On Wednesday evening, Michigan State University will officially open Big Ten conference play in basketball with a road game at Minnesota. Before the Spartans tip off, I wanted to do a deep dive into the entire Big Ten schedule in order to better understand what we can expect from the Spartans and from the Big Ten race overall.
The method the I use to perform this analysis is based heavily on efficiency margin data generated by Ken Pomeroy. In my opinion, "Kenpom" data is the gold standard of basketball data analytics. It correlates very well with point spreads and therefore the data can be used to generate win probabilities for any future basketball match-up. With this tool in hand, it is possible to simulate the results of the entire Big Ten regular season as well as the Big Ten Tournament.
Big Ten Regular Season Odds
To start this analysis, let's first take a look at how the 14 members of the Big Ten stack up according to Kenpom efficiency margin data that I pulled a few days ago, just prior to the start of Big Ten season. Note that efficiency margin is defined as the point differential that a team is expected to have against an average Division 1 team in a game of 100 possessions. Also, note that for this entire analysis, the results of the first few Big Ten games that have occurred this week are not being considered.
 |
| Figure 1: Kenpom efficiency margin data for all 14 Big Ten team prior to the start of conference play |
Currently, Purdue grades out as clearly the best team in the conference. After the Boilermakers, there is a cluster of five teams, including Michigan State, that have an efficiency margin right around 20.0. After that, the quality of teams in the Big Ten trails off until we reach the bottom four teams in the conference (Penn State, Rutgers, Nebraska, and Minnesota) who are all markedly weaker than the rest of the conference.
The data shown in Figure 1 is the core input to all of the analysis to follow. That said, this data should be taken with a grain of salt. This early in the season Kenpom still seems to be biasing his data based on the assumed strengths of teams in the preseason.
For example, Kenpom had several Big Ten teams in his preseason top 10, including Michigan (No. 2), Illinois (No. 5), and Ohio State (No. 8). All three of these teams have struggled so far in 2021, and it does not appear that Kenpom's analysis has quite caught up with this reality. As the season continues to play out, the predictions should become more accurate as that preseason biasing is phased out.
Based on the data shown in Figure 1 and the Big Ten schedule, it is possible to simulate the results of the regular season. As discussed above, the data above does carry with it some uncertainty. Fortunately, this uncertainty is quantifiable based on historical data, and I account for this variance in my simulation.
Table 1 below gives the result of my 100,000-cycle Monte Carlo simulation of the full Big Ten regular season.
 |
| Table 1: Preseason Big Ten win probability matrix based on the results of a 100,000 cycle Monte Carlo simulation which accounts for the known variance in team strength at this point in the season. |
Based on this data, Purdue is expected to win the Big Ten by roughly three games, but second place appears to be up for grabs. As for Michigan State, the Spartans are expected to win right around 12 games in conference play. There is about a 60 percent chance that MSU will win between 10 and 14 games and a 20 percent chance of winning 15 games or more.
Table 2 below summarizes the odds for each team to win the Big Ten regular season title.
 |
| Table 2: Results of the full Big Ten season simulation showing the odds for each team to win or share the Big Ten regular season title and the number of wins needed to do so. |
As expected, Purdue has a clear edge in odds. Specifically, my simulation suggests that the Boilers have over a 56 percent chance to at least share the title. Ohio State and Iowa both graded out with a 12 percent chance, while Michigan's and Michigan State's odds are virtually identical at just over 10 percent. I will also note that the odds suggest that a record of at least 17-3 may be needed to hang a regular season banner this year.
Big Ten Tournament Simulation
As a part of my simulation of the Big Ten regular season, it is also possible to calculate the seeding for the Big Ten Tournament as well as the results of the tournament in each scenario. Table 3 shows the current odds for each team to earn a given seed in the Big Ten Tournament.
 |
| Table 3: Big Ten Tournament Seed distribution matrix |
Obviously, Purdue is the clear favorite to earn the No. 1 seed, but after that, the field is pretty wide open. Most likely, the more important fight will be to secure one of other tournament double byes that are given to the top four seeds. Right now, Ohio State, Iowa, Michigan, Michigan State, and Illinois all look like strong candidates to make the top four, but two of those teams will not make it. The Spartans currently project as the No. 5 seed.
As for the results of the Big Ten Tournament, those odds are shown below in Table 2.
 |
| Table 4: Odds for each team to advanced past each round of the Big Ten Tournament, based on the results of 100,000 Monte Carlo full season simulations and a calculation of the tournament odds in each scenario. |
Purdue's odds to win a tournament title (37 percent) are a bit lower than the odds to win a regular season title (56 percent). The Spartans' odds to hang a tournament banner currently project at eight-and-a-half percent. I will note that the odds in Table 4 map almost exactly to the current Kenpom efficiency data and is therefore only as accurate as that data set.
