Paper trading is the practice of executing a trading strategy in a simulated environment that uses live market data but does not involve real money. It serves as an intermediate validation step between backtesting on historical data and deploying a strategy with real capital. While backtesting answers the question "would this have worked in the past?", paper trading answers "does this work right now, in real time?"
Why paper trading matters
Backtesting, no matter how rigorous, has limitations. It relies on historical data and simulated execution. Paper trading exposes a strategy to the realities of live markets: real-time data feeds, actual latency, genuine order book dynamics, and current market conditions. Issues that do not surface in backtesting, such as data feed interruptions, unexpected latency spikes, or strategy bugs that only manifest in real-time processing, become apparent during paper trading.
Paper trading also validates the operational infrastructure. It tests the full pipeline from data ingestion to signal generation to order submission, ensuring that all components work together correctly before real capital is at risk.
Paper trading versus backtesting
Backtesting provides statistical validation over a long historical period but assumes perfect infrastructure. Paper trading provides operational validation in real-time but over a limited time window. Both are necessary. A strategy that passes backtesting but fails in paper trading likely has implementation issues. A strategy that works in paper trading but was never backtested might be succeeding due to current favorable conditions that will not persist.
Limitations of paper trading
Paper trading does not fully replicate live execution. Simulated orders do not affect the market, so there is no market impact. In live trading, a large order can move the price, but in paper trading, the price is unaffected. This means paper trading results tend to be slightly more optimistic than live results for strategies that trade significant volume.
Paper trading also runs for a limited time, typically weeks to months. This window may not include important market events like earnings seasons, economic crises, or volatility regime changes. A strategy that performs well during a calm paper trading period might struggle during a subsequent volatile period.
Best practices
Run paper trading for long enough to capture a meaningful number of trades and ideally a range of market conditions. Track the same performance metrics used in backtesting (Sharpe ratio, drawdown, win rate) and compare them to backtest expectations. Significant discrepancies between backtested and paper-traded performance warrant investigation before going live.
Practical example
A trader backtests a momentum strategy over five years with a Sharpe ratio of 1.8 and a maximum drawdown of 12%. Before deploying capital, the trader runs the strategy in paper trading for three months. The paper trading Sharpe ratio is 1.5 and maximum drawdown is 9%. The slight performance degradation is expected due to realistic execution timing, and the results are close enough to backtest expectations to justify moving to live trading with a small initial allocation.
How Tektii helps
Tektii supports paper trading through its live market data connections and proxy service. Strategies can connect to real-time market feeds and generate orders that are simulated rather than routed to exchanges. Because Tektii strategies are Docker containers that communicate via WebSocket, the same strategy code used in backtesting runs unchanged in paper trading and eventually in live trading. This eliminates code translation errors that can occur when moving between backtesting and live platforms.