Trading costs are everything that separates a gross backtest result from a net one: the bid-ask spread you cross, the slippage between your intended and actual fill, and commissions or fees per trade.
Target audience: Traders validating a system who need to know whether the edge survives real-world costs.
Trading costs are everything that separates a gross backtest result from a net one: the bid-ask spread you cross, the slippage between your intended and actual fill, and commissions or fees per trade.
Costs are subtracted from every trade, so they fall hardest on high-frequency, small-edge systems. A strategy that looks profitable gross can be a loser net, and the only number that matters is net. Modelling costs honestly is what keeps a backtest from promising an edge that the broker quietly takes back.
The spread is the gap between bid and ask; a market order crosses it, paying half on entry and half on exit in effect. Slippage is the difference between the price you intended and the price you got, worst in fast or thin markets and on stop orders. Commission and exchange fees are a fixed charge per contract or share. Together they form a cost that is paid on every round trip, win or lose, and that must be subtracted from the gross result.
To see whether costs matter, put them in the same unit as your edge: R. If your average risk per trade is 10 ticks and your round-trip cost (spread plus slippage plus fees) is 1 tick, costs are about 0.1R per trade. A system with +0.3R gross expectancy keeps +0.2R net, still good. A system with +0.08R gross expectancy is a net loser. The smaller your edge and the wider your costs relative to your stop, the more costs decide the outcome.
Cost is paid per trade, so a system that trades ten times a day pays ten times the daily cost of one that trades once. High-frequency, small-target systems are the most cost-sensitive: their edge per trade is tiny, so a fraction of an R in cost can erase it. Lower-frequency, larger-R systems are more forgiving. When in doubt, model costs on the pessimistic side; live slippage in the trades you most want to take is usually worse than average.
A scalping system shows +0.12R gross expectancy over 5,000 trades, a beautiful curve. Round-trip cost is estimated at 0.15R (tight stops make costs a large share of risk). Net expectancy is -0.03R: across 5,000 trades the system loses money, and the gross curve was an illusion created by ignoring the spread and slippage it pays on every one of those trades.
Reporting gross results and never subtracting costs.
Assuming average slippage on exactly the fast trades where slippage is worst.
Ignoring that tight stops make a fixed cost a large fraction of R.
Forgetting that costs are paid on losers too, not just winners.
Underestimating cost on the highest-frequency part of the system.
Myth
That a profitable gross backtest is a profitable system.
Reality
No paired reality note provided.
Myth
That slippage is symmetric and small on the trades you care about.
Reality
No paired reality note provided.
Myth
That commissions are negligible at high frequency.
Reality
No paired reality note provided.
Estimate your round-trip cost in R and subtract it from your gross expectancy.
This lesson is educational content only and is not financial advice. Trading involves substantial risk. A tested process improves decision quality and survivability; it does not predict the market or guarantee any outcome. Trade only with risk you can afford to lose.