[NOTE: I originally posted this on Sunday afternoon, but I some realized that one of my formulas had an error. I accidentally used MSU's power ranking at the end of the 2019 season instead of the preseason ranking in my calculations of point spreads. This drove up MSU's win probabilities, especially for the ESPN data set. This also explains why my calculation of expected wins based on ESPN's data was different than their calculation. Those calculations now match.
I have now updated this post with the more accurate data and updated the narrative and my analysis accordingly. My overall predictions remain essentially the same, but the outlook for MSU is a little worse than in the original version, so my picks are now a bit more optimistic. But, that's how I roll...
Anyway, I am sorry for the confusion. I HATE putting what I consider to be bad data out there. To the best of my knowledge and mathematical ability, the following analysis is now "correct".]
As the temperature rises and the number of COVID-19 cases fall in the Great Lakes States, many are starting to dare think optimistically about the next few months and the possible return of college football and some semblance of normalcy. Across the country, teams are starting to ramp up again, starting with voluntary work outs on many college campus, including East Lansing starting on Monday. Quite honestly, literally no one knows the trajectory of the next few months. Things could get significantly better, and they could get worse.
The future is unknown. But, in times like this, it is important to hope for the best, yet plan for the worst. For people in leadership positions, they should focus on the later, while keeping the former in mind. But as for optimistic sports fans like me, I am going to simply focus on hoping for the best. I am going to assume that football season will begin on time and play out as planned. My fingers are firmly crossed.
In this spirit, it is that time of year when I usually start to look at various season previews to get a feel for how the upcoming college football season will play out. My general methodology is to try to find as many different preseason rankings as possible and to use those to estimate a consensus preseason ranking for all 130 FBS teams. I can use this ranking to do a full mathematical analysis of the entire season.
In most years, a good cross section of full rankings are not available in early July. But this year these ranking may be late or hard to find. Either way, I still plan to perform a full analysis of the upcoming season in July or August. Today, I want to provide a sneak peak at what might be in store for MSU in the 2020 season.
Opponent Overview
While preseason data is a bit limited, I did find a full set of preseason rankings from both CBS and ESPN. I then used these rankings to project a preliminary power ranking for each team (using last year's final rankings as a guide) which I can then use to project point spreads and win probabilities for each of MSU's planned 12 games. This should give us a feel for what to expect. Let's start by looking at where CBS and ESPN rank MSU as well as each of MSU's 12 opponents for 2020.
 |
| Figure 1: Rankings of MSU's 2020 Opponents based on CBS and ESPN's Preseason Rankings |
The bad news is that both CBS and ESPN have six teams on MSU's schedule ranked significantly higher than the Spartans: Ohio State, Penn State, Michigan, Iowa, Minnesota, and Indiana. CBS and ESPN are at near consensus in their rankings for the first four of those teams, while they differ a bit more on Indiana and Minnesota. To make matters even worse, MSU plays all six of those teams in a row in the middle of the schedule, and the first five of which occur before MSU's first and only bye. The schedule makers certainly did not do MSU any favors here.
We can also see that according to CBS, MSU is better than six of its 2020 opponents: BYU, Miami, Toledo, Northwestern, Maryland, and Rutgers. However, ESPN has both Miami and Northwestern ranked higher than MSU. Either way, three of MSU's twelve opponents fall generally in the middle of Figure 1: BYU, Miami, and Northwestern. At this point I think that we can consider all three of those games as toss ups.
Of the three "toss-up" games, the opener vs. Northwestern appears to be the biggest question mark. CBS is not high on the Wildcats at all, ranking them No. 111 of 130 total FBS teams. In contract, ESPN has Northwestern ranked No. 25. I am not sure who to believe here, but based on what we saw last year on the field, CBS's prediction seems much more likely.
Finally, three of MSU's games appear to be more likely to wind up in the win column: Maryland, Rutgers, and Toledo. That said, ESPN has a much smaller gap between MSU and Maryland than CBS does. Considering that this game is also on the road should give any MSU fan some pause.
