Paper trading is the practice of running a strategy with simulated orders while observing current market prices.

Paper trading is the practice of running a strategy with simulated orders while observing current market prices. Its purpose is operational validation. It lets a team check whether the strategy produces the expected forecasts, applies portfolio constraints, follows its scheduled rebalance logic, and triggers risk controls when live prices evolve in real time.

This differs from backtesting. In a backtest, the strategy is evaluated on historical data. In paper trading, the market data are live, but the orders are not actually sent for execution. That distinction matters because paper trading can confirm that the decision process works as intended under current conditions, while leaving execution outcomes untested.

For a quantitative trading system, the useful question in paper trading is not whether the strategy would have traded profitably. The useful question is whether the system behaves correctly as a process. A sound paper-trading run should therefore produce detailed logs. Those logs should show when forecasts were updated, when constraints became binding, when scheduled rebalances were triggered, what orders the system would have generated, and when risk limits raised alerts or blocked actions. This kind of record helps verify that the strategy logic is operating correctly against live prices.

Paper trading is especially useful for checking implementation details that can fail even when a research idea is sound. Forecast production may be delayed or missing. Constraint logic may not bind when expected. Rebalance scheduling may drift from the intended clock. Risk limits may fail to trigger or may trigger too often. Running the full process in a simulated mode against live prices exposes these operational issues before live deployment.

What paper trading does not test is execution quality. Because the orders are simulated, paper trading does not establish what would happen in the market once an order interacts with liquidity. It does not measure fill rates, slippage, or realized order latency. Those outcomes depend on execution conditions and market microstructure, which are not resolved by a process-only simulation.

That limitation is consistent with a broader principle in quantitative research: the mechanism being tested must match the claim being made. Sonar’s research materials distinguish between model evaluation concepts and warn against overstating what a given test can prove. A process test can validate operational logic. It cannot by itself validate execution-specific outcomes.

For that reason, paper trading and execution analysis answer different questions. Paper-trading logs can show that forecasts, constraints, rebalance timing, and risk controls are functioning correctly against live prices. A separate analysis of live-trading execution data is required to study fill rates, slippage, and latency. Without that second layer, execution-specific conclusions are not supported.

In practice, paper trading is best treated as a rehearsal of the trading workflow. It is a check that the strategy responds to live prices in the intended way and that the surrounding control systems behave correctly. It is not a substitute for evidence on how orders are actually filled in live markets.

Covered in depth in the Strategy research fundamentals pillar hub.