Transaction Costs & Slippage

A backtest that ignores the costs of trading is dangerously optimistic, and transaction costs are where many good-looking strategies quietly die. Real trading is not free, and the gap between a strategy's performance before costs and after costs can be the difference between a profitable system and a losing one. This lesson examines the costs of trading and why accounting for them realistically is essential to honest backtesting.

The costs of trading

• Commissions and fees are the explicit charges for executing a trade. Although they have fallen substantially over time, they still matter, especially for strategies that trade frequently.
• The bid-ask spread, from Unit 1, is a cost on every round trip, because you typically buy at the higher ask price and sell at the lower bid price, losing the spread each time you enter and exit.
• Slippage is the difference between the price you expected when you decided to trade and the price at which the trade actually executes, since the market can move in the interval between decision and execution.
• Market impact is the effect of your own order on the price, because a large order consumes available liquidity and pushes the price against you, so you cannot always trade significant size at the quoted price.

Why costs are so consequential

Transaction costs matter because they directly erode the thin edges that real strategies depend on. As the signals lesson stressed, genuine edges are usually small, often just a slight tilt in the odds, and such a small edge can easily be entirely consumed by the costs of trading. A strategy that appears clearly profitable before costs can become unprofitable once realistic commissions, spreads, slippage, and market impact are subtracted. The costs are not a minor adjustment to be waved away, they can flip the entire conclusion about whether a strategy works.

The tyranny of turnover

The damage from transaction costs scales with how often a strategy trades, a relationship sometimes called the tyranny of turnover. Every trade incurs costs, so a high-frequency strategy that trades constantly pays those costs over and over, and the cumulative drag can be enormous. A strategy that trades rapidly must have an edge per trade large enough to exceed the costs of that trade, a demanding requirement that many fast strategies fail to meet once costs are honestly counted. This is why low-turnover strategies are often more robust to costs, and why the appealing-looking results of high-frequency backtests so often collapse when realistic costs are applied. The more a strategy trades, the higher the bar its raw edge must clear.

Transaction costs are where great-looking backtests go to die. A thin edge measured before costs can vanish entirely once the real bill of trading comes due.

The danger of ignoring costs

A backtest run without realistic transaction costs is not just slightly optimistic. It can be profoundly misleading, presenting a losing strategy as a winning one. This is one of the most common reasons that backtested strategies fail in live trading: the backtest assumed costless or cheap execution, while reality imposed spreads, slippage, and market impact that the strategy's thin edge could not survive. Including conservative, realistic estimates of all the relevant costs is therefore not an optional extra but a basic requirement of any trustworthy backtest. A strategy must be profitable after realistic costs and slippage, not merely before them, and any analysis that omits these costs has answered the wrong question entirely. This connects directly to the broader theme that a backtest must reflect the real conditions of live trading as faithfully as possible, a theme that culminates in the lessons on going from backtest to live.