Define the decision point
Execution quality starts with the decision timestamp. The arrival price is commonly the prevailing benchmark when an order is released to the market, but the applicable benchmark depends on mandate, instrument and data quality. A mid-price benchmark is informative only when bid and ask observations are sufficiently synchronized.
For an executed purchase of quantity Q, one simplified signed measure is Q × (average fill price − arrival price). Fees, rebates, spread, partial executions, opportunity cost and market drift require additional, separately reported components. This formula describes measurement structure, not IPAFRI's realized execution history.
Event-time integrity
Each event should carry a normalized clock and immutable identifier: market-data arrival, signal creation, risk decision, order submission, acknowledgment, partial fill, cancellation and final reconciliation. Clock drift and message reordering are ordinary engineering risks, not edge cases.
Microstructure and attribution
Spread crossing, queue position, market impact, venue fees and price movement during the order horizon can all affect realized cost. A single average slippage figure conceals tails, regimes and incomplete fills. Conditional distributions and confidence intervals are more meaningful than isolated screenshots.
Reference architecture, not a performance claim
The diagram below is a conceptual map of research, limits, gateway state and reconciliation. It is not presented as evidence of institutional connectivity, active client accounts or executed orders.
Minimum reporting standard
- Declare benchmark, currency, timestamp resolution and trading calendar.
- Report both executed quantity and unfilled opportunity cost.
- Separate fees, impact, spread and timing where identifiable.
- Publish sample size and uncertainty before interpreting cost improvements.