User Guide

Use this guide to build a backtest from prepared market data and strategy decisions, then inspect what the engine actually executed. The quickstart gives you a complete first run. The topic pages below explain the choices you make when adapting that run to your own data and strategy.

The code in the Getting Started tutorials runs from the installed package against bundled synthetic data. The topic pages describe individual interfaces and settings. When a topic page shows only a method call, use its linked tutorial for a complete program and expected output. The API reference gives the signatures and source docstrings for the released package.

Find the task you need

Task	Guide and runnable example	API contract and book notebook
Prepare bars, signals, context, and timestamps	Data Feed Run: Bundled example data	DataFeed Book: Futures feed
Write decisions and carry state between callbacks	Strategies and Stateful Strategies Run: First backtest and Risk and State	Strategy.on_data Book: Stateful strategies
Choose order types and determine when fills can occur	Order Types and Execution Semantics Run: Orders and Timing	Broker.submit_order Book: Single asset backtest
Allocate across assets and deal with different daily close times	Rebalancing Run: Multi-asset Rebalancing	Broker.rebalance_to_weights Book: Allocator comparison
Set cash, margin, contract, and sizing rules	Account Policies and Configuration Run: Accounts and Constraints	BacktestConfig Book: Conformal sizing
Model commission, slippage, impact, and funding	Market Impact and Execution Costs Run: Costs and Funding	BacktestConfig Book: Gross versus net
Add position exits and portfolio limits	Risk Management Run: Risk and State	StopLoss Book: Library risk demo
Compare explicit execution settings	Profiles Run: Profiles and Parity	BacktestConfig.from_preset Book: Engine divergence
Export fills, trades, equity, and portfolio state	Results and Analysis Run: Results and Analysis	BacktestResult Book: Performance reporting
Move an existing Zipline strategy to prepared bars and explicit settings	Migrate from Zipline Run: Complete migration example	Engine.run Book: Engine divergence

Choose an asset example

All five bundled panels are synthetic. They test the library workflow offline; they do not estimate a strategy's live performance or reproduce a book case study. The data page runs a round trip for each panel and states its price and volume assumptions.

Asset setting	Example	Setting to check before using your own data
Equity	Equity round trip	Share precision, trading calendar, and corporate-action handling in the input
ETF	ETF round trip	Weight targets, rebalance timing, and cash available for the next fill
Futures	Futures round trip	Contract multiplier, margin, and the timestamp of each market close
FX	FX round trip	Pair quotation and the account currency used for valuation
Crypto perpetual	Perpetual round trip	Funding timestamps, rate sign, and the position held before each event

The examples use declared settings rather than implied market defaults. Start with the configuration guide when you replace a bundled panel with another instrument or venue.

From a book notebook to a library run

The Book Guide maps checked companion notebooks to the matching library workflows. Begin with the library quickstart, then compare the notebook's data, decision time, fill time, account rules, and costs with the settings in your run. Keep the original notebook's research question separate from the engine configuration. The book links point to a pinned companion revision, and each linked tutorial states what its notebook adds.

Check a result

Inspect fills, trades, and equity before interpreting summary metrics. A missing fill can mean that an order was rejected, an order type did not trigger, the next eligible bar was absent, or buying power was insufficient. The result tutorial exports these records and joins event timestamps back to the input data. The diagnostic handoff shows the optional analysis package used for downstream metrics.