How Post-Trade Cost Analysis Improves Trading Performance

Post-trade cost analysis compares what you intended to trade with what actually happened. It measures commissions and fees, execution prices relative to a chosen benchmark, delays, and the effect of orders that were only partly filled. The aim is to identify changes that improve results after costs—not simply to find the cheapest-looking fill.

LuxAlgo’s native charts, Journal, and Quant, our coding agent, support different parts of that review. The Journal helps organize actual fills and net trade results. Charts provide market context. Quant lets you test strategy rules with explicit cost assumptions. Detailed broker, venue, and order-level transaction cost analysis (TCA) requires additional execution and benchmark data.

• Choose the benchmark before measuring: Decision price, arrival price, and interval VWAP answer different questions.
• Use a consistent sign convention: In this guide, a positive cost means worse execution relative to the benchmark.
• Include incomplete orders: Attractive fills can conceal costly missed exposure.
• Compare similar trades: Adjust for size, liquidity, urgency, volatility, and session.
• Test the proposed improvement: More passive orders, smaller slices, or later execution each introduces trade-offs.

TCA Market Impact Webinar

OneMarketData’s webinar explains the relationship between execution, liquidity, and market impact. It is an older educational presentation; use current documentation for product capabilities and regulatory requirements.

Core Elements of Cost Analysis

Key Performance Metrics

CFA Institute’s trade execution framework distinguishes implementation shortfall from individual execution benchmarks. Implementation shortfall compares the outcome of the intended investment at the decision price with the actual outcome, including costs of trading and failing to trade.

Market VWAP is the sum of eligible trade price × trade volume divided by eligible volume over the chosen window. Your own quantity-weighted average fill price uses only your fills; it is not the market’s VWAP. TWAP instead averages observations or interval prices with equal time weighting under a defined sampling method.

For a benchmark price B and average execution price P, use buy cost in basis points = 10,000 × (P − B) ÷ B. For a sell, reverse the price difference: 10,000 × (B − P) ÷ B. One basis point is 0.01%. State whether fees are included. A negative result under this convention indicates price improvement, not automatically an execution problem.

A Worked Example: Partial Fills and Opportunity Cost

Suppose you decide to buy 1,000 shares at a decision price of $50.00. The order reaches the execution process when the reference price is $50.05. You buy 600 shares at $50.10 and 200 at $50.20, pay $8 in explicit costs, and leave 200 shares unfilled. At your predefined evaluation endpoint, the stock is $50.40. These are hypothetical figures.

The average fill price is $50.125 and the fill rate is 80%. The $100 filled-share price cost can be split into $40 of decision-to-arrival delay, using 800 × $0.05, and $60 of arrival-to-fill execution cost. Do not add those components to the $100 again. The $80 opportunity cost is a comparison with intended exposure, not an extra debit on the broker statement.

If interval market VWAP were $50.15, the fills would look $20 better than VWAP before fees: 800 × ($50.125 − $50.15) = −$20. Yet total implementation shortfall is still positive. Both calculations can be correct because they use different benchmarks and quantities.

Likewise, observed price movement after an order is not all caused by that order. Separating market impact from broader market movement, information, and timing requires a model or a suitable comparison group.

Getting Clean Trade Data

Begin with the records needed for your question. A broker statement can establish fills and charges. Quote-level or tick-level data is needed for finer analysis of spreads, arrival benchmarks, and short-horizon price changes. Authentication and compressed storage are operational details; they do not make a dataset accurate.

• Order records: Parent and child identifiers, side, intended quantity, decision time, submission time, amendments, cancellations, and completion status.
• Fill records: Execution time, price, quantity, venue or broker where available, commissions, fees, rebates, currency, and contract multiplier.
• Market references: Time-aligned quotes and trades, session calendars, benchmark intervals, corporate actions, and the precise source coverage.
• Reconciliation: Remove duplicates, account for corrected or canceled fills, normalize timezones and units, and match totals to statements.

Keep missing timestamps or quotes explicit. Do not assign an unavailable decision price from a later candle close. Synchronize clocks closely enough for the horizon being measured, and preserve the original records alongside the analysis.

Equity and FX Cost Analysis

Neither market is automatically “easy” to analyze. A reference feed that was unavailable to the trader can still be informative, but should not be treated as a guaranteed executable alternative.

Reading Analysis Results

Finding Trading Problems

Review costs in both currency and basis points. A $10 cost matters differently on a $2,000 order than on a $200,000 order. When summarizing a group, calculate total cost divided by the relevant total benchmark notional; a simple average of per-order basis points answers a different, equally weighted question.

Segment by order size relative to market activity, spread, volatility, urgency, instrument, session, and strategy. Then compare brokers, venues, or order methods within reasonably similar groups. Sending one broker the hardest orders can make an unadjusted ranking misleading.

