Periscøpe
Request access
Docs menu Execution Modes

Execution Modes

Backtest, Paper, and Live: what changes, what stays the same, and when to promote.

An execution mode is a setting on a run, not on a strategy. The same strategy version can run in all three modes. Only the data feed and the broker change.

Backtest

Fast historical simulation for strategy development. Market data is replayed from history and fills are simulated. You set a start and end time on the run's environment config, and the simulation steps through that window as fast as it can.

Use Backtest to develop and validate strategy logic, sweep parameters, and build intuition. There is no capital at risk and no real orders go anywhere.

Paper

Real time simulated trading with no capital at risk. The data feed is live, but orders are filled by a simulator rather than sent to a broker.

Use Paper to rehearse a strategy under real market conditions: verify it handles live data shapes, doesn't misbehave across session boundaries, and produces the kind of activity you expect.

Live

Production execution against a real broker account. Real market data, real orders, real capital.

What changes between modes

  • Data feed. Backtest uses historical data for the run's time window. Paper and Live use the real time feed.
  • Broker. Backtest and Paper both use simulated fills. Live sends orders to a real broker account.
  • Time. Backtest steps through simulated time; Paper and Live run on real time.
  • Environment config. Backtest takes a start and end time. Paper and Live do not.
  • Reset. Available for Backtest and Paper runs that have started. Not available for Live.
  • Kill copy. The confirmation dialog is stricter in Live and reminds you that open positions are not auto closed.

What stays the same

  • The strategy code. The same strategy.py runs in every mode.
  • The strategy parameters. They work the same way across modes.
  • The SDK surface. Every event, action, and enum behaves identically in every mode. self.simulation_time tracks historical time in Backtest and real time in Paper and Live.
  • The run surface. Status lifecycle, tabs, metrics, and issues look the same no matter the mode.

Promotion guidance

A common path.

  1. Develop in Backtest. Iterate fast, sweep parameters, confirm the strategy produces the trade cadence you expect and that P&L looks sensible. Do not skip this step.
  2. Rehearse in Paper. Run for a session or more against the live feed. Watch for live only issues: market open and close behavior, session gaps, symbol availability, order rejections that never happened in backtest.
  3. Promote to Live. Only after Paper looks clean. Start small. Treat the first Live run as an extension of Paper, not a validation of the strategy itself.