at the 20-yard line, the staff opted for a draw play on fourth and three. I was not inspired.

Finally, can I talk about everyone’s favorite former Detroit Lions General Manager, Matt Millen? If my scoring from home is correct, Matt clearly displayed this weekend that he: 1) Cannot calculate the distance of a field goal based on the line of scrimmage. 2) Believes that an eleven-point deficit in football is a three possession game and 3) Has no idea when the clock stops, starts, or how long plays take. Let’s just say the job security may be an issue in the booth as well.

That all said, I am still the Chief Optimism Officer here and therefore it’s my job to offer the following positive spin. While losing to Rutgers is very, very bad, it could have been slightly worse. We could have been there in person, listening to the fans boo instead of sitting at home and perhaps enjoying a 16-oz double IPA before the end of the first quarter. Thank the Lord for small mercies. All levity aside, there are several things to point to which give me some hope.

First, the new coaching staff took over this team just after signing day, didn’t have a spring practice, and had to install a new offense and defense during a global pandemic. Second, Rutgers did not look like the usual raging dumpster fire that we are used to seeing. Under Coach Schiano, they might wind up being semi-competitive. Third, seven turnovers are enough to get you beat by pretty much any Power Five team (and most Group of Five teams) and turnovers tend to be a fluky stat. It is extremely unlikely that a repeat of that performance will be seen any time soon.

Whether it is justified or not, I still have faith in this new coaching staff. I like Coach Tucker’s attitude and energy, and I like his resume and pedigree. Judging him and his staff on the job that they are doing after one game, or even one season (especially this one) is foolish. I still like the mix of talent on this team. I see potential, and still believe that they will provide us with some good memories this fall. I am not secure in this belief, but I have faith.

Even if they don’t, I am excited to see what Coach Tucker can do after a few years in East Lansing. While his debut did not go the way that any Spartan fan would have wanted, he is still our coach and deserves the support of Spartan Nation. This is the hand that we have been dealt in 2020 and just like real life, it is our job to try to make the best of a bad situation and try to secure the best possible future.

Besides, it’s 2020 and the whole season has a feel of an exhibition anyway. The whole operation might not even make it past mid November. So, let’s just all raise and glass and enjoy for a moment the privilege that it was just to watch the Green and White take the field on a beautiful fall afternoon. Make mine a double.

Week Eight Results and Betting Review

In the Bad Betting Advice preview of this week’s action, I provided a summary of my algorithm’s and the EPSN FPI’s computer picks for the week. In the Against All Odds series, I will provide a summary of the week’s action and see how the picks for the week faired. Figure 1 shows the final margin of victory for each game relative to the opening Vegas Line.

Figure 1: Summary of the action in Week 8, based on the initial Vegas Line

Similar to the figure in the Bad Betting Advice column, this figure shows the results of all games involving two FBS teams this week, relative to the opening Vegas line. At a glance, this figure shows the teams that beat the spread (above the solid diagonal line), the teams that did not cover (below the solid diagonal line) and the favorites who lost straight up (below the red line on the x-axis).

The dotted diagonal lines represent score differentials that are one standard deviation from the average (which is the spread itself). As I explained in detail previously, the standard deviation in college football is almost exactly 14 points. Therefore, we can expect that roughly one third of all games in a given week will have a margin of victory greater or less than 14 points from the spread.

Week Eight had 44 total games involving two FBS teams, so it is expected that about 15 of those games will fall outside of the dotted lines. In Figure 1, I count nine games above the dotted diagonal line and seven games below it, so the math checks out, as it does most weeks.

As for the teams that got the job done this week, those include Notre Dame, Northwestern, Kansas State, Louisville, LSU, Wisconsin, Tulsa, Michigan, and Boston College. As for the underachievers, that would include Clemson, and five of the teams that were upset, including a certain favorite team of mine from East Lansing.

Let’s just take a second to ponder the paragraph above. Based on the normal distribution, each team has only about a one-in-six chance of either beating the spread by 14 points or more (as Michigan did) or not covering by 14 points or more (as MSU did).

So, the odds of both of those events happening on the same weekend is roughly 1-in-36 (or 2.5 percent if I use the exact value). So, if it feels like the results of the games involving the teams from Michigan were particularly unlucky for Spartans fans who also enjoy a little schadenfreude, it appears that you are correct. It was the virtual equivalent of rolling snake eyes.

Speaking of things that are upsetting, Table 1 below summarizes the upsets in Week Eight, including a comparison to my picks from last week.

Table 1: Upset results for Week Eight

As the table shows, there were a total of 14 upsets this weekend, which is slightly higher than the value of roughly 11 that simulation predicted. All but two of the upsets were games with a spread of a touchdown or less. Relative to the Vegas line, Rutgers’ upset of MSU was the biggest upset of the week. Neat.

As for prognostication, my algorithm did an OK job, getting two of the five picks correct (40 percent). The FPI did a much better job, going a perfect three-for-three. As for the recommended bets, those results of summarized in Table 2 below.

Table 2: Summary of the betting recommendation results from Week Eight

In this case, my algorithm did not get the job done, going only one-for-four (25 percent). The FPI did a much better job, hitting on six-of-nine, which is good for 67 percent. If I take the combined predictions, the record for the week is 7-6 (54 percent) which is equivalent to the historical performance of this system.

In total, my algorithm went 20-24 against the spread (ATS), or 45.5 percent. The FPI fared better, going 24-20 (54.5 percent). While I did not officially post the picks from Week Seven, I did run the numbers as a trial, and for the record, it did a better job last week. Year to date (starting last week), my algorithm is 35-36 overall ATS (49 percent), while the FPI is 36-35 (51 percent) based on my accounting.

