01Research systemsActive research + live options

Ashlock Trading

Systematic trading research platform spanning live options, intraday crypto microstructure, and statistically rigorous strategy validation.

← Back to projects

Ashlock is a three-strategy trading research stack: separate capital per strategy, live options, order-book research, and a shared validation layer that keeps the numbers honest.

System proof

Validation proof

The codebase already shows a serious research boundary rather than a pile of ad hoc notebooks. Ashlock separates strategy capital and state, exposes a unified portfolio-status entrypoint, and ships a shared validation layer for bootstrap Sharpe intervals, Monte Carlo permutation tests, and walk-forward consistency. The headline is not that the strategies are finished; it is that the repo is built to make weak evidence harder to hide.

Strategy model

Three independent strategies with separate capital, risk, and position state

Live boundary

Daily options are in production while intraday perps remain an explicit paper-trading research line

Validation stack

Bootstrap Sharpe CI, Monte Carlo path significance, and walk-forward consistency checks in shared tooling

Unified reporting

`shared.portfolio_status` and `shared.strategy_report` summarize books through one interface

Research stance

The repo explicitly warns against trusting point estimates or post-hoc bucket filtering without rerunning the backtest correctly

Ashlock Trading

What it is

Ashlock matters here less because it is finance-specific and more because it shows a style of engineering I care about: disciplined experimentation, explicit uncertainty handling, and systems designed to survive contact with noisy real-world data.

The repo is structured around three independent strategies: intraday Hyperliquid L2 research, a not-yet-built daily swing portfolio, and a live daily options book. They share infrastructure, but they do not share capital or position state.

Ashlock Trading

Why the architecture matters

The repo is opinionated about separation. Different alpha sources, different volatility regimes, and different blow-up modes should not be flattened into one portfolio just because they live in one codebase.

That leads to a cleaner operating model: separate capital pools, separate circuit breakers, shared dashboards and validation utilities, and one command to inspect the whole stack without pretending every strategy has the same risk profile.

/Intraday perps L2 research in paper-trading mode
/Daily options strategy already in production
/Shared portfolio status and reporting utilities
/Per-strategy caps rather than one blended risk bucket

Ashlock Trading

What makes the research honest

The strongest part of the repo is the validation layer. Instead of quoting a single Sharpe or profit factor and moving on, the shared tooling runs bootstrap confidence intervals, Monte Carlo permutation tests, and walk-forward consistency checks before treating a result as meaningful.

That mindset translates well beyond trading. In any system where the sample is small, path dependence is real, and the environment shifts, point estimates are dangerously easy to over-believe.