Artificial Madness: How Sweet it is! Proprietary Model Simulates Sweet 16 Games

Original article can be found here (source): Artificial Intelligence on Medium

Artificial Madness: How Sweet it is! Proprietary Model Simulates Sweet 16 Games

Oh, how sweet it is! We have 16 teams remaining in our Artificial Madness after simulating the second round of games yesterday. Which teams have moved one step closer to the Final Four?

Stats Perform’s award-winning research group is simulating the final scores of our tourney matchups using a model it created with advanced data from our AI team along with other metrics.

How are we running this? Our model takes that advanced data and calculates each team’s offensive and defensive ratings from the 2019–20 season. Then we’re also able to project scores of the matchups our model has produced from the seeding process that’s based on group tiers.

Keep in mind that there will be some variability, as we are running the model only once. This leaves room for upsets — just like there should be in March.

Think of it as a weighted coin flip, and the weights are the probability of each team winning. For example, Kansas has a 96.7% chance of beating Robert Morris, but Virginia only has a 55.9% chance of beating USC.

Below are also hypothetical scenarios on what might have happened in each game given the score:

Maryland is SO close to pulling off the upset! Kansas survives on what either was a game-winning shot from Devon Dotson or a missed attempt from Anthony Cowan Jr. Perhaps Dotson missed his shot, but Azubuike was there to tip it in as time expired. This would have been a fun one to watch, especially the matchup between 7-foot Azubuike and 6-foot-10 Jalen Smith. The Jayhawks had a 66.2% chance at beating Maryland, but the Terps nearly knocked them off. This would be Kansas’ fourth trip to the Elite Eight in the last five years.

Wichita State beat Hofstra to advance to the Sweet 16, but the Shockers’ journey ends here with a loss to the Spartans. Michigan State was excellent defensively this season, holding opponents to 37.9% from the field this season, fourth-lowest in Division I. That defense, combined with excellent three-point shooting from Cassius Winston and impressive scoring from freshman Rocket Watts, allow the Spartans to blow past the Shockers and reach the Elite Eight for the second straight year. Could it be back-to-back Final Four appearances for Michigan State?

Our model adjusts for the pace of both teams to come up with a score that would be similar to the outcomes that the team and its opponent would have. Wisconsin and LSU have different styles of play — the Badgers play a slow game that uses the shot clock while the Tigers go quick and score on the fast break. Based on the score, it appears that this game looked more like LSU’s style of play, and yet Wisconsin won. The Badgers averaged 67.2 points in the regular season but are averaging 84.3 points in their three tournament games. This is the furthest that Wisconsin has gone in the NCAA Tournament since reaching the championship game in 2015.

Though Seton Hall is a 3 seed, our model sees this as a big upset. Duke had a 64.2% chance at winning this game, but the Pirates survive and reach the Elite Eight for the first time since 1991. Remember that Seton Hall was not even the favorite in the second round against West Virginia. It will be interesting to see how the model thinks the Pirates, who will be the higher seed, will do against Wisconsin in the next round. Myles Powell wants to make sure his last few games are heard around the country.

What a run for John Groce and Akron, pulling off the biggest upset of the tournament. Of all the Sweet 16 games, the Zips had the lowest chance of moving on at 20.1%. This is why simulating a model just once can have such variance — that 20.1% proved to be the victor this time around. Baylor barely loses, too — it could have been the Bears’ first Elite Eight appearance since 2012. Seemingly everybody is on the Akron bandwagon after this win — LeBron James, a native of Akron, Ohio, tweets his support, while everyone who loves MAC-tion is all in on this team.

Akron will have to face Creighton, whose three-point shooting dooms the Hoosiers in the Sweet 16. Though Indiana was a 10 seed, the chances of third-seeded Creighton advancing were smaller than one may think — the Jays had a 56.4% chance at winning, according to our model. Indiana has not made the Elite Eight since 2002, while Creighton reaches the Elite Eight for the first time in school history.

Gonzaga and Ohio State last met in the second round of the 2018 NCAA Tournament, with the Zags winning, 90–84. This time, the Bulldogs make sure that the Buckeyes do not get close to winning, catching fire from three while having one of their best defensive efforts of the season. This is the fourth time in the last six seasons that Gonzaga has reached the Elite Eight — the Zags made it to just one Elite Eight in school history prior to that (1999). What a job welldone by Mark Few.

Illinois had a great run to the Sweet 16, knocking off San Diego State and reaching the second weekend for the first time since 2005. This Illini team gave fans a lot of hope for the future. Losing to an impressive Oregon team shows what this team can do instead of what this team failed to do. The Ducks had a 64.0% shot at beating the Illini, so this win is not much of a surprise.

We had six Big Ten teams reach the Sweet 16, but only two moved on to the Elite Eight. In addition, a Cinderella story has emerged thanks to Akron, the first team seeded 13 or lower to reach the Elite Eight (seeding started in 1979). There is only one 1-versus-2 matchup in the Elite Eight, and at least one team seeded 3 or higher will reach the Final Four. That’s Artificial Madness.

Come back on Thursday as we will reveal the Elite Eight in our Artificial Madness. You can check us out @StatsBySTATS for the latest news and information in the sports world.