Introduction

What ClearEdge Algo is

ClearEdge Algo is a self-hosted, single-user automated trading system. It runs entirely on your own machine: you bring your own market-data and broker keys, your data and results never leave your box, and you are the only account. It combines four things most retail tools keep separate:

1. A data layer — vendor bars (Tiingo daily, Databento intraday, crypto) ingested into a local Parquet catalog that is the single system of record for every backtest.
2. Strategies as code — a library of built-in strategies, plus a Studio that turns a plain-English idea into a real, sandboxed strategy.
3. A validation gauntlet — a deliberately strict, anti-overfit pipeline (rule test, walk-forward, deflated Sharpe, cost stress, cross-sectional breadth) that tells you, honestly, whether an edge is real or just a lucky fit.
4. Execution — paper and live trading through Interactive Brokers (and a local crypto simulator), with a hard risk engine, a kill switch, reconciliation, and an audited approve-to-live gate.

The engine under the hood is NautilusTrader; the same strategy code runs in a backtest and on a live account, so what you validate is what you trade.

Who it’s for

A technically capable individual — someone comfortable running a couple of services locally — who wants to research systematic trading strategies honestly, without the survivorship bias and curve-fitting that make most retail backtesting worthless. It is a research-first tool that can trade real money, not a black box that promises returns.

The philosophy: honest research

The product’s center of gravity is the validation gauntlet, and its guiding principle is that most apparent edges are noise. ClearEdge is built to disprove your strategies, not to flatter them:

• No look-ahead, by construction. Backtests read only from the local catalog, never a live API, and strategies act on closed bars that fill on the next bar. Runs are reproducible.
• A positive backtest is not enough. Every claimed edge must beat a random-entry twin (the rule test), survive out-of-sample re-optimization (walk-forward), clear a deflated Sharpe that discounts for how many variants were tried, hold up under doubled costs, and be judged against simply buying and holding the same instrument.
• Breadth over cleverness. A real edge shows up on many independent instruments, not two or three lucky ones — so cross-sectional breadth is a load-bearing rung, and the honest finding across the built-in research is that simple edges win and added machinery rarely adds signal.
• The results are recorded faithfully. Refuted ideas are documented as refuted. The strategy library and the research notes say plainly what did and did not work.

This honesty is also the product’s marketing position: it never makes performance claims; it sells the discipline.

What you can do with it

• Ingest daily and intraday market data for equities/ETFs and crypto.
• Run single backtests with full metrics, equity/drawdown charts, round-trip trade tables, and a buy-and-hold benchmark.
• Put a strategy through the full anti-overfit gauntlet and get a 0–100 readiness score.
• Combine validated strategies into portfolios and books, and stress their correlations.
• Describe a strategy in plain English in the Studio, then iterate on it with a research copilot that reads your backtest results.
• Promote what survives to paper trading, and — behind an explicit, audited gate — to live.
• Learn the methodology through an in-app, structured Learn course.

Next

• Architecture — how the pieces fit together.
• Concepts & glossary — the vocabulary used throughout these docs.
• Getting started — install and run it.