Most trading strategies don’t fail because they’re wrong.

They fail because they’re fragile.

This newsletter is about building robust systematic trading strategies. Not shortcuts, not overfitted backtests, and not recycled trading tips. The focus here is on structure, research, and what can actually survive out-of-sample.

WHAT YOU’LL FIND

You’ll find practical research and clear thinking on:

• strategy design

• robustness and overfitting

• feature engineering and target construction

• realistic backtesting

• machine learning applied to trading, with a strong focus on generalization

ABOUT

I’m Lucas, an independent quantitative researcher focused on systematic trading.

My work centers on extracting signal from noisy financial data, building robust feature and target pipelines, and designing strategies that can move from research to live trading.

Over the years, I’ve worked across the full process: data processing, model design, backtesting, validation, and execution.

CURRENT RESEARCH

My current research focuses in particular on three areas:

• Feature selection under noisy market conditions.

• Fast and scalable feature computation.

• Robust feature and target engineering for financial machine learning.

INFRASTRUCTURE

To support this work, I’ve also developed tools to accelerate research workflows, standardize feature engineering, and improve consistency between research and live trading.

Oryon Python library: https://oryonlib.dev

FINAL THOUGHT

“Most strategies don’t survive. The goal here is to build ones that do.”