One engineer. 433 strategies. Documenting the journey to a systematic trading income.
QuantMechanica is a one-person project, run by an engineer based in Europe with a background in engineering — not from a trading floor or a hedge fund. I've been fascinated by financial markets for years, but this project isn't about investing.
For long-term wealth building, index funds are simply the best option — I'm not trying to beat the market with stock picks. The goal here is different: building an additional income stream through systematic Forex trading.
Prop firms are ideally suited for this. They let you trade with their capital, and on success, you earn a disproportionately large return relative to your own investment. If the strategies work on prop firm accounts, they'll work on real CFD accounts too — generating additional income with minimal capital at risk. And over time, that compounds.
A project of this scale — 433 strategies, 30,000+ backtests, a 10-phase validation pipeline — would be impossible for a single person without AI. I use Claude as my assistant across the entire workflow: researching academic papers and trading books, developing the MQL5 and Python code, orchestrating and managing the automated backtests, and running the subsequent statistical analyses. AI doesn't replace the engineering thinking and decision-making, but it multiplies what one person can accomplish by orders of magnitude. That's what made this project possible in the first place.
I'm using every scientific method available — statistical validation, walk-forward analysis, Monte Carlo simulations, stress testing — combined with AI-assisted research and development to get there. This site documents the entire journey: what works, what fails, and why.
I approach financial markets the same way I approach any engineering problem: with structured methodology, automated testing, and zero tolerance for unverified claims.
Systems thinking, process design, and quality control — applied to financial markets instead of production lines. Every strategy goes through a rigorous 10-phase pipeline before it touches real capital.
Walk-forward analysis, Monte Carlo simulations, Deflated Sharpe Ratio, False Discovery Rate correction, multi-seed overfitting detection, and HARSH stress testing with extreme market conditions.
MQL5, Python, and Claude AI. Automated backtesting across 23 symbols, parameter sweep engines, commission validation, and a pipeline that runs 30,000+ tests while I sleep.
Every result is published — including the 92% of strategies that failed. The full Strategy Archive, equity curves, commission reports, and stress test outcomes are open for anyone to review.
I publish every result — wins and failures. If a strategy dies at Phase 6 HARSH stress testing, you see exactly why.
I don't predict markets. I build systems, test them rigorously through 10 phases, and deploy only what survives.
Index funds beat most active investors. This isn't about beating the market — it's about building an additional income stream through systematic edge exploitation.
433 strategies, 9 years of tick data, 23 symbols, real commissions, HARSH stress tests. No shortcuts, no promises — just data.
81+ edge types, family caps, symbol caps, and correlation analysis. The portfolio is built to survive — no single strategy or market can take it down.
Strategies don't get a lifetime pass. Forward tests, live monitoring, and periodic re-validation ensure that what worked in backtests keeps working in reality.
Every strategy goes through a 10-phase pipeline. Every result is published — including failures. Claims are backed by data, assumptions are stated, and nothing is deployed without surviving stress tests. No exceptions.