# Execution Cost Modeling Also known as: slippage, spread, partial fills, requotes/latency. A Risk, Sizing & Exits concept (Orders & execution) in the LuxAlgo Library, with 1 indicator implementation. ## What is Execution Cost Modeling? Execution cost modeling is the practice of estimating the gap between the prices a strategy assumes and the prices real orders achieve. The gap has several components: the bid-ask spread paid when crossing the book, slippage between the decision price and the fill price, commissions and exchange fees, partial fills that leave intended size unexecuted, and, in dealer-quoted markets, requotes and latency effects where the quoted price is gone by the time the order arrives. Together these are the frictions that separate a backtest from a brokerage statement. The discipline exists because untreated costs are one of the most common ways a profitable-looking system fails live. A backtest that fills every signal at the printed close pays no spread, suffers no slippage, and never misses size. For a low-frequency system with large average wins the distortion may be tolerable; for a high-turnover strategy whose average trade earns a few ticks, realistic costs routinely erase the entire edge. Modeling costs explicitly, and stress-testing the strategy's [sensitivity to cost assumptions](https://www.luxalgo.com/library/concept/cost-sensitivity/), is therefore a core part of honest strategy evaluation rather than an accounting afterthought. Cost models range from crude to elaborate. The simplest charge a fixed amount per trade. Better models scale costs with the instrument's spread and volatility, condition slippage on [relative volume](https://www.luxalgo.com/library/concept/relative-volume/) and time of day, and treat stop orders (which chase price) more harshly than resting limits (which risk non-fill instead). Institutional models add market impact, the adverse price movement caused by the order itself, which grows with order size relative to available liquidity. No model is exact; the practical goal is a conservative estimate that the live strategy tends to beat rather than an optimistic one it tends to miss. ## How it's calculated There is no single canonical formula; the standard per-round-trip accounting is: ``` round_trip_cost = commissions + spread_cost + slippage_entry + slippage_exit spread_cost = (ask - bid) / 2 per side crossed slippage = fill_price - decision_price (signed against the trade) cost_bps = round_trip_cost / notional * 10000 net_expectancy = gross_expectancy - expected_round_trip_cost commissions: broker and exchange fees for both sides of the trade decision_price: price at the moment the signal fired or the order was sent fill_price: volume-weighted average price actually received notional: dollar value of the position, used to normalize costs gross_expectancy: average profit per trade before frictions ``` Market-impact terms for large orders are often modeled as increasing with the square root of order size relative to daily volume; parameters are venue and instrument specific. Limit-order strategies substitute a fill-probability model for spread cost, since the cost of passivity is missed trades rather than paid spread. ## How traders use it - In backtesting, as a deduction applied to every simulated trade, ideally scaled to spread and volatility rather than fixed; this is the difference between [realistic and decorative cost assumptions](https://www.luxalgo.com/library/concept/cost-model-realism/). - As a viability screen: computing the strategy's average gross profit per trade in ticks and comparing it to modeled round-trip cost quickly reveals whether an edge survives friction at all. - In live monitoring, by recording decision price versus fill price on every order so the realized slippage distribution can be compared with the model and the model corrected, a feedback loop that also feeds [walk-forward evaluation](https://www.luxalgo.com/library/concept/walk-forward-analysis/). - In execution design: modeled costs guide choices between market and limit orders, between taking liquidity and posting it, and whether large orders should be sliced algorithmically via [TWAP, VWAP, or POV schedules](https://www.luxalgo.com/library/concept/twap-vwap-pov-execution/). - With honest limits: slippage in fast markets, around news, and at stop-loss prices is fat-tailed, so average-based models understate worst cases and should be paired with stress scenarios. ## Execution cost modeling vs adjacent concepts - **Cost Model Realism** (https://www.luxalgo.com/library/concept/cost-model-realism/): The evaluation standard this modeling serves: whether a backtest's assumed frictions plausibly match live trading. Cost modeling supplies the numbers; realism judges them. - **Cost Sensitivity** (https://www.luxalgo.com/library/concept/cost-sensitivity/): The stress test built on top of a cost model: re-running results across a range of cost assumptions to see how quickly the edge degrades. - **TWAP/VWAP/POV Execution** (https://www.luxalgo.com/library/concept/twap-vwap-pov-execution/): Execution algorithms designed to reduce the very costs being modeled, chiefly market impact, by slicing large orders across time or volume. ## FAQ ### How much slippage should I assume in a backtest? There is no universal number; it depends on instrument, order type, size, and session. A common starting point is one tick or spread per side for liquid futures market orders, then replacing assumptions with your own measured fills as soon as live data exists. ### Why did my live results underperform a backtest that already included costs? Usual suspects are optimistic fill logic (limit orders assumed filled on a touch), costs that ignore volatility spikes, stop orders modeled at their trigger price, and partial fills reducing size on winners. ### Do costs matter for swing and position traders? Less than for day traders, but they still compound: spread, fees, and overnight financing across dozens of trades a year can consume a meaningful share of a modest edge. ### What are requotes and why do they belong in a cost model? In dealer-quoted markets such as retail forex, the dealer may reject your order and offer a new price when the market moves. Modeling them as extra adverse slippage during fast conditions keeps the backtest honest. ## Implementations in the Library - Execution Cost Modeling (LuxAlgo): https://www.luxalgo.com/library/indicator/execution-cost-modeling/ ## Related concepts - Order Types: https://www.luxalgo.com/library/concept/order-types/ - Order Constructs: https://www.luxalgo.com/library/concept/order-constructs/ - TWAP/VWAP/POV Execution: https://www.luxalgo.com/library/concept/twap-vwap-pov-execution/ - Exchange Order Flags: https://www.luxalgo.com/library/concept/exchange-order-flags/ --- Source: https://www.luxalgo.com/library/concept/execution-cost-modeling/ (LuxAlgo Library, the encyclopedia of trading & technical analysis). Free to use with attribution: https://www.luxalgo.com/library/license/