Skip to main content
Pulse is a Python backtesting framework for options and futures strategies. You write a Strategy, run it over historical data, and read back equity, trades, greeks and metrics.

Where you run it

Pulse runs in your Coder workspace. Pulse and the market data are already set up there, so you can start writing a strategy immediately.

Your first backtest

On Coder this runs as-is. A short put on Brent crude (COM.BRN), rolled monthly:
What each piece does:
  • on_update() runs once per pulse; it places orders, which fill the same pulse (see Architecture and the pulse loop).
  • The resolvers turn “30-delta put, monthly, at least 30 days out” into a concrete strike and expiry (see Resolvers).
  • run_sequentially assembles the environment from local defaults (the data under PDC_PATH, a timeline built from the underlying’s settlement calendar) and runs the loop; the result carries the finished strategy with its stats. It takes a list, so the same call runs a whole sweep.
Set PDC_PATH before import pulseMarket data is read from the directory named by the PDC_PATH environment variable, and the data layer binds it at import time. Setting it inside a running session or notebook kernel is too late. On Coder it is preconfigured.
If a strategy can’t find data, see Gotchas & FAQ.
Where to go next
  • Read results properly: Metrics and Stats, and the charts + StrategyReport.
  • Build other shapes: the Structures recipes (spreads, strangles, condors, …), each with a manual and a helper version.
  • Understand the model: the Concepts section.
  • Working examples: the Notebooks.
Last modified on November 4, 2026