A Digital Hub Insights Publication
The Signal
Reimagining finance from first principles
The tools have changed. The questions haven't. We go back to fundamentals and ask — now that AI is in the workflow, what's actually possible that wasn't before?
166%
XGBoost Return
| 45K+
Practitioners
| PhD
First Principles
|
Our Thesis
Every generation of tools opens doors that were previously invisible. AI doesn't just optimize the old way of doing finance — it makes entirely new approaches possible.
For decades, quantitative finance has operated within well-established frameworks — frameworks built when computational power was scarce and data was expensive. Those constraints shaped everything: how we construct portfolios, how we validate models, how we identify risk.
The constraints have changed. But most of the approaches haven't.
The Signal exists to explore what happens when a practitioner goes back to first principles — not to promote AI, but to ask better questions. Each issue takes a real finance problem, strips away assumptions, and rebuilds with the tools now available. Including what didn't work. Especially what didn't work.
Use Cases
One problem. Rethought from scratch.
Each issue follows a consistent method: take a well-established finance problem, question the assumptions, rebuild with modern tools, and report honestly on the results.
● Live
Momentum Strategy
Can ML models identify momentum signals that traditional cross-sectional approaches miss? XGBoost delivered 166% over five years vs. SPY's 90%. Walk-forward validated. No lookahead bias.
| ◐ Next
Mean Reversion
Do overreactions create predictable opportunities? Rethinking classic contrarian strategies with ML-driven regime awareness and adaptive thresholds.
|
○ Upcoming
Factor Timing
Can you predict which factors will outperform next month? Exploring ML-driven factor rotation with honest assessment of what's signal and what's noise.
| ○ Upcoming
Pairs Trading
Statistical arbitrage reimagined. Using ML to identify cointegrated pairs dynamically rather than relying on static historical relationships.
|
What The Signal is
→ First-principles research with real data
→ Practitioner to practitioner
→ Honest results — including failures
→ Use cases that rethink, not recycle
→ AI as enabler, not headline
| What it isn't
✕ A newsletter
✕ Market commentary or hot takes
✕ Cherry-picked backtests
✕ Another "AI will revolutionize" pitch
✕ Surface-level content
|
Current Issue
Momentum, Reimagined
Traditional momentum strategies rank stocks by past returns and hope the trend continues. We asked: what if an ML model could learn which momentum signals actually persist — and which ones are noise?
Our XGBoost momentum strategy delivered 166% total return over five years, nearly doubling SPY's 90%. Validated with walk-forward optimization, temporal cross-validation, and zero lookahead bias.
The full breakdown — methodology, code, performance attribution, and what we tried that didn't work — is available free to subscribers.
Built For
Practitioners who think in first principles
If you work in quantitative finance, portfolio management, risk, or data science in finance — and you're tired of surface-level "AI will change everything" content — The Signal was built for you.
We don't assume AI is the answer to every problem. We assume every problem deserves to be re-examined with the tools now available. Sometimes AI transforms the approach. Sometimes it doesn't. We report both.
Get The Signal
New use cases published monthly. The momentum strategy breakdown — including methodology, code, and full results — is free for subscribers.
Subscribe to The Signal
Get the current issue free — including full methodology and code.
We respect your privacy. Unsubscribe at any time.
The Signal is a publication of Digital Hub Insights LLC
Questions? support@cybersafeid.com