Skip to content

Introduction

Trading starts with an idea: perhaps momentum persists, prices revert after an extreme move, or risk can be controlled by changing position size. The difficult part is not describing the idea; it is turning it into explicit rules and finding out where those rules fail.

Backtide is a local-first research workspace for that process. It helps you collect market data, express a strategy, test it against history, inspect the trades and risk, and then observe the same logic with current market data through simulated live simulation. The application and Python API use the same saved strategies, indicators, metrics, sizers, configuration, and results.

From an idea to evidence

A useful research loop has five stages:

  1. State the hypothesis. Define the market behavior you expect and why it might persist.
  2. Make the rules precise. Choose inputs, entry and exit conditions, position sizing, fees, slippage, and risk limits before looking at the result.
  3. Backtest. Run the rules over historical data and inspect the full result—not only the final return, but also drawdown, rejected orders, individual trades, and unstable periods.
  4. Challenge the result. Try other instruments and time windows, vary assumptions, and keep unseen data for confirmation. A good-looking chart is a reason to investigate, not proof.
  5. Run live. Observe the strategy on current market updates with simulated fills before deciding whether the idea deserves any real-world use.

Backtide makes this loop reproducible: experiments retain their configuration and outputs, while custom library objects let you reuse the same definitions in later tests.

Choose a workflow

  • Start with Backtest introduction to learn what historical testing can and cannot tell you.
  • Use Experiments to configure and run a complete test.
  • Read Results and Plots to inspect returned objects and visual evidence.
  • Continue with Live introduction and Live simulation when you want to observe a strategy against current exchange data.
  • Build reusable logic in the library: strategies decide, indicators transform data, sizers choose quantities, and metrics summarize finished runs.

Warning

Backtide is a research and education tool. Backtests and live simulation simplify execution and cannot predict future returns or guarantee that a real order would fill at the simulated price.