A cost z-score can flag unusually large residuals: (observed cost − model-expected cost) ÷ estimated residual standard deviation. A value above 2 may be a chosen review threshold, but is not a universal failure rule or a reliable probability statement without distribution and model checks. Keep sample size, model version, and prediction uncertainty visible.

If you track alpha capture, define the expected return, reference horizon, and realized outcome explicitly. Otherwise the label can mix strategy quality, market movement, execution timing, and costs.

Spotting Cost Patterns

• Time of day: Compare similar trades around the open, midday, close, and announcements.
• Passive versus aggressive execution: Review price improvement alongside missed fills, waiting time, fees, and adverse moves after a fill.
• Order size: Check whether costs rise with participation or size relative to available liquidity.
• Outliers: Inspect the original order and quote history before attributing a bad result to a broker or algorithm.

There is no universal −10 or −15 bps threshold that identifies a broken execution process. The sign convention, asset, order difficulty, and benchmark all matter. A higher passive fill percentage also does not prove that total implementation costs fell.

Making Changes Based on Data

Improving Trade Execution

Change one rule at a time, preserve the previous method as a baseline, and compare later results on similar orders. A limit order controls its acceptable price; it cannot guarantee execution or lower total shortfall.

Selecting and Reviewing Brokers

Evaluate price improvement, fill rates, completion times, available order types, market access, operational reliability, and reporting quality together. Keep difficult and easy orders comparable, and investigate how each broker defines its metrics. Lowest commission or fastest fill alone is an incomplete selection criterion.

Review Actual Trades with LuxAlgo, Then Stress-Test Costs

Use the Journal for Real Trade Records

Open Journal beside Panels in LuxAlgo’s workspace header. Add a manual account, import a supported statement, or connect a supported broker. The Journal belongs to your account and stays separate from shared workspaces.

The Journal derives flat-to-flat round trips from fills, including average entry and exit, fees, net P&L, and duration. Review individual trades with their chart context and use Breakdown to examine categories such as symbol, hold time, weekday, and tags.

These views help identify trades worth investigating. Complete TCA still requires intended orders, decision and arrival references, and unfilled quantity; a net-profit chart alone does not reconstruct them. Preserve benchmark details in your review records and reconcile any charges that were not included in imported fees.

Test Cost Sensitivity with Quant

Use Quant to build a precisely defined strategy, review its code, and run it on the intended symbol and timeframe. In strategy settings, Properties controls simulation assumptions including commissions, slippage, and order size. The backtest report shows the resulting historical performance.

Keep the entry and exit logic fixed while comparing a base cost assumption with a stressed one. For example, a hypothetical average gross result of 12 bps per round trip leaves 8 bps after 4 bps of all-in costs, but −2 bps after 14 bps of costs. That sensitivity is useful even though it does not estimate the actual cost of every future trade.

Enter costs in the units the settings require: a slippage input in ticks is not a percentage or basis-point input. Avoid charging the same spread or slippage twice. Candle-based backtesting also cannot reproduce every queue position, routing delay, incomplete execution, or live market-impact effect.

Native chart volume tools can help explain the market backdrop, but volume profiles and liquidation-level indicators are not substitutes for execution records, stop orders, or a TCA engine.

Where Specialist TCA Platforms Fit

Institutional requirements may call for dedicated market-data integration, order attribution, counterparty comparisons, and audit workflows. Current examples include:

• Bloomberg BTCA: Multi-asset execution analytics and benchmarks, including activity from Bloomberg and third-party trading systems.
• OneTick BestEx and TCA: Execution benchmarks and analytics supported by tick-data infrastructure.
• ION LookOut TCA: Post-trade equity execution analytics with more than 80 performance and cost metrics, according to ION.
• TS Imagine’s Fixed Income Transaction Cost & Liquidity Analytics: Fixed-income analysis available within TradeSmart and through post-trade data workflows.

Choose by asset coverage, benchmark definitions, input data, and the review process you need. An award or a long metric list does not establish which service will improve a particular trading strategy.

Regular Review Methods

Connect Pre-Trade Expectations to Post-Trade Evidence

Before trading, record expected costs, urgency, and the intended benchmark. During execution, preserve order events and fills. Afterward, reconcile results, inspect exceptions, and compare costs with the estimate. Review individual exceptions promptly and broader patterns on a consistent schedule, with enough comparable trades to support a decision.

AI in Cost Analysis

Machine learning can help estimate expected costs or prioritize unusual executions when suitable training data and validation exist. Require chronological testing, realistic inputs available at decision time, and monitoring for changing market conditions. Predictions should not silently replace missing execution evidence.

Quant’s role here is strategy coding and historical testing with specified assumptions. Treat a change it helps you test as a hypothesis to evaluate, rather than evidence that broker routing, live fills, or regulatory reporting have