What Slippage and Market Impact Actually Measure
Slippage is the difference between an expected execution reference and the actual result. The term is often used loosely, so useful analysis must name the benchmark and decompose the cause. A fill can differ because the order crossed the bid-ask spread, the market moved before arrival, the order consumed order-book depth, routing took time, the order was only partially completed, or fees and financing were omitted.
Market impact is the portion associated with the order's interaction with available liquidity and other participants' response. Implementation shortfall is a broader framework that compares a paper decision with the realized portfolio result, including execution cost and the opportunity cost of unfilled quantity. These measures turn "the fill was bad" into testable components.
For background on quotes, spreads, and liquidity and how order routing and execution quality affect outcomes, see those companion pages. The execution cost calculator lets you apply these formulas to your own trades.
Key Takeaways
- Every slippage number needs a side, benchmark, timestamp, quantity, and unit.
- Spread crossing and market impact are related but not identical.
- Delay before order arrival should be separated from execution after arrival.
- Large size relative to liquidity increases impact risk nonlinearly.
- Unfilled quantity creates opportunity cost that completed-fill statistics omit.
- Backtests should use sensitivity ranges, not one optimistic slippage constant.
Choose the Benchmark Before Looking at the Result
Possible benchmarks include decision price, arrival midpoint, arrival quote on the relevant side, opening or closing auction price, volume-weighted average price, and a schedule-specific benchmark. Each answers a different question. Decision price captures delay between the strategy decision and order release. Arrival price focuses on execution after the order is ready. VWAP compares with market activity over an interval but can be gamed by choosing the interval afterward.
A benchmark must be observable, available at the relevant time, and aligned with the objective. Do not compare a risk exit with the day's best price. Do not compare an auction order with a random continuous quote after the cutoff.
| Benchmark | What it measures | Main limitation |
|---|---|---|
| Decision price | The market reference when the strategy decided to trade | Requires a precise timestamp at the decision moment |
| Arrival midpoint | Mid of bid and ask when the order became available | Midpoint may not be executable; does not include spread cost |
| Arrival quote (side) | The executable price on the buy or sell side at arrival | Reflects spread cost but not depth beyond top of book |
| VWAP | Volume-weighted average over an execution interval | Can be selected after the fact; inappropriate for point-in-time decisions |
| Open / close auction | The official auction print for orders designated to participate | Only valid for auction-directed orders; cannot be used for continuous fills |
Signed Slippage and Dollar Cost
Use direction so unfavorable movement has the same sign for buys and sells. One convention is side × (fill price − benchmark), where side is +1 for buys and −1 for sells. Multiply by quantity for dollar cost, then divide by benchmark notional for basis points.
A buy of 2,000 shares benchmarked at $30.00 and filled at an average $30.06 has six cents per share of unfavorable slippage, $120 total, or 20 basis points. A sale benchmarked at $30.00 and filled at $29.94 has the same signed cost. State whether fees are included.
| Quantity | Benchmark | Fill price | Per-share slippage | Dollar cost | Basis points |
|---|---|---|---|---|---|
| 2,000 shares (buy) | $30.00 | $30.06 avg | +$0.06 | $120 | 20 bps |
| 2,000 shares (sell) | $30.00 | $29.94 avg | +$0.06 (signed) | $120 | 20 bps |
Formula summary: Slippage (bps) = [side × (fill − benchmark) / benchmark] × 10,000
Decompose Implementation Shortfall
Start with the paper portfolio created at the decision benchmark. Compare it with the actual portfolio after fills. Completed-order cost can be separated into delay, spread, impact, and fees where data permits. Unexecuted shares contribute opportunity cost if the market moves away, or opportunity benefit if not buying avoids a decline. The latter should not be treated as execution skill without a defined cancellation rule.
The decomposition prevents a fast, expensive fill from looking better merely because it completed, and prevents a cheap partial fill from looking better while ignoring missed quantity.
Implementation shortfall formula:
IS = (paper portfolio value at decision) − (realized portfolio value after fills)
= delay cost + spread cost + market impact + fees + opportunity cost of unfilled sharesHow Market Impact Develops
Impact depends on order size relative to available and replenishing liquidity, urgency, volatility, information content, participation rate, and the market's response. A visible aggressive order can remove quotes and signal demand. Other participants may raise offers or withdraw. After completion, some movement may reverse if it was temporary liquidity pressure; some may persist if the order conveyed information or coincided with broader repricing.
Simple linear assumptions often fail for very small or very large orders. The first 500 shares may trade at one level, while the next 10,000 consume increasingly expensive depth. Stress tests should include nonlinear cost as size grows.
- Temporary impact — price movement that reverses after liquidity pressure subsides.
- Permanent impact — the persistent component associated with information or lasting repricing; difficult to identify precisely from one order.
Modeling Slippage in Backtests
A fixed cents-per-share model is easy but distorts securities with different prices. Fixed basis points scale with price but ignore liquidity. Half-spread plus a size-impact term is more realistic when spread data and volume proxies exist. Event strategies may require separate models for open, close, earnings, halts, and extended hours.
