Stuyvesant Wins National Math Forecasting Tournament

Stuyvesant teams take first and second place at the National Forecasting Tournament.

Reading Time: 5 minutes

Cover Image
By Andrew Wille

Stuyvesant sophomores Andy Huang, Christopher Louie, Ryan Radwan, and Christopher Shen and juniors Rebecca Bao, Bingde Jiang, Edwin Lin, and Gloria Yip participated in the National Math Forecasting tournament from February 6 to April 14. The teams placed first and second respectively, winning a total of $4,000 for the school. Students on the winning teams also each received $500 scholarships, and all students who completed the whole tournament earned $100 Amazon gift cards.

The National Math Forecasting Tournament is a competition hosted by the Alliance for Decision Education in which students apply their mathematical skills to predict the likelihoods of real-life events. The competition was based around mathematical forecasting, a technique that uses data and research to make informed predictions. Students were presented with problems predicting real-life statistical trends, and they needed to use their research, math, and appropriate measures of confidence to accurately predict the future.

The tournament consisted of five two-week rounds, with five questions asked per round. Each member of the team had to select a probability from zero to one of a specific event occurring and state their confidence level in their answer. The questions were generated based on common teenage interests and had a wide range of topics, such as the likelihood of songs hitting top 100 charts, concert ticket prices, and COVID-19 data. “We worked really hard to think about the teenage brain in terms of selecting those questions,” Alliance of Decision Education Director Megan Roberts said.

The 12 Stuyvesant teams that participated in the competition were mentored by Stuyvesant math teachers Andrew Wille and Patrick Honner. “Our role was to help students understand what forecasting is, some of the mathematical concepts that are important for forecasting, and how to apply those concepts,” Honner said. “Wille and I both have classes full of enthusiastic math students, so we asked students to express interest if they thought this would be something they would enjoy. The contest was [also] a great way to see math in a new way, to see math applied to fields that interest different people.”

Both teams expressed their gratitude to their coaches and teammates for their valuable guidance and collaboration. “Mr. Honner and Mr. Wille were really supportive during the contest, and we could tell that they were passionate about what they were doing,” Bingde (Charles) Jiang said. “It was definitely a lot of fun to be able to do all this stuff with my current group mates, especially with all the work that they did.”

The teams met after school during the tournament weeks to practice and work on their predictions. “It was just such a fun experience to get to interact in a math-type way with these students in a different context than a classroom. They would show up, bring their research and their energy, and we would guide them and they were quite successful,” Wille said.

For the junior team, research methods and strategies depended on the type of trend that was being questioned. “A lot of the questions that they gave us were time-based. For example, with gas prices, you can go along with the trend. And if the gas price reaches a certain height, it’s impossible to drop four cents in a day. Usually, for questions [like] those, we would wait it out, play the waiting game. For other ones like movies, we based it off of the opinions of other people by using movie reviews and critics,” Gloria Yip said.

In terms of strategy, students also had to decide between using bolder but riskier predictions and less variant but safer predictions. The sophomore team believed that taking bigger risks, while not a sure victory, greatly contributed to their success. “We [made] the bold decisions to go absolute,” Christopher Shen said. The team chose more extreme predictions as opposed to numbers in the middle.

On the other hand, the junior team chose to make safer predictions, believing that the scoring system incentivized playing it safe. “We decided to avoid making bolder decisions, unless we knew the prediction was guaranteed to be true,” Edwin Lin said. “And even if we were confident in our decision, we would still keep our predictions within [an] 80 to 90 percent [range].”

Team scores were measured using a Brier score, which calculates how far a forecast is from the truth; the lower the score, the better. Each member had to pick a number between zero (zero percent chance) and one (100 percent chance), which represented the probability of a certain event happening. The further away a team’s prediction was from the truth, the higher their Brier score was. A safe choice, such as 0.5, would have prevented them from achieving a high Brier score, while riskier choices, like 0.2 or 0.7, had the potential to result in more drastic losses or gains. A team’s Brier score was the average of each of their individual scores.

For many participants, this tournament also allowed them to develop their math interests and provide insight into future pursuits. “I want to invest in stocks later, in [the] future, and that’s basically forecasting because you are trying to see if it will go up or down,” Andy Huang said.

Many students also felt as if they were also able to learn more about fields that were previously unfamiliar to them. “There were a series of questions on Biden’s approval rating. I’ve never really been a politician but having researched all of that, I found out that U.S. presidents in the past don’t usually have a good reputation in terms of approval [ratings], so that was something interesting to witness,” Jiang said.

Overall, Honner and Wille attribute Stuyvesant’s success in the tournament to the hard work and competitive spirit of the participants, as well as their genuine interest in the mathematical field. “Everyone had a really good time but I think the students who really liked it had a tendency to want to research something in a quantitative or mathematical way, which I think is a lot of Stuyvesant students,” Wille said. “It was students who wanted to take a look at real world events that haven’t occurred yet and applied some sort of quantitative reasoning to those events.”

The Alliance of Decision Education hopes that the competition will continue to help students develop effective decision-making skills. “We [wanted to help] students [learn] the concepts and skills of how to be great decision-makers,” Roberts said. “So much of that content has to do with the skills of forecasting, being able to understand base rates (a percentage of a population that has a certain characteristic), being able to think probabilistically, being able to see the inside-outside view, being able to collect and evaluate data, all these sorts of concepts and skills.”

Ultimately, participants enjoyed the tournament and were able to develop and practice making well-informed, confident decisions through forecasting. “I like to say [that] not everybody wins, but everybody learns. And so it’s a really fun way to make yourself a better decision maker while also learning the skill of forecasting,” Roberts said.