.
The largest finance competition in the country, the Quant Challenge, organized by Itaú Asset, Itaú Unibanco’s investment manager, had as its winning project a robot called Persistence, developed by students from USP’s School of Engineering (Poli). At the end of December, the group, consisting of Alexandre Barsam Junqueira, Giovane Azevedo, Matheus Kanaan, and Eduardo Parri Cuperman, received R$15,000 as a prize.
This was the fifth edition of the challenge, which had the participation of 1,281 students enrolled in 126 universities located in 20 Brazilian states, which formed 462 groups. Those interested had to be enrolled in a bachelor’s or technologist’s degree course to participate. The goal is to stimulate innovation in the investment world through the development of quantitative strategies that help uncover asset performance and optimize portfolio composition.
Persistence uses Topological Data Analysis (TDA) instead of traditional classical statistics to optimize portfolios of shares listed on B3, Brazil’s main stock exchange. To illustrate Persistence’s work, the group uses the following analogy: while classical statistics only looks at the brightness of each star in a starry sky, TDA evaluates the stars in groups and how they relate to each other to form constellations. In this way, the robot balances the portfolios using the stock returns over a period prior to the investment, hoping that the returns will follow predictable trends, obtaining alpha (extra return) on the market.
“During the competition, we had the opportunity to put together the strategy, backtest it, and present the results to a panel of Itaú Asset managers with varied backgrounds and valuable insights,” said Eduardo Parri Cuperman, a student from the winning group, on Linkedin. “This competition was not only a learning milestone but also the start of the Poli Quant project, a quantitative finance league we are creating at Poli.”
Another group from Poli took third place. Students Luis Fernando Nangino Ferreira, Lucca Torchia Terra Kherlakian, Lucas Scaramelli, and Leticia Miwa Yamashita created the Coincierge project, a robot that analyzes news sentiment to identify times when currencies are overvalued or undervalued. In this way, he acts as a concierge for investors in the foreign exchange market, offering precise recommendations by analyzing the news from different countries. The group won a prize of R$4,000.
“After five editions of this competition, we have been able to see how rich Brazil is in quantitative finance talent, in all regions, and the potential that the subject has for application not only in the investment industry but in other segments of society,” says Rafael Lavrado, responsible for organizing the Quant Challenge. Students from Insper’s Economics course came in second place.
A panel of professionals from Itaú Asset evaluated the projects anonymously, without the evaluators knowing the authors of the works. The criteria analyzed were Robot Presentation, Strategy Concept, Modeling, Backtest, Results, and Project Conclusion.
.
With information from Itaú’s Corporate Communications Department
English version: Nexus Translation