Use multiple scenarios: low, base, and stress. Re-run performance under delayed entry, wider spread, reduced fill probability, and capped participation. Report how much gross edge survives. A strategy that only works with same-bar midpoint fills has not demonstrated a robust edge.
| Model type | Formula / approach | Strength | Weakness |
|---|---|---|---|
| Fixed cents per share | cost = N × c |
Simple, transparent | Distorts across price levels; ignores liquidity |
| Fixed basis points | cost = notional × bps / 10,000 |
Scales with price | Ignores spread, depth, and size effects |
| Half-spread + size impact | cost = 0.5 × spread + k × (size / ADV)0.5 |
Captures spread and nonlinear impact | Requires spread data and calibrated impact coefficient |
| Session-segmented | Different parameters for open, midday, close, earnings | Reflects real liquidity variation | More parameters, harder to calibrate |
For a deeper look at building robust backtesting frameworks that account for realistic execution costs, see the backtesting guide.
Reducing Execution Cost Without Hiding Risk
Reduce size, extend execution time, use price constraints, avoid thin sessions, participate at a lower rate, or select more liquid instruments when those choices fit the strategy. However, patience creates opportunity risk and information can decay. Splitting an urgent exit into tiny pieces can be more dangerous than paying impact promptly.
The decision rule should compare marginal impact with marginal delay risk. For each additional unit of patience, ask what price improvement is plausible and what loss occurs if the market moves away. This turns execution style into an economic choice.
This tradeoff is central to risk management: the cost of a patient exit can exceed the cost of a fast one when urgency is high.
Execution-Cost Decomposition
Separate causes before choosing a remedy. Each component requires a different diagnosis and a different fix.
| Component | Benchmark transition | Possible remedy |
|---|---|---|
| Decision delay | Decision price to arrival price | Automate workflow, simplify approvals, improve data timing |
| Spread cost | Midpoint to executable quote | Use patient limits when non-fill risk is acceptable |
| Depth / impact | Top quote to average fill | Reduce size, stage order, select liquid time |
| Fees | Execution plus explicit charges | Compare all-in broker and venue economics |
| Opportunity cost | Unfilled quantity versus later price | Adjust limit, duration, or completion rule |
| Post-trade drift | Fill versus later benchmark | Study adverse selection; do not automatically call it routing error |
Worked Scenarios
Small liquid order
Situation. A 100-share marketable buy crosses a one-cent spread and fills at the ask.
What the evidence says. Most cost may be ordinary half-spread relative to midpoint; size impact is probably small.
Practical response. Do not overengineer routing conclusions from a one-dollar economic difference.
Large participation rate
Situation. An order represents 35% of the stock's volume over ten minutes.
What the evidence says. The trader may be a major source of demand and can cause or amplify price movement.
Practical response. Lower participation, extend horizon, or accept that urgency has a measurable price.
Limit order partly fills
Situation. Half the target fills cheaply and the remainder misses a rally.
What the evidence says. Completed shares show low cost while unfilled shares create opportunity cost.
Practical response. Evaluate the total decision, not only average fill of completed quantity.
News-driven exit
Situation. A position must be reduced after material negative news.
What the evidence says. Delay risk may dominate spread and impact; minimizing measured slippage could increase total loss.
Practical response. Use a benchmark and urgency rule appropriate to risk containment.
Backtest uses 5 bps everywhere
Situation. The same cost is applied to large caps, microcaps, opens, and midday trades.
What the evidence says. The model ignores actual spread, depth, session, and size differences.
Practical response. Segment by liquidity and market phase and test wider ranges.
Practice Lab: Turn the Concept into a Repeatable Process
Each exercise follows a three-step format: state the case, identify competing explanations for the evidence, then apply the practical response. Write out your answers before checking the guided response.
Exercise 1: Small liquid order
Case. A 100-share marketable buy crosses a one-cent spread and fills at the ask.
Do not begin by choosing an order or judging the outcome. First write the exact objective, the information available at the decision timestamp, the quantity, and the maximum acceptable adverse result.
Next, identify which evidence is observable and which is inferred. The key interpretation is: most cost may be ordinary half-spread relative to midpoint; size impact is probably small. Convert that interpretation into at least two competing explanations. This prevents a single screenshot or fill from becoming a false certainty.
Finally, apply this response: do not overengineer routing conclusions from a one-dollar economic difference. Record what would cause you to keep, modify, cancel, or escalate the plan. A complete answer includes the benchmark, market phase, price boundary, completion rule, and post-event review field.
Exercise 2: Large participation rate
Case. An order represents 35% of the stock's volume over ten minutes.
Do not begin by choosing an order or judging the outcome. First write the exact objective, the information available at the decision timestamp, the quantity, and the maximum acceptable adverse result.
Next, identify which evidence is observable and which is inferred. The key interpretation is: the trader may be a major source of demand and can cause or amplify price movement. Convert that interpretation into at least two competing explanations.
