Dispatch from the Edge - February 2024
Humble beginnings, planting seeds, looking forward
“Give me six hours to chop down a tree, and I will spend the first four sharpening the axe.”
― Abraham Lincoln
Hi everyone! The past two months have been a whirlwind of activity here at the newsletter. We have been dutifully sharpening and sharpening and sharpening that axe, getting ready for the work to come. Preparatory work like this is just as important as the work itself.
And we are very, very excited about what’s to come.
We published four pieces explaining our approach to algorithm backtesting and strategy evaluation:
In-Sample and Out-of-Sample Testing: The backbone of our methodology, helping us find models that shine in the future, not just in the past
Cross Validation and Walk Forward Optimization: Exploring advanced techniques for robustness testing, their potential pitfalls, and outlining our probabilistic approach
Monte Carlo Simulation: Powerful reality checks via “what if?”-style simulations further cementing significance of strategies
Why do we need millions of backtests?: Our flag in the ground, highlighting metrics that we focus on to give you the best assessments of strategy effectiveness
These foundational posts culminated in version 1 of our "Really, Really Thorough Backtesting Methodology." This methodology is designed to be a guide for how we look at trading strategies. We want to be, well, very thorough and informative in backtesting. We see this methodology as an iterative guide. We intend to update it and engage with you all on how to best execute.
Armed with this methodology, we published results from two first research projects: the first analyzing 18 different moving averages across the 28 FX major pairs and the second conducting a sweeping backtest of the entire S&P 500, involving over 86 million simulations. This level of comprehensive research is rare outside the realms of academia and top-tier institutional firms, and we're just getting started. This is a powerful toolkit with a lot of potential. We know it, and there is a lot coming down the pipelines.
In addition to our focus on algorithm backtesting, we've also dipped our toes into the fascinating world of deep learning. While this exploration was momentarily paused to prioritize algorithmic trading strategy evaluation, the allure of deep learning in financial analysis is too potent to sideline for long. Rest assured, we will revisit this topic soon.
Looking ahead, Alpha on the Edge is brimming with potential harvests from our metaphorical seed planting. We are deeply grateful that you have decided to join us. Don’t hesitate to reach out and connect.
Until next time, keep on the cutting edge, everyone.