New Election Prediction Model by Students at Brown University Outperforms Industry Leaders

Leading-Edge Machine Learning Technology Sets New Standard in Election Forecasting with a Focus on Transparency.

PROVIDENCE, R.I., June 28, 2024 /PRNewswire/ — Brown Political Review proudly announces the launch of, an innovative election prediction platform developed by a team of 13 dedicated Brown University students. The team has created state-of-the-art machine learning models that consistently outperformed the leading election prediction models in back-tests on the 2020 general elections and the 2022 midterms, setting a new benchmark for accuracy in election forecasting. invites voters and campaigns nationwide to use the tool in the current 2024 election season.

“We are thrilled with the results of the technology we’ve built,” states Asher Labovich, founder of “We made it open-source so that anyone can see how our models arrive at their conclusions. Many voters believe election prediction models are biased and don’t trust them. So it was important to our team that the results be data-driven with minimal assumptions. In our model, the data speaks for itself.”

Over the past year, the team has dedicated themselves to developing a more accurate and interpretable election forecast than any other model currently available. utilizes an ensemble of 1,000 different models and over 500,000 data inputs that allow the tool to dynamically understand the interactions between various electoral influences.

Using its innovative machine learning model, the back-tests on the 2020 general elections and the 2022 midterm elections consistently outperformed the leading election prediction models such as FiveThirtyEight, The Economist, Split Ticket, and Race to the WH. In the current election year, predicts, in contrast to its peers, that most of the swing states in the presidential election will vote in unison–either all red or all blue.

Currently, the model predicts that Trump has a 56% chance to win the electoral college, while the Senate is predicted to be evenly split. (The model updates every 24 hours). The model also provides the likeliest outcomes for the presidential map.

“We want to reach the ‘edge of predictability’ and provide an election prediction model that not only yields the most accurate results but also offers unparalleled transparency and accessibility. We believe that by demystifying the election prediction process and making our models open-source, we can empower an even more informed and engaged citizenry,” declares Labovich.

Guided by principles of transparency, innovation and accessibility, all of the election predictions come with simple graphical explanations to help users understand the underlying data and analysis. Ultimately, the team hopes to transform the election prediction landscape and inspire political engagement.

About by Brown Political Review aims to demystify the world of election predictions for politically engaged citizens and stakeholders. Developed by 13 college students from Brown University, consists of state-of-the-art machine learning models that predict key U.S. elections with minimal assumptions. The result is the most accurate election prediction technology available. Transparent and accessible to all,’s models are open source with graphical explanations for all predictions to help users understand how the predictions were reached. Working to achieve the “edge of predictability,” the team is on a mission to empower informed political engagement by educating the public on the power and usefulness of prediction technology. To learn more visit

Anita Lane

SOURCE Brown Political Review