What “execution quality” means for a broker account
Execution quality is the degree to which orders placed through a broker are filled in a way that matches the trader’s intent (timing, price, and completeness). In practice, you cannot judge it from marketing descriptions; you judge it from what happened between order placement and execution.
For assessment, use neutral targets:
- Price alignment: how close the realized fill price is to the intended execution reference (for example, the last quoted price at submission time, or a mid-price definition you choose).
- Timing: whether the broker reports execution with a delay that is meaningful for your order horizon.
- Fill completeness: whether you receive the full requested size, or partial fills.
- Cost transparency: how fees, commissions, and spreads affect the realized outcome.
Core mechanics: what to measure and how
A practical measurement approach starts by defining an observation window and the fields you will use.
- Record the timeline For each order, capture:
- order submission time (as reported by your platform/logs)
- modification time(s), if any
- execution time(s) for each fill
- intended side (buy/sell) and requested quantity
- realized fill price(s) and filled quantity
- Choose a price reference and state the assumption Because “good” execution is relative, you must pick a consistent reference and explain it. Examples of references (choose one and apply it consistently):
- the mid-price you compute from bid/ask at the moment of submission
- the best available quote at submission time
- the quoted spread you observed at submission time
- Compute simple, comparable metrics Use metrics that do not require forecasts:
- Slippage (relative): realized fill price minus your reference, with sign adjusted for buy vs. sell.
- Average realized price: if multiple partial fills occur, compute a size-weighted average.
- Fill ratio: executed size divided by requested size.
- Latency proxy: difference between order submission time and execution time.
- Separate components you can control from components you cannot Stable assessment depends on your method. Variability depends on:
- market movement during the order life
- liquidity available at the moment of execution
- account-specific costs (commissions/fees) and how they are reflected in net price
Evidence and realistic limitations (what can fail)
Even with careful measurements, there are material limitations and failure modes:
- Reference mismatch: if your chosen reference does not reflect the true decision point (for example, quotes update), “slippage” may describe your reference choice more than execution quality.
- Partial fills and re-pricing: multiple fills can occur at different prices; comparing only the first fill can mislead.
- Time-stamp differences: platform logs may not be perfectly synchronized with broker execution reporting, so latency comparisons can be noisy.
- Survivorship bias in samples: if you only analyze trades that executed promptly, you may miss the conditions where quality deteriorates.
- Market-condition confounding: wide spreads or thin liquidity can dominate results, even if a broker’s process is stable.
A crucial limitation: historical relationships between order conditions and slippage do not guarantee future results, because microstructure and liquidity conditions change.
Verification: how to check independently
To verify what you are measuring, use a repeatable checklist:
- Consistency check: apply the same reference and metric definitions to all orders in your sample.
- Sensitivity check: repeat calculations using an alternative reasonable reference definition (for example, mid-price vs. best-quote) and see whether conclusions change.
- Stratify samples: compare metrics across different liquidity or spread regimes you can compute from your own recorded quotes.
- Look for outliers: identify orders with extreme slippage, long delays, or low fill ratio, then examine their recorded timeline for plausible causes.
If you can clearly explain your assumptions, show the raw order/execution fields used, and summarize results with confidence intervals or simple ranges, you can assess execution quality in a way that is evidence-based and does not rely on promises about future performance.