The most reliable graveyard in quantitative finance is the gap between a backtest and a live account. Strategies arrive there looking healthy: strong gross returns, clean signals, persuasive statistics. What kills them is rarely the idea. It is the accumulated cost of touching the market: the spread paid on every entry, the slippage on every fill, the impact of the strategy’s own footprint. Paper alpha dies in transit.
The Tax on Turnover
Transaction costs behave like a tax levied on activity. The bid-ask spread is the visible portion, charged on every round trip regardless of whether the trade was wise. Slippage is the difference between the price a model assumed and the price the market actually offered when the order arrived. Market impact is subtler still: the act of trading moves the price against the trader, and it scales with size and urgency.
The arithmetic is unforgiving because the tax compounds with turnover. A strategy that trades often must clear its cost hurdle on every single trade, and a modest per-trade friction, multiplied across thousands of executions, can consume an edge entirely. Gross results describe the quality of a forecast. Net results describe a business. We are only interested in the second number.
Costs Belong Inside the Backtest
A common failure pattern is to run a frictionless backtest, admire the result, then subtract an estimated cost at the end. This ordering is backwards. Costs applied after the fact cannot change what the strategy would have done, yet in reality they should: a signal too weak to clear the spread should never generate an order at all. Cost models therefore live inside our simulations, shaping every decision the strategy makes.
We then stress the assumption. Every candidate is re-evaluated under cost scenarios deliberately harsher than our estimates, including wider spreads, worse fills, and degraded liquidity. A strategy whose economics survive only under optimistic execution assumptions has not passed validation; it has merely postponed its failure to the live environment, where the tuition is paid in real capital.
Execution as a Source of Edge
The reality gap also has a constructive face. If costs can destroy an edge, reducing them can create one. Order placement is a research problem in its own right: whether to cross the spread or rest passively, how to time entries within a session, how instruments behave around opens, closes, and liquidity transitions. Execution engineering earns its return in basis points that arrive with unusual consistency.
Our promotion pipeline reflects this hierarchy. No strategy at Bountify is judged on gross performance at any stage; every backtest, every walk-forward fold, and every paper-trading session is scored net of modelled costs, and paper trading exists largely to test whether those models match reality. The market does not pay for forecasts. It pays for forecasts that survive the journey through the order book.