Projected Spreads and Win Probabilities
More quantitatively, we can look at MSU's schedule in detail, based on both CBS and ESPN's predicted rankings. Tables 1 and 2 below show MSU's schedule, including the rank of each team, the projected point spreads, and the associated win probabilities. At the bottom of each table is the expected total number of wins, which is equal to the sum of the win probabilities for all 12 games. I have color-coded each game based on the three categories identified in Figure 1 (likely win, likely loses, and toss-up).
 |
| Table 1: Summary of MSU's 2020 Schedule, based on CBS's preseason rankings. |
 |
| Table 2: Summary of MSU's 2020 Schedule, based on ESPN's preseason rankings. |
As noted above, there is a large difference between the predictions of each outlet on a game-by-game basis, as well as in the total expected win count. CBS's rankings suggest MSU will win between five and six games, while ESPN's ranking translate to MSU only winning four.
I think that it is also instructive to look at the three different categories of games to get a feel for what to expect. First, for the "likely win" games (in green), my analysis suggests that MSU has over a 70 percent chance to win each of these games (with the exception of the game at Maryland using ESPN's rankings, where MSU is projected to be a slight underdog). Taken together, MSU's expected number of wins for these three games is between 2.1 wins (ESPN) and 2.6 wins (CBS). Despite my categorization of these games as "likely wins" MSU might be slightly fortunate to win all three.
As for the "toss-up" games (in yellow), the analysis using ESPN's data suggest MSU's odds to win these games varies from 26 to 47 percent. CBS's rankings suggest MSU's odds to beat Northwestern and Miami are actually much higher (94 and 73 percent, respectively). Taken together, the expected number of wins are 1.2 (ESPN) and 2.1 (CBS). If you average the two predictions, MSU is expected to win about one-and-ahalf out of the three games, which is why I have them categorized as "toss-ups."
The fact that the three toss-up games occurs in the first month of the season puts additional pressure on MSU to start strong in the Mel Tucker era. In order to stay on track to at least make a bowl game, MSU will be well served by winning at least three of their first four game. If MSU is slow out of the gate, MSU could wind up in a very deep hole very quickly, because the meat of the Big Ten schedule starting in October looks very challenging.
That challenging run starts with a road trip to Iowa City and ends in mid-November in Happy Valley. According to both services, MSU will face six consecutive Top 35 teams in that stretch. Neat. MSU is projected to be an underdog in all six games, but the math suggests that more likely than not, MSU will likely win at least one game in this stretch. The CBS rankings result in an expected win total of 0.77 in this stretch, while the ESPN rankings give an expected win total of 0.71.
With the CBS data, a win at Indiana is most likely, while the ESPN data suggests the Minnesota game is the most winnable. I would also like to note that the spread for the Michigan game projects to be between eleven and thirteen points (18 to 22 percent odds to win). This makes a satisfying home upset of the Wolverines in Mel Tucker's first season one of the more plausible possible scenarios.
Season Overview
If all of this data is taken together, it is possible to calculate the odds that MSU wins anywhere from zero to all twelve games based on the preseason rankings of either CBS or ESPN. Figure 2 below gives the win probability distribution using both sets of data.
 |
| Figure 2: Win probability distribution for the 2020 MSU football season based on the preseason rankings by CBS and ESPN |
As the expected win totals in Table 1 and 2 show, MSU will most likely between four and six games total, depending on which service is closer to the actual strength of each team. The CBS win distribution is shifted up compared to the ESPN distribution by about a game and a half. In total, the CBS rankings suggest MSU's odds to get to six wins and make a Bowl Game are 49%, while ESPN's rankings peg those odds at only 13%.
So, with all of this taken together, here is how I view the season. As stated above, MSU's first four games are all winnable, but MSU could easily go 1-3 in September if things don't go well. In order to stay on schedule to make a Bowl Game, I think MSU needs to go at least 3-1.
The opening game vs. Northwestern is just an enigma right now. It will be the first time anyone has seen a Mel Tucker coached MSU team and he will be facing a Northwestern team that is either pretty good (if you ask ESPN's computer) or really bad (if you ask CBS). This game could be a blowout either way, and it wouldn't tell us anything.
MSU then travels to BYU, where the Spartans will likely be a slight underdog. When was the last time a Big Ten team went out West in the regular season and won? (I am actually too lazy to research this one right now, but it's been a while...) Let's just say I have a bad feeling about this one.
