Transaction costs and slippage: the line that turns a paper edge into a real one

Most backtests that look brilliant die the moment they meet reality, and the usual killer is not a wrong signal — it is the cost of acting on the signal. Every trade pays a spread, a commission, and a little extra because your own order moved the price. A strategy that trades often pays that toll often, and a thin edge that survives on paper can be entirely consumed by it.

The three costs, named

The spread is the gap between the price you can buy at and the price you can sell at; crossing it is a cost even before any fee. The commission is the explicit fee per order or per unit. Slippage is the difference between the price your backtest assumed and the price you actually got, because the market moved or your size ate into the available orders. The first two are knowable; the third is the one optimists forget.

Model the cost you would really pay, not the advertised one

Subtract a realistic per-trade cost from every simulated fill — not the best-case promotional commission, but the spread you would actually cross plus a slippage allowance that grows with how big and how fast you trade. A strategy that trades a hundred times a month needs a far more honest cost model than one that trades twice a year, because the toll is paid a hundred times. If your edge per trade is smaller than your cost per trade, no amount of winning direction saves you.

Why turnover is the hidden variable

Two strategies with the same gross signal can have opposite net results purely because one trades ten times more. That is why cost modelling and the overfitting question are linked: a hyperactive strategy is often both more overfit (see degrees of freedom) and more cost-sensitive. Stress-test the result by doubling your assumed cost — if the edge vanishes, it was never much of an edge, and the validation checklist is where that stress test belongs.