SIAM Financial Mathematics Student Programming Competition

Danny Kurban and I participated and placed 4th place in the SIAM FM21 Student programming competition, and the specific problem details are here. The written report is here and below is a short summary of our solution.

We devise a simple momentum-based stock trading strategy that maximizes a mean-variance objective function by dividing the portfolio into a “winner-basket” that increases expected return and a “diversification-basket” that reduces risk. For investors with low risk aversion, the strategy overweights the “winner-basket”, while for risk averse investors, the strategy converges to an equal weight strategy. We use a market simulator in Matlab to show that our strategy consistently beats the benchmark equal weight strategy as well as more advanced strategies based on deep neural networks, especially when the time horizon is short relative to the number of available stocks.