Execution Algorithm Simulator

The arrival price is the market price at the moment the order was released for execution, before any of it was worked. It is used as the benchmark because it is the price the decision was made against, so measuring against it captures everything that happened between deciding and finishing. Benchmarks measured over the execution window itself, such as the interval volume-weighted price, flatter an algorithm that simply tracked whatever the market did.

It targets a fixed share of whatever volume is actually trading, so its own pace follows the market rather than the clock. When volume rises it trades more, and when volume dries up it slows down. That makes it adaptive to liquidity, at the cost of an uncertain completion time: an order can finish early on a busy day or fail to complete at all on a quiet one, which is the tradeoff against a schedule-based approach.

Because impact rises faster than order size. A large order consumes the available liquidity at the best prices and then reaches worse ones, and it also signals the presence of a large buyer or seller, which invites others to trade ahead. Splitting the order lets displaced liquidity replenish between slices and makes the total demand harder to detect. The cost of doing so is time, during which the price can move for unrelated reasons.

It runs the slices against a simulated price path, which means the path is not affected by the order being executed against it. Real execution moves the market, and it also leaves information behind that other participants react to. Both effects work against the trader and neither appears here. The comparison between algorithms remains informative because the omission applies equally to all of them; the absolute cost figures are optimistic.

Because prices drift during the execution window for reasons unrelated to the order. A buy order worked into a falling market fills below the arrival price and shows a favorable result, and the same algorithm in a rising market shows the opposite. That component is direction, not skill, and it dominates single-order comparisons. Judging execution quality requires many orders across both directions, where the drift component averages out and the impact component does not.

It trades one risk for another. A short window concentrates the order into less liquidity, raising market impact, and finishes before the price has time to move much. A long window spreads the order thinly, reducing impact, and exposes the unfilled remainder to price movement for longer. Neither is universally better: the appropriate window depends on order size relative to normal volume and on how urgent the underlying decision is.

The comparison focuses on how the scheduling of slices interacts with the price path, so spread and fee costs are not the variable being isolated. In a real execution they are a material component, particularly for orders in less liquid instruments and for algorithms that cross the spread rather than posting passively. An algorithm that appears cheapest on schedule alone can be more expensive once the passive or aggressive posture of each slice is priced in.

Every algorithm degrades, and the choice between them matters less than the size decision does. An order representing a substantial share of normal daily volume cannot be executed in one session without moving the price materially, so the realistic options are extending across multiple days, seeking a block counterparty, or reducing the size. Slicing an order that is too large simply distributes the impact rather than avoiding it.

Some brokers offer scheduled execution algorithms to retail clients, and availability varies widely by firm and by account type. For most individual order sizes the question is largely academic, since an order small relative to the market's normal volume can be filled at the touch without meaningful impact. The concepts matter more than the access: they explain why a large order costs more than its quoted price implies.