Fall 2022 Recap
Dear SIF Community,
This past semester has been a semester of great progress for the Smith Investment Fund Quantitative Team. From an entire team structure revamp to an impressive class of new members, we have a lot of exciting news to share.
Recruitment
We opened the Fall semester with our established recruitment campaign. Our application was open for the first 4 weeks of the semester, during which we held a variety of recruiting events. We received a wealth of qualified applicants with a variety of different backgrounds. Choosing which candidates to interview was a challenge, and making the final selection required extensive debate and discussion. Our recruitment campaign concluded with the addition of 16 new driven and talented members to the Quant team. All of these new members participated in our “Intro to Quantitative Finance” curriculum, where they honed their data analysis skills and learned more about financial markets. The semester was capped off by an alpha competition, wherein the new members put to use everything that they learned over the past several weeks. Their alphas displayed a strong understanding of the curriculum material, particularly the Markowitz efficient frontier, and represented some of the most creative and impressive alphas we have seen from first year members.
For example, Ocarina Lin, Julia Chen, and Elizabeth Qiu combined fundamental ideas of momentum and rolling double average to categorize the lowest and highest percentile of the universe and their moving averages to create an enhanced moving average strategy. For their alpha, Aaquib Syed and Phillip Guo developed infrastructure to optimize a portfolio based on the index/factor models that were covered during lectures, including Fama-French, CAPM, and Carhart four-factor. Their contributions allow users to easily select a few available factors from the SIF factor database or create their own factors, quickly build a factor model based on them, and then build a portfolio by maximizing estimated Sharpe ratio with Bayesian optimization.
Overall, we were extremely impressed with all of the new members’ progress this semester and look forward to seeing their contributions to the club in coming months.
Team Structure
As the Quant team has grown, we have discovered that our previous team and meeting structures are not optimized for larger teams. In response, we adopted a new team structure that was originally proposed by members Sam Galita and Ravi Panguluri. While previously we were divided into Research and Infrastructure, we now have four verticals – Live Trading and Alpha Development, Cryptocurrencies, General Research, and Internal Tools. Each vertical is made up of researchers and infrastructure engineers that are focused on similar projects. With this new structure, we have seen increased collaboration, bleeding edge projects, and greater contributions by new members.
Projects
SIF Quant members across all verticals embarked on exciting and ambitious projects this past semester. Summaries of these projects can be found below.
Kaushik Vejju continued his work on SIFSearch, an internal search engine that allows SIF members to upload and search for SIF-related media, including code repository links, research articles, and past SIF projects. This Django-based application uses Algolia's Search API to provide an efficient search experience and allows users to upload media in the form of links and multiple file types.
The live trading infrastructure team spent the semester designing a system for high frequency alpha development. Python classes were created for individual strategies and their order submission and position tracking systems. Their goal is to build functional backtesting and live trading classes by the end of the Spring semester.
Sam Galita worked on performing exploratory analysis on the higher frequency data, planning HFT research for future semesters, and designing the HFT infrastructure to interface with researchers.
Over the winter break, Sam Galita and James Zhang implemented an enhanced pairs trading strategy that uses a variety of machine learning methods and pair validation techniques in order to more efficiently select profitable pairs for trading from a larger universe than previously possible. In addition to weight optimization on the strategy's portfolio, they also plan on redeveloping the trading method of the strategy to better time entries and exits and make it compatible with the new HFT infrastructure.
Abhinav Modugula did some preliminary analysis on the minute candle data with Sam and found some promising insights we could use for alpha development. In addition, Abhinav developed some sector based alphas involving momentum and reversion.
According to the paper Bunn, Gianfreda 2018, Richard Zha worked on implementing a stochastic model for California day ahead electricity price formation. He took a distributional approach with the GAMLSS R library, statistically modeling the first four moments of day ahead prices as a function of known fundamental exogenous drivers of electricity price, specifically focusing on the impact of renewable generation. Richard also made plans for future research using this model in an operational context, dealing with problems such as optimal battery charging and generator risk management.
Along with starting to onboard macroeconomics factors, Vikas Reddy explored CPI and Inflation data to see how these signals were reflected in the stock market and vice versa.
Over winter break, Suryaa Rajinikanth started working on GoBackTester, a project written in Go and Python that is able to test trading strategies with forex. Using TrueFX’s historical data, it is able to parse through millisecond data to test strategies that aim to take advantage of forex’s volatility to make a profit. The program leverages GoRoutines and concurrency to quickly produce outputs that are much faster than any Python equivalent. An example alpha that capitalized on volatility was able to create 3.4x profit in two days while trading on EUR/USD.
Conclusion
We look forward to building on the success of this past semester with even more ambitious projects this Spring. We would like to thank our past presidents, Uro Lyi and Ryan Downing, who graduated last spring after making immense contributions to the Quant Team. Peter Geertsema, a long-tenured, true leader of the Quant Team graduated this Fall. We wish Uro, Ryan, Peter, and all of our alumni the best for their bright futures. Thank you for your continued support of the Smith Investment Fund. We look forward to sharing more exciting news with you soon.
Best,
SIF Quant Executive Team