Finally, apply this response: lower participation, extend horizon, or accept that urgency has a measurable price. Record what would cause you to keep, modify, cancel, or escalate the plan.
Exercise 3: Limit order partly fills
Case. Half the target fills cheaply and the remainder misses a rally.
Do not begin by choosing an order or judging the outcome. First write the exact objective, the information available at the decision timestamp, the quantity, and the maximum acceptable adverse result.
Next, identify which evidence is observable and which is inferred. The key interpretation is: completed shares show low cost while unfilled shares create opportunity cost. Convert that interpretation into at least two competing explanations.
Finally, apply this response: evaluate the total decision, not only average fill of completed quantity. Record what would cause you to keep, modify, cancel, or escalate the plan.
Exercise 4: News-driven exit
Case. A position must be reduced after material negative news.
Do not begin by choosing an order or judging the outcome. First write the exact objective, the information available at the decision timestamp, the quantity, and the maximum acceptable adverse result.
Next, identify which evidence is observable and which is inferred. The key interpretation is: delay risk may dominate spread and impact; minimizing measured slippage could increase total loss. Convert that interpretation into at least two competing explanations.
Finally, apply this response: use a benchmark and urgency rule appropriate to risk containment. Record what would cause you to keep, modify, cancel, or escalate the plan.
Exercise 5: Backtest uses 5 bps everywhere
Case. The same cost is applied to large caps, microcaps, opens, and midday trades.
Do not begin by choosing an order or judging the outcome. First write the exact objective, the information available at the decision timestamp, the quantity, and the maximum acceptable adverse result.
Next, identify which evidence is observable and which is inferred. The key interpretation is: the model ignores actual spread, depth, session, and size differences. Convert that interpretation into at least two competing explanations.
Finally, apply this response: segment by liquidity and market phase and test wider ranges. Record what would cause you to keep, modify, cancel, or escalate the plan.
Common Failure Modes
No benchmark timestamp
Without it, the expected price can be chosen after the result.
Correction: Capture the reference automatically at decision and arrival.
Calling all price movement impact
The market may move independently of the order.
Correction: Use controls, participation data, and humility about attribution.
Ignoring opportunity cost
Partial-fill strategies look artificially cheap.
Correction: Include unfilled shares in implementation shortfall.
Using gross strategy returns
Execution cost can consume a small edge.
Correction: Report net sensitivity across cost scenarios.
Optimizing cost without urgency
The cheapest-looking order can fail the strategy objective.
Correction: Value delay and completion explicitly.
Decision Checklist
- Name the objective. Completion, benchmark tracking, risk exit, or price improvement.
- Capture decision and arrival benchmarks. Separate workflow delay from execution.
- Use signed cost. Make unfavorable buys and sells comparable.
- Include quantity and fees. Convert cents to economic dollars.
- Measure completion. Track unfilled opportunity.
- Estimate participation rate. Relate size to current liquidity.
- Run low/base/stress costs. Avoid one-number confidence.
- Feed results back into sizing. Execution uncertainty changes maximum position.
Key Terms Used on This Page
- Decision price
- The market reference when the strategy or manager decided to trade.
- Arrival price
- The reference when the order became available for execution.
- Implementation shortfall
- The difference between a hypothetical decision-price portfolio and realized outcome under a defined methodology.
- Participation rate
- Order volume as a share of market volume over the relevant interval.
- Temporary impact
- Price movement that reverses after liquidity pressure subsides.
- Permanent impact
- The persistent component associated with information or lasting repricing; difficult to identify precisely from one order.
Frequently Asked Questions
What is slippage in trading?
It is the difference between an expected benchmark and actual execution, stated with side, quantity, time, and methodology.
Is slippage always bad?
A fill can improve relative to the benchmark. Use signed favorable and unfavorable results rather than assuming every difference is a loss.
What causes slippage?
Spread crossing, market movement, delay, insufficient depth, impact, partial fills, routing, volatility, and fees can all contribute.
What is market impact?
It is price movement or execution cost associated with the order consuming liquidity or changing other participants' behavior.
How much slippage should I use in a backtest?
Use data- and strategy-specific ranges. At minimum, test spread-aware base and stressed assumptions by liquidity and session.
Is VWAP always the right benchmark?
No. It is useful for some interval-execution objectives but can be inappropriate for point-in-time decisions, urgent exits, or auction orders.
Can limit orders eliminate slippage?
They cap eligible execution price but can create non-fill and opportunity cost. They do not eliminate all execution uncertainty.
Educational Disclaimer
This page explains slippage and market impact for general educational purposes. It does not evaluate your financial circumstances, recommend a security, select a broker, or tell you which order to place. Quotes can change before an order reaches a market. Examples use simplified assumptions and exclude taxes, fees, financing, borrow costs, corporate actions, and other account-specific factors unless stated. Brokerage capabilities, exchange procedures, market-data entitlements, tax treatment, and regulatory requirements can change. Verify operational and legal details with the broker, exchange, regulator, or tax professional responsible for the decision.