Updated Expected Wins and Season Odds

Over the summer and just prior to the start of the Big Ten season, I presented the results of several full season simulations that provided the odds for each Power Five team to win their respective division and conference. Those simulations were based solely on the preseason rankings and the known historical uncertainty of those rankings.

Now that the season is underway for the Big Ten, there is more data and thus an opportunity to update those odds based on a revised simulation. The twist and improvement that I added to my simulation this year is the uncertainty in the quality of each team on a week-by-week basis. This uncertainty decreases the more games a team has played, based on historical data.

So, for each Big Ten team there are now two “observations” about how good each team really is: the preseason ranking and the performance this past weekend. We likely know a little more about each team now compared to a week ago, but two observations are barely enough for a correlation. So, the uncertainty in the Big Ten data is still quite large.

In addition, the data is still not “connected” for the Big Ten teams. For example, Indiana upset Penn State this weekend in dramatic fashion. (I am sure that Penn State fans feel that they got jobbed.) Is this because Indiana is better then expected? Is Penn State worse than expected? Did Indiana just get lucky or vice versa?

Based on only one game, it is impossible to tell. The math dictates that we assume an average. In other words, Indiana was a little under-rated, Penn State was over-rated, and luck was neutral. In the coming weeks, these individual factors will begin to shake out in more detail.

That said, Table 3 below gives the updated expected win totals, power rankings, strength of schedules, division, and conference title odds for the Big Ten, based on all the data that we have to date... which is not much.

Table 3: Updated power ranking, expected win totals, strengths of schedule, and odds for the Big Ten following Week eight

Note that the strength of schedule calculation is a bit complicated. I have reviewed the method previously, but I made a small modification based on the crazy season that is 2020. In the first column, I list the strength of schedule for a team’s full schedule. In the second column, it is the strength of schedule based on games already completed, year-to-date.

The unit in each case is the expected number of wins by an average Power Five team playing the same schedule. To adjust for the fact that not all teams are play 12 games this year, I am scaling each number as if each team were playing (or had already played) 12 games.

From the MSU point of view, the update is grim. In my system, MSU’s power ranking dropped almost 30 slots to No. 83 overall and the expected win total was more than cut in half. My math now suggests that we can only expect MSU to win one game total in the current eight-game schedule.

Not only did MSU’s ranking go down, but the strength of schedule calculation suggests that MSU’s schedule actually got a bit tougher on balance over the weekend. This is mostly likely due to the fact that both Michigan and Northwestern over-achieved significantly. That said, I still calculate an extremely unlikely 0.05 percent chance (one in about 2,100) that MSU wins the Big Ten East. Fingers remain securely crossed.

As for the rest of the Big Ten, in the East, Ohio State had a great opening weekend, as did Michigan. The odds for both teams to win the Division went up by 12-14 percentage points, with the Buckeyes now having a 58 percent chance to repeat as East Champs. Indiana’s odds got a small boost, up to six percent, while Penn State’s odds dropped by 26 percentage points to 11 percent.

In the West, it was really good weekend for the Badgers. Not only did Wisconsin look strong in their opener against Illinois, but the presumed challengers, Iowa and Minnesota, both lost. As a result the odds for Wisconsin to return to Indy are now over 50 percent as well. The most likely challenger at this point appears to be Northwestern (28 percent) based on the strong performance in securing a win over Maryland. Northwestern’s strength of schedule now appears to be the second easiest in the conference behind Ohio State.

Table 4: Big Ten win distribution matrix following the action in Week Eight

Table 4 above shows the updated win distribution matrix for the Big Ten following this weekend’s action. Once again, the data for MSU is grim. My simulation now gives MSU a 63 percent chance of winning no more than one game. The odds of getting to 4-4, somehow, are now at four percent. I could go on, but I don’t want to depress anyone any further.

What’s Next for MSU

The expected win distribution matrix for MSU can be deconstructed by looking at the odds for MSU to win each of the remaining seven scheduled games. The simulation provides these odds, which can then by used to back out a projected spread. Here are the updated projections:

  • MSU at Michigan (-27.5, 2.7 percent). Note that the actual line opened slightly lower than that. Note also that the odds of an MSU win are virtually identical to the low probability of the combined outlier results from the Spartans and Wolverines this past weekend.
  • MSU at Iowa (-14, 16 percent). This actually seems more promising after Iowa’s loss this weekend, except that these are actually longer odd than I projected a few weeks back.
  • Indiana at MSU (+11.5, 21 percent). This games used to be a toss-up. It no longer looks like a toss-up.
  • MSU at Maryland (+5.5, 66 percent). I have good news: this game actually looks more winnable now.
  • Northwestern at MSU (+14, 16 percent). I am not sure that I believe in Northwestern so much after one game against Maryland, but this is what the math says.
  • Ohio State at MSU (+24.5, 4 percent). Yep.
  • MSU at Penn State (-24, 5 percent). Sure.

So, it’s not great. MSU projects to only be favored in only one of the remaining games (at Maryland). Keep in mind as well that these projections are based on a fairly small set of data. Next week, we will know a little bit more about the entire Big Ten. The situation could look a bit better. It could also look worse.

As a small tidbit of probability-based optimism, I offer the following calculation. Based on the odds above, there is an 11 percent chance that MSU secures at least one victory against one of Michigan, Penn State, or Ohio State. Furthermore, there is a 44 percent chance that MSU gets a win over one of Iowa, Indiana, and Northwestern.

That is all for this week, Stay tuned on Wednesday for my next round of Bad Batting Advice for Week Nine. In the mean time, like it or not, it’s Michigan Week. So, get out there and troll like you mean it. Until next time, Go State; Beat the skunk bears!