Live vs Backtest: Slippage, Latency, and Psychology Gaps

A system that returns 30% in backtest often returns 10-15% live. The gap is not bad luck; it is the sum of slippage, latency, and execution psychology that backtests ignore. Quantify each to predict the real number.

Slippage Modeling

Backtests usually assume fills at the signal price. Live fills differ.

• Market orders: assume slippage equal to half the spread plus 1 tick of adverse movement on entry, 1 tick on exit. For ES futures at a 0.25-point spread, that is roughly 1-2 ticks per round turn.
• Limit orders: assume a fill rate, not guaranteed fills. If your backtest assumes every limit fills, but live fill rate is 70%, your trade count and returns drop.
• Stop orders: in fast markets, stops slip 2-5 ticks routinely. Model stop slippage as 2x normal slippage.

Practical rule: subtract 1-2 ticks per round turn from every backtest trade before trusting the equity curve. If the system cannot survive this haircut, it cannot survive live.

Latency

Signal-to-execution latency includes:

• Indicator computation time.
• Order transmission to the broker.
• Broker-to-exchange routing.
• Fill acknowledgment.

For swing trading, 1-2 seconds of latency is irrelevant; for 1-minute systems, 1 second can move price a full tick against you; for scalping, 100ms matters.

Measure your real latency by logging signal time and fill time over 50 live trades. Add the average adverse price movement to your slippage model. If latency cost exceeds 10% of expected per-trade profit, your timeframe is too fast for your infrastructure.

The Psychology Gap

The largest gap is behavioral. Backtests execute every signal; live traders do not.

• Missed signals: fatigue, distraction, or fear cause 10-20% of signals to be skipped. Each missed winner hurts more than each missed loser, because winners are rarer.
• Discretionary overrides: pulling a stop or doubling size after a loss. A single override can erase weeks of system profit.
• Drawdown panic: traders abandon systems during normal drawdowns, missing the recovery that the backtest promised.

Quantify the psychology gap by forward-testing on a small live account. Compare the trades you actually took to the signals the system generated. The difference in returns is your personal execution drag. Most retail traders find 30-50% of the backtest edge disappears in live execution.

Closing the Gap

• Automate execution wherever possible. Automation removes missed signals and emotional overrides.
• Use limit orders with a fallback market order after a timeout to control slippage and fill rate.
• Size down so drawdowns do not trigger panic. If a 10% drawdown makes you interfere, your size is too big.
• Forward-test for 50-100 trades before scaling. The live track record is the only honest predictor of live performance.

The Real Number

Expect live returns of 50-70% of backtest returns after slippage, latency, and behavioral drag. A thin backtest edge turns negative live; only backtests with substantial margin survive the translation to reality.