Slippage and Execution Quality

Scope. This article explains what slippage is, how the total cost of getting a trade done decomposes, and why every measure of execution quality depends on a benchmark that somebody chose. MarketClue publishes no execution quality score, selects no benchmark on a reader's behalf, and does not characterise any execution as good or bad. It accepts, routes and executes nothing. Prices are small non-canonical illustrative figures describing no actual security. United States market structure, checked 17 August 2026.

The professional term for the full measure is implementation shortfall — the gap between the value of the decision and the value of what actually happened. It is more useful than slippage because it accounts for things that occur before an order is even sent.

Where the cost comes from

Suppose someone decides to buy 1,000 shares when the midpoint is $40.00. By the time the order reaches the market the midpoint is $40.06, and the order fills at an average of $40.1010 — the book-walking result computed in Market Orders.

Worked example — most of the cost happened before the order existed. The total shortfall is $0.1010 per share, or 0.2525% of the decision price — $101.00 on 1,000 shares. And 59.4% of it is delay: the market moved while the decision was being made and the instruction was being entered. Only 40.6% of the cost arose from anything the execution mechanism did. That proportion is specific to this illustration and would differ with any other numbers — the point is not the split but the fact that a split exists. A reader who examines only the fill price is auditing the smaller component and cannot see the larger one, because the delay component leaves no trace anywhere in the trade record. It is the same invisibility that Limit Orders identified in unfilled orders: the costs that show up on statements are not the costs that matter most. Figures are illustrative and describe no actual security.

The measurement problem

Every statement about execution quality is a comparison, and the comparison has to be chosen. Here is the identical fill measured four ways.

Worked example — one execution, four benchmarks, and the verdict changes sign. Against three benchmarks this fill was worse than the reference; against the fourth it was better. Nothing about the execution differs between rows. The choice of benchmark did not measure the quality of the execution — it determined the answer. This is precisely the failure mode set out in Benchmarking and Measuring Performance, arriving here at the level of a single trade rather than a portfolio: a decision made before the arithmetic determines the arithmetic's conclusion, and it is invisible in the result. Note also which benchmark flatters the execution. The day's volume-weighted average is the one measure in the table that could not have been known at the time of trading — so a favourable comparison against it says something about the rest of the day, not about the execution.

What the published statistics do and do not settle

In the United States, execution quality statistics are published under Rule 605 and order routing under Rule 606, as Payment for Order Flow describes. They are real, public and useful.

Two limits are worth understanding. They are aggregated across many orders, so they describe a population rather than any particular execution. And the price-improvement figures within them are measured against the prevailing published quote — which is a real benchmark and not the only conceivable one.

The unavoidable point, which this portal has made about payment for order flow and repeats here without softening: the comparison a reader most wants is against what would have happened otherwise, and that outcome does not exist. Slippage is always measured against a counterfactual, and the counterfactual is constructed rather than observed.

What can be reduced and what cannot

Some of the shortfall is mechanical. Spread and impact are properties of size against available depth, set out in Liquidity and Depth.

Some of it is a consequence of urgency. A faster execution consumes more depth; a slower one bears more delay. The two components trade against each other, and reducing one increases the other — which is why there is no setting at which both are minimised, and why this portal offers no guidance on where to sit between them.

And some of it is not a cost at all. If the price moved because of information that would have moved it regardless, the "delay cost" measures the market working rather than the trader failing. Distinguishing the two requires knowing why the price moved, which is not recoverable from the trade record.