Next up is Toledo, who MSU should beat, followed by Miami, which also looks to be a near toss-up. Can MSU win both while still breaking in a new staff and system following the most bizarre summer on record? I would like to say yes, but my gut says a 2-2 record in September is a very real possibility. If MSU can go 3-1 or even 4-0 to start the season, I think MSU fans should feel pretty good.
Then, things could get interesting. As the calendar turns to October, hopefully the new staff and the old players will find some sort of rhythm, and they will need it for what is to come. The next three games are at Iowa, vs. Ohio State and vs. Michigan. The first two look really tough, but in the Michigan game, nearly anything can happen. Can MSU steal a win here? Maybe.
Following the battle with the Wolverines, MSU travels to Indiana and then hosts Minnesota. Both teams are supposed to be good, but maybe not great. Can MSU steal a win here as well? Maybe, and that would be great. Two wins would frankly be pretty awesome. Then again, will the team be mentally and physically beat down after facing Iowa, OSU, and Michigan back-to-back-to-back? That might also be true. Again, the schedule makers were not kind here.
MSU finally gets their bye in early November and then makes the trip to Happy Valley to face Penn State. On paper here in June, this looks like the toughest game on the schedule (since MSU faces Ohio State in East Lansing.) While this is very likely a loss, any time you look across the field and see James Franklin, you know your team has a chance.
That then brings MSU to the final two games versus Rutgers and at Maryland. While MSU should handle the Scarlet Knights, the Maryland game could get a bit scary. I would not want MSU to be sitting at 5-6 and needing a win to get to bowl eligibility, but that is a very real possibility. Actually, that is what I think is most likely.
Based on all of this, if I were forced to make a prediction, I am going to go with a slightly optimistic prediction of 6-6. I expect MSU to get a bit of a bump in energy from the new coaching staff, but this schedule does look challenging, and the start could be bumpy. In my opinion, the most likely scenarios are:
1) MSU starts out slow and finishes 2-2 in September, but then hits their stride enough to steal two games in October (two of vs. Michigan, at Indiana, and vs. Minnesota), and then beats Maryland and Rutgers to finish 6-6
OR
2) MSU starts out September at 3-1 (with a loss at BYU), only to pick up only one win in the tough six-game stretch (one of vs. Michigan, at Indiana, and vs. Minnesota) before finishing with two wins at the end of the season to get to 6-6.
At this point, I would be remiss if I didn't comment that this entire analysis hangs on the idea that the preseason rankings listed above are close to accurate. History has shown time and time again that they are not. Based on my analysis, preseason rankings tend to be off by an average of about 20 places.
So, MSU might wind of being a lot better than these "experts" believe. MSU might even be a borderline Top 25 team in a few key players emerge and the team stays healthy. On the other hand, MSU could also be a lot worse. My historical data suggests this uncertainty in how good MSU actually is will impact the final record by plus-or-minus two full wins. The difference between the predictions derived from the ESPN and CBS rankings show exactly how sensitive these numbers are to small changes in a team's ranking or in the rankings of some of their opponents.
For example, teams like Iowa, Minnesota, Indiana, and Michigan might wind up with a lot of injuries or just might be a lot worse than expected. Or, teams like Northwestern, Miami, and Maryland could be better than expected. MSU's actual strength of schedule could wind up being better or worse than we think here in June. The uncertainty of a team's schedule strength typically has a plus-or-minus one win impact on a team's final record compared to preseason expectations.
The final factor that tends to impact wins and losses I classify simply as "luck." Basically, this measures a team's ability to win toss-up games. This factor tends to impact a team's final win / lose record to a similar extent as schedule uncertainty: plus-or-minus about one win. Historically speaking my preseason expected win predictions are accurate on average, but the standard deviation is about two wins for the reasons stated above (uncertainty in team strength, schedule, and luck.)
So, as we sit here in June, the optimists may feel that a seven or an eight win season is possible. On the other hand, the pessimists likely see only three or four wins on the schedule. I can justify both arguments. I just hope we get a chance to find out.
That is all for today. Until next time, enjoy, and Go Green.
Comments
Post a Comment