PyBacktester

An event-based backtester written in Python. The backtester is built with a consumer/producer pattern where the consumer is the Controller class and the producer is the DataSource class. The trading algorithm is contained in the Algorithm class.

Custom Portfolio

A custom portfolio may be created by instantiating the Portfolio class and adding stocks, prices, and shares. This is demonstrated in the many of the unit tests in the tests.py file.

Algorithmic Trading Interface

The "Algorithm" class defines a strategy that the backtester uses. To make your own, you must implement a "generate_orders" method. An example algorithm is included in the base class.

Alternative DataSource

The basic DataSource included is built on top of pandas DataReader. This source may be modified to be any realtime data feed. The DataSource's single requirement is to fill a Queue class with data from the feed. The data should be in the form of a tuple (Timestamp/Id, Ticker str, Price float).

Run

Enter "python backtester.py" for a default run. Output is logged to run.log.

Disclaimer

This code is meant for educational purposes only. The included algorithm should NOT be used to inform real trades in any way. In fact, it randomly liquidates positions.