What “execution quality” means in STP-style dealing
Execution quality describes how closely the real transaction outcome matches the order intent at the moment an order is sent. In an STP (straight-through processing) context, the key idea is that the broker’s systems attempt to pass orders to liquidity providers with minimal manual intervention. That framing is about process mechanics, not a promise of better results.
A practical assessment should separate stable mechanics from variable conditions:
- Stable mechanics are things you can define once and then measure repeatedly, such as reporting accuracy, latency measurement availability, and how fills are generated and recorded.
- Variable conditions include market volatility, liquidity at that moment, and changes in spreads and execution costs.
Measurable factors to evaluate
Focus on execution-related metrics you can compute from trade and order data. The main goal is to see whether the broker’s execution behavior is consistently close to the order intent.
1) Price slippage vs. quoted intent
Define slippage for your analysis. A common definition is the difference between the expected price used in your order decision (for example, the displayed quote at the time you triggered the order) and the actual fill price. Because quotes and order handling differ across platforms, your assumption must be explicit:
- Assumption example: use the last available mid or bid/ask shown at order submission time as “intent.”
- Then compute:
slippage = fill_price − intent_price(sign depends on buy/sell).
This metric is not a predictor; it is a descriptive measure of execution differences.
2) Fill characteristics and completion
Execution quality can degrade even if average slippage looks acceptable. Look for:
- Partial fills (multiple fills for one order)
- Order rejection or cancellation patterns
- Re-quotes / requotes-like behavior (where applicable)
Material limitation: different venues and order types can legitimately cause partial fills or different completion behavior. That is why you must compare like with like (same order type, size, and session conditions).
3) Cost components beyond spread
Two orders with the same spread can still have different effective costs due to:
- Commission or fee structure
- Execution price impact (captured by slippage)
- Timing effects (fast markets can move between intent and fill)
A measurable approach is to estimate effective execution cost per trade using the same assumptions across your sample window: implied cost = spread at intent time + (slippage converted into price terms). Keep your conversion consistent.
4) Timing and reporting consistency
Assess whether the broker’s platform data provides enough information to support timing analysis:
- Does the record include timestamps for order submission and fill?
- Are timestamps consistent across events (order created, accepted, filled)?
Even without “live” latency measurements, you can often evaluate internal consistency: for example, whether fill timestamps align logically with order states.
Evidence and example: a repeatable comparison
A realistic scenario-based method:
- Choose a narrow set of times (for example, the same trading sessions) and keep market instruments consistent.
- Use a consistent order template (same type, same size category, same time-in-force if applicable).
- For each executed order, record the intent reference you chose (quote shown at submission time) and the fill price and any commission.
- Compute descriptive statistics per condition: mean slippage, median slippage, and the distribution tails.
What you should watch for as an evidence limitation:
- If results only look good during calm periods but degrade during volatility, the “quality” claim is actually conditional.
- If your chosen intent reference is not stable (for example, different quote sources for different orders), comparisons become unreliable.
Limitations and risks to acknowledge
1) Outcomes vary with market conditions
Execution is influenced by liquidity availability at the moment of execution. A broker can follow a stable processing path but still face worse outcomes during sudden moves because the market moved or because depth thinned.
2) Historical relationships do not guarantee future results
Even if a broker’s execution was “tighter” in a past sample, that does not establish what happens next. Execution quality is context-dependent, and your next window may differ in volatility, spread regime, or liquidity.
3) Multiple failure modes can hide inside averages
Common failure modes include:
- Slippage spikes during rapid price changes
- Partial fills that change the cost profile
- Cost drift where fees or effective costs differ from what users assume