Market impact is the price movement caused by an order itself.
Market impact is the price movement caused by an order itself. When a strategy tries to buy, its demand can push the execution price upward. When it tries to sell, its supply can push the execution price downward. This creates an execution cost that is separate from the signal, separate from fees, and separate from the spread.
The key property of market impact is that it increases with order size. Small trades can often be absorbed with limited disturbance. Larger trades consume more available liquidity and force execution into less favorable prices. That means a strategy can look clean in a small backtest or a light simulation while becoming materially more expensive once traded at real size.
Cross venue data is useful because it helps separate general market movement from the effect of a specific execution. Sonar Sciences describes cross venue data as synchronized information across exchanges and venues, used to study how prices and liquidity interact across fragmented markets. In that setting, an execution can be compared against contemporaneous conditions across venues rather than against a single print on one venue. That makes it easier to identify the part of slippage associated with the order interacting with available liquidity, instead of confusing it with ordinary market drift or isolated venue noise.
This matters because backtests usually model the world at a level of detail that is too coarse to capture impact. A small sample backtest often assumes fills at observed prices, or applies a simple cost estimate that does not scale correctly with size. If the test only uses small hypothetical orders, the strategy may appear robust even though the same logic would face much higher costs in live execution. The missing component is not the alpha model. It is the price concession required to complete the trade.
Sonar Sciences' backtest overfitting audit is designed to test whether a backtest result is likely to be overstated by repeated selection and multiple testing. That is a different problem from market impact. The audit addresses statistical inflation in reported performance from trying many variants, while market impact is an execution effect that arises when orders meet real liquidity. A strategy can pass an overfitting audit and still face execution costs that were not modeled. In that sense, overfitting diagnostics and impact analysis are complementary rather than interchangeable.
The deflated Sharpe ratio glossary reinforces this distinction. It explains a method for adjusting a Sharpe ratio for selection bias and non normality so that apparent performance is judged against the number of trials and the distributional properties of returns. That helps evaluate whether an observed backtest edge is statistically credible. It does not by itself estimate the cost of moving the market during execution. A backtest can therefore look statistically disciplined and still omit a size dependent implementation cost.
Taken together, these sources support a practical lesson for systematic traders. Market impact is an implementation cost created by the order itself. It tends to rise with order size because larger orders must consume more liquidity. Cross venue analysis is a better framework for isolating that effect than a single venue view. And statistical checks for backtest overfitting, including tools related to the deflated Sharpe ratio, do not substitute for explicit modeling of execution costs. If impact is absent from testing, small sample backtests can understate the true cost of trading at scale.
Covered in depth in the Cross-venue market data & signals pillar hub.