Skip to main content

Parallel Universe March Madness: Round of 32

If the many-worlds interpretation of quantum mechanics is correct, there are an infinite number of universes without the with COVID-19 pandemic. In those universes, the NCAA Tournament just wrapped up its first weekend.  I have been using the results of a single Monte Carlo simulation of the tournament to see how one mathematically consistent version of the tournament might have played out.  Here are those results through the second round.

Midwest Region:
  • No. 1 Kansas beat No. 9 Marquette, 76-63
  • No. 5 BYU eliminated No. 4 Louisville, 70-67
  • No. 3 Villanova sent home No. 6 Iowa, 75-64
  • No. 2 Michigan State pummeled No. 10 Texas Tech, 77-61

(I also put together a simple box score generator, so perhaps the game went like this...)

After their offensive struggles in the first round, Cassius Winston and Rocket Watts more than made up for it in the second round game against Texas Tech. Watts started hot and stayed that way, putting up a game-high 21 points on 9-of-13 shooting (including three of five from three).

In the second half, the Red Raiders tried to contain Watts, but then Cash went off and added 19 points on 8-of-13 shooting (three of four from three) and seven assists.  Aaron Henry also contributed a workman-like 11 points, 10 rebounds, and five assists.

Tillman struggled a bit from floor, hitting only 1-of-5 shots, but did finish with eight points and nine rebounds.  Bingham and Kyle Ahrens provided some punch off the bench with six points each, while Loyer, Brown, and Hall added two points each.  Overall, MSU shot 53 percent from the field and 40 percent from three, although Cash and Watts were the only players to score from deep.  MSU moves on to face Villanova in Indianapolis on Thursday.

As for my original Midwest Region picks, I got three of the four, Kansas, BYU, and MSU.  I did pick Iowa to upset Villanova, but I will take 75 percent.

South Region:
  • No. 8 Saint Mary's upset No. 1 Baylor, 67-65
  • No. 4 Wisconsin ended the season of No. 12 New Mexico State, 68-61
  • No. 6 Michigan eliminated No. 14 South Alabama, 72-64
  • No. 2 Creighton defeated No. 7 Colorado, 71-67

Of course, the upset of Baylor in the second round would have been big news, but it would not have been a shock.  Also, Michigan advancing to the Sweet 16 due to facing weaker than expected opponents in exactly the type of thing that probably would have happened.  (That's the Michigan difference.)  All of a sudden, the South Region looks ripe for the picking.

As for my pre-simulation picks, Creighton was the only team that I would have correctly picked.

West Region:

If you think that the action in the South was wild, here is what my simulation spit out for the West:

  • No. 9 Rutgers upset No. 1 Gonzaga, 71-69
  • No. 5 Ohio State beat No. 4 Oregon, 69-64
  • No. 3 Kentucky defeated No 11 Xavier, 77-60
  • No. 7 Illinois upset No. 2 San Diego State, 67-66
Clearly, this would have been a fantastic result for the Big Ten, and based on how strong the Big Ten is in Kenpom, it is not at all unreasonable.  While it was tempting in my analysis to knock either Baylor or Gonzaga out in the 2nd round, I choose not to. The simulation took out both.  As for my predictions, I only got Ohio State correct as a Sweet 16 participants.

East Region:

Results in the East Region were a bit more tame, with one exception
  • No. 1 Dayton outlasted No. 8 Providence, 71-67
  • No. 4 Maryland took down No. 5 Auburn, 72-66
  • No. 3 Seton Hall sends home No. 11 East Tennessee State, 75-64
  • No. 7 Arizona edged No. 2 Florida State, 72-68
I actually got three of these four picks as well, including the "correct" pick of Arizona over Florida State.  The only difference was I had Penn State advancing to the Sweet 16 instead of getting upset in the first round.  In total, I got correctly "predicted" exactly half of the Sweet 16.

In summary, here is the updated Parallel Universe NCAA Tournament Bracket



The main lesson here is that the variance in the tournament results is very real.  For both the predictions that I made about the tournament and the results of the simulation, I was using the same source data set (Kenpom efficiencies) to generate the results, and yet I still got only half of the picks correct.  There was a Method to my Madness, yet it still didn't work that well.

For now, just like the actual NCAA Tournament, we will take a break for a few days until "play" resumes on Thursday night.  Stay tuned.  


Comments

Post a Comment

Popular posts from this blog

March Madness Analysis: Did the Selection Committee Get it Right in 2025?

I will be assembling my "usual" stats-based analysis of the bracket, complete with picks a little later this week (but before Thursday). For now, I had some thought on the bracket. In general, MSU's draw is about as good as fans could expect. I will go into more detail on that later. As for the job that the committee did... I am far from impressed. Once again, there are multiple errors in team selection, seeding, and bracketing as a whole. Let's look at each one in turn. Did the Committee get the right 68 teams? More or less. This is the area where I am the least concerned. As I mentioned yesterday, my metrics had UNC safely in the field and not even in the First Four, but I swapped them out for WVU at the last minute. UNC's single Q1 win gave me too much pause. I felt slightly vindicated when UNC made it.  My biggest beef is with Texas making it in at 19-15. That's just too many loses. Yes, they had 7 Q1 wins, but that also had 5 loses outside of Q1 and an ov...
Post a Comment
Read more

Dr. Green and White Helps You Fill Out Your Bracket (2025 Edition)

For my money, we are all of the cusp of the best three weeks of the entire year. We just wrapped up two weeks of conference tournaments, but those were just an appetizer to the main course that is yet to come.  The powers that be gave us the menu on Sunday evening for the feast that is to come. Now it is time to enjoy a brief break and palette cleaner before we all make our selections. But what shall we choose? Which tasty little upset looks the best in the first round? Which teams are most likely to be sweet in the second weekend? Which quartet will comprise the final course? Over the years I have developed a set of analytics and computational tools to gain a better understanding of the mathematical underpinning of the NCAA Basketball Tournament. My methodology has a solid track record of correctly identifying upsets and sometimes doing more than that. In 2023, I used data to correctly predict that No. 4 seed UConn win the National Title. There is no foolproof way to dominate your...
Post a Comment
Read more

2025 College Football Analysis, Part Two: A Deep Dive into MSU's Schedule

In part one of this year's math-based preseason analysis of the college football season, we looked back at the 2024 season. Through that analysis, we learned about the historical accuracy of preseason polls (plus-or-minus 25 positions) and regular season win totals (plus-or-minus 2.5 wins). We also explored the impact of changes in ability, schedule, and luck. Now it is now time to shift focus to the 2025 season. Over the years I have developed and refined a way to simulate the entire college football season using schedule information and preseason rankings as the only inputs. I will soon go through the full details of what I learned from this exercise.  For today, I will focus exclusively on what it says about the Michigan State Spartans. We will take a close look at the Spartans' 2025 schedule from three different points of view. Opponent Overview The best place to start this analysis is with the simulation's inputs. Figure 1 below summarizes the preseason rankings (w...
Post a Comment
Read more