Jakub Michańków

Application of Deep Reinforcement Learning to At-the-Money S&P 500 Options Hedging This paper explores the application of deep Q-learning to hedging at-the-money options on the S\&P~500 index. We develop an agent based on the Twin Delayed Deep Deterministic Policy Gradient (TD3) alg...

Our new paper is out: “Application of Deep Reinforcement Learning to At-the-Money S&P 500 Options Hedging.”
arxiv.org/abs/2510.09247
We apply a TD3-based DRL agent to hedge S&P 500 options using 20 years of intraday data. The agent outperforms Black–Scholes delta-hedging.

A new post on substack about a strategy I've been testing recently - a Dual Momentum strategy:

triplesun.substack.com/p/strategy-d...

First of the series, I'll be playing around and explaining some concepts in the upcoming follow-ups.

Alternative Loss Function in Evaluation of Transformer Models The proper design and architecture of testing of machine learning models, especially in their application to quantitative finance problems, is crucial. The most important in this process is selecting ...

Our latest paper is now on arXiv: arxiv.org/abs/2507.16548
We compare Transformer and LSTM models for forecasting financial returns and building algorithmic trading strategies.

Here's my latest substack piece: triplesun.substack.com/p/classic-ec...

Generalized Mean Absolute Directional Loss as a Solution to Overfitting and High Transaction Costs in Machine Learning Models Used in High-Frequency Algorithmic Investment Strategies Regardless of the selected asset class and the level of model complexity (Transformer versus LSTM versus Perceptron/RNN), the GMADL loss function produces superior results than standard MSE-type loss ...

Our new paper is out, arxiv.org/abs/2412.18405. In it, we introduce updates to our new loss function designed for algo trading with ML.

New post on our newsletter is out, feel free to have a look!
open.substack.com/pub/triplesu...