Five Andover students participated in the international M3 challenge.
After spending an entire day in early March crammed into a basement working, a team of five Andover students has advanced to the Final Round of the MathWorks Math Modeling (M3) Challenge, placing in the top six of 655 competing teams. Angeline Zhao ’25, Tianyi Evans Gu ’25, Yifan Kang ’24, Eric Wang ’25, and Anthony Yang ’25 will represent Andover in the international math competition on April 29 in New York City, competing for a share of 100,000 dollars in scholarships.
Zhao, one of the original coordinators of the team, described the M3 Challenge and previous topics that the competition has focused on. She highlighted applied mathematics’ presence in the world, and how the M3 Challenge embodies that spirit.
“[The M3 Challenge] is an annual competition for high school students in the [United States of America] and the [United Kingdom] to practice dealing with a real-life situation using different mathematical model techniques. [The goal] is to make predictions for the future, the way that it’s framed is that you’re supposed to give recommendations to a U.S. government entity… The topics are typically things that are impacting our world today. Our topic this year was the housing crisis and homelessness. Last year, it was about E-bikes. So, they’re all relevant topics and we have to find different modeling techniques to predict how these things will change in the future and factors that will influence it… Mathematics is always something that I’ve enjoyed, and applied math is something that I’ve always wanted to get more into especially since I really enjoy statistics,” said Zhao.
Over the course of 14 hours, the team tested four different statistical models and ultimately submitted a 31-page paper describing their findings. Kang, the oldest member on the team, discussed challenges the team faced to achieve reliable and productive results.
“The hardest part [of the process] was reading past papers and generally getting a sense of what [statistical] models we can use to analyze the data provided by the competition. Since [the M3 Challenge] is a math modeling challenge, it’s mostly about applying reasonable models and getting the results to be as accurate as possible,” said Kang.
Zhao also commented on the topic of preparing for the Challenge, noting the learning curve that came up in the process. She delved into how the team decided which mathematical processes to use and how to apply them in the most efficient way.
“The hardest part is the fact that there is a learning curve to getting used to how you’re supposed to develop all these different models because a lot of the models are quite complicated and involve a lot of programming or different complex modeling techniques. Some of them involve machine learning, which can be quite time-consuming to figure out, so that’s definitely a learning curve. You have to figure out when you should be using certain models and how to use the simplest model for certain situations so you’re not overcomplicating them, all while still coming up with a unique and accurate solution that fits the circumstances of the problem… The best part of [this process] was getting to work together and seeing everything come together in the end,” said Zhao.
Wang discussed the approaches that he and the rest of his team took to solve this year’s problems. He mentioned his surprise regarding how many speculative guesses the team had to make while using models.
“Our approach was to really just see how much information we could gather in 14 hours. It was a very limited period of time, and we didn’t have that much experience with modeling. We looked at some of the past teams’ papers, how they decided to approach the challenge, and what types of models they used, and we took that information and tried to synthesize our own model to use for the challenge. I learned that there are a lot of assumptions you have to make when modeling something. Obviously, there are a lot of unpredictable factors, and a lot of the assumptions may seem quite sketchy, but it is necessary when trying to make a model for predicting future scenarios,” said Wang.
Khiem DoBa, Instructor in Mathematics, served as the Faculty Mentor for the Andover M3 Challenge Team. DoBa expressed his pride in the work done by the team, praising how much they had managed to accomplish in just 14 hours.
“I’m proud of their talent and their work ethic and determination. In order to produce these solutions from the kind of open-ended challenge [they’re provided], it was an incredible amount of work, and the team produced a thirty-one page paper that used multivariate [Linear Regression Model] to come up with a model to address the question the challenge provided,” said DoBa.
Editor’s note: Tianyi Gu is a Managing Editor for The Phillipian.
