Slippage is the difference between the price you expect and the price your order receives. In quieter August markets, thinner liquidity can widen that difference, turning a strategy that looks profitable on paper into one that loses money after execution.
On August 1, 2012, Knight Capital began sending erroneous orders into US equity markets after a faulty software deployment. The outcome was still unfolding while the orders accumulated. Knight could not assume that displayed prices would remain available long enough to exit cleanly.
Within about 45 minutes, the company had lost roughly $440 million. The US Securities and Exchange Commission documented the incident in its 2013 administrative order against Knight Capital Americas. The failure was far larger than ordinary retail slippage, but the mechanism carries a useful warning: intended prices and executable prices are different facts.
For a trader in August, that gap rarely arrives as one dramatic number. It appears as a few cents on entry, another few cents on exit, and a backtest that quietly assumes every order fills where the signal appeared.
The cost hidden between signal and fill
Suppose a strategy buys 100 shares when the quoted price reaches $20.00. The order fills at $20.04 because only a limited quantity was available near the quote. The position later receives an exit signal at $20.20, but the sell fills at $20.16.
The chart shows a 20-cent move. The account captures 12 cents before commissions and fees:
- Expected gross profit: $20
- Entry slippage: $4
- Exit slippage: $4
- Realized gross profit before other costs: $12
Slippage consumed 40 percent of the expected profit.
These figures are illustrations, not a claim about typical execution. The actual result depends on order type, order size, available liquidity, volatility, venue, and how quickly the market moves. That uncertainty is precisely why a strategy should model execution costs rather than treat them as a rounding error.
The effect becomes harder to notice across repeated trades. Ten trades that each lose $8 to combined entry and exit slippage create an $80 execution cost. A strategy expecting $100 of gross profit has little room left for spreads, commissions, or losing trades.
Why quiet August markets can mislead
Lower activity can produce calm-looking charts without guaranteeing deep order books. A quoted price may represent fewer available shares or tokens than your intended order requires. Once that displayed quantity is consumed, the rest of the order can fill at progressively worse prices.
Market orders make this trade-off explicit. They prioritize execution, while the final price remains uncertain. Limit orders cap the acceptable price, but they introduce another uncertainty: some or all of the order may remain unfilled.
Automated strategies can repeat the problem faster. A backtest may record a buy at the signal price, assume the full position filled, and subtract no spread or slippage. Live trading then encounters three differences at once: the signal took time to reach the broker, the quoted price moved, and the desired quantity was unavailable at one level.
That does not make automation inherently defective. It means the execution assumptions deserve the same scrutiny as the entry rule.
A useful starting point is to rerun a backtest with several cost assumptions. If a strategy works with zero slippage, weakens at two basis points, and fails at five, its apparent edge is fragile. Pair that test with max drawdown analysis, as explained in this beginner’s guide to testing survival, not just returns.
Measure the execution you actually received
Record four values for every trade: the signal price, submitted price, average fill price, and filled quantity. Then calculate slippage separately for entries and exits.
For a buy:
`slippage = average fill price − decision price`
For a sell:
`slippage = decision price − average fill price`
Multiply the result by the filled quantity to express the cost in account currency. Keeping the raw price difference matters too, because a $5 cost on a $500 position has a different meaning from the same cost on a $50,000 position.
Review the data by instrument, time of day, order type, and position size. Averages alone can conceal the trades that matter most. The median may show normal execution, while a small group of poor fills erases the month’s profit.
Spread also deserves its own field. Slippage and spread are related, but they measure different costs. The spread exists before the order arrives. Slippage measures how execution differs from the chosen reference price. A 12-cent spread can raise planned risk materially before additional slippage enters the calculation.
Put execution inside the approval decision
An approval gate creates a pause between an AI-generated signal and a real order. Use that pause to inspect the conditions the signal cannot control: the current spread, available quantity near the quote, expected slippage, and whether the position still fits the planned loss if the fill is worse.
Consider a planned entry at $50 with a stop at $49.50. The intended risk is 50 cents per share. If the actual entry becomes $50.10 while the stop remains unchanged, the risk rises to 60 cents per share, a 20 percent increase. The position size should change, or the trade should be rejected.
That is the practical lesson from Knight Capital at retail scale. Execution deserves its own control because an order can behave differently from the instruction that created it. Before approving an August trade, write down the worst acceptable fill. If the market cannot provide it, leave the order unfilled.
Educational content, not financial advice.
Comments
No comments yet.