Big Ten Strength of Schedule
While the current efficiency margin data has a very strong influence on the projected final results of the Big Ten season, it does not match perfectly. For example, Figure 1 suggests that Michigan is the second best team in the conference, but Table 1 suggests that the Wolverines have only the fifth highest number of expected wins. The reason for this difference is due to the unbalanced 20-game conference schedule.
Table 5 below summarizes the overall Big Ten schedule and highlights the games where two Big Ten opponents play each other only once. The cells highlighted in green are a home game only for the team listed in the row, while the cells highlighted in orange represent a road game only for the team listed in the row.
 |
| Table 5: 2021 Big Ten schedule matrix. |
Based on this table, Michigan State does appear to have a fairly favorable schedule, based on the current Kenpom efficiency margins of each team. Most notably, the Spartans only face Purdue once, and that game will be played in the Breslin Center. In addition, Michigan State plays Iowa and Ohio State only once as well, although both of these games are on the road.
Over the years I have developed my own formula to quantify strength of schedule. In short, I use efficiency margin data to calculate the number of expected wins that an average power five team should should get if they were to play each team's schedule.
I also perform a second calculation where I attempt to correct for one quirk when it comes to comparing strengths of schedules among a group of teams in the same conference. Good teams tend to have easier schedules because they do not play themselves and vice versa for the weaker teams. Figure 2 below shows the results of my strength of schedule calculation for the 2021-22 Big Ten season using both methods.
 |
| Figure 2: Big Ten conference strengths of schedule. The left panel shows the raw strength of schedule calculation. The right panel shows the adjusted values are correction for the strength of the team in question. |
Figure 2 confirms that Michigan State does, in fact, appear to have a slight schedule advantage this year. The Spartans' schedule currently ranks second easiest behind Ohio State. A glance back at Table 5 shows us why. Ohio Sate plays Michigan twice this year, but they only play the other four teams currently projected in the upper tier of the conference once. This explains why Ohio State's odds are slightly better than the other Big Ten teams with a similar efficiency margin.
Iowa has the third easiest schedule based on this method. Illinois and Michigan, however, grade out as having schedules that are in the bottom half of the conference. Based on this analysis, MSU's schedule is easier than Michigan's schedule by about half of a game. As for Purdue, the raw strength of schedule data suggests that the Boilermaker's schedule is relatively easy.
However, the right-side panel of Figure 2 shows that Purdue's schedule is actually the toughest in the conference when the the benefit of them not having to play themselves is removed. Illinois and Michigan have similarly tough schedules after this correction.
MSU Schedule Overview
Finally, I would like to take a look at Michigan State's schedule in some detail. Figure 3 gives the currently projected point spreads and victory probabilities for all 20 of the Spartan's conference games.
 |
| Figure 3: Projected point spread and victory probabilities for Michigan State's 20 Big Ten conference games. |
There are a few different ways to digest this data. As shown above in Table 1, the expected win total for Michigan State (i.e. the sum of the probabilities in Figure 3) is just under 12 wins. That said, MSU is also currently projected to be favored in 14 of the 20 conference games. Of the games where Michigan State currently projects as an underdog (at Michigan, at Wisconsin, at Illinois, at Iowa, versus Purdue, and at Ohio State) the projected point spread is three points or less in all of the games except the one against Purdue.
Figure 3 also gives a preview of the likely ebb and flow for the Spartans' season. The front end of the schedule is fairly light. The Green and White are projected to be comfortable favorites in the two conference games played over the next week. Then, the first five games of January are fairly manageable as well, with the exception of road games at Northwestern and at Michigan.
The odds suggest that MSU should go at least 5-2 in this stretch to stay on target to win 12 games. If MSU has aspirations to challenge for the Big Ten Title (or at least a top four finish), getting six, if not all seven wins in this stretch might be necessary.
In mid January, the Spartans get a week off to prepare for a nasty stretch where they will play four-of-five games on the road, starting with a trip to Madison on Jan. 22. The road game at Rutgers is the easiest game in that stretch, as the rest of the games are near toss-ups. The Spartans need to win at least three of those games games to stay on target.
On Feb. 9, Wisconsin plays the return game at the Breslin Center which is the start of a slightly lighter four game stretch, three of which are at home. Winning three of those four games would keep MSU on track.
This leaves just the final four games, three of which are fairly brutal. Over a span of nine days in late February and early March, the Spartans travel to Iowa, play Purdue at home, and then travel to Ohio State. The season then ends on a softer note with a home game against Maryland. If Michigan State can just split those four final games, it would again keep MSU on track.
Based on the data that we have in early December, this is what MSU and Big Ten fans should expect from this year's basketball season. That said, as the games are played and the data accumulates, the odds will change. I will be popping back in throughout the season to provide an update on these odds. Stay tuned, and Go Green.
Comments
Post a Comment