Slippage

Slippage is the difference between an order's decision or arrival price and its average execution price. Adverse slippage raises a buy fill or lowers a sell fill. One direction-aware convention is:

adverse slippage (bps) = side × (fill − decision) / decision × 10,000
side = +1 for a buy, −1 for a sell

A decision to buy at $64,000 that averages $64,019 therefore incurs about +2.97 basis points. A sell filled at $63,981 against the same benchmark also incurs about +2.97 adverse basis points.

Where it comes from

Price can move during latency, and an order larger than the available top level can consume progressively worse prices in the order book. Bybit's own execution documentation describes market orders filling across multiple book levels when the best level does not hold enough size.

The bid-ask spread and slippage are related but not interchangeable. Whether crossing half the spread appears inside “slippage” depends on the benchmark. A model that charges spread separately should define its slippage input consistently to avoid charging the same cost twice.

Why backtests must charge it

A common misconception is that one fixed basis-point setting predicts live slippage. It is a stress assumption. Actual execution varies with order size, liquidity, volatility, latency, order type, and venue. Frequent strategies are especially sensitive because small shortfalls repeat across many fills.

On AlphaProve

AlphaProve exposes market slippage, additional stop slippage, spread, and order latency as separate run settings. Recorded L2 depth can price size-dependent fills when the selected market and period have coverage. Rerunning a strategy across plausible assumptions shows sensitivity; it does not discover the one future cost. The historical comparison in four levels of backtest honesty shows how fees, spread, slippage, funding, and fill rules interacted in one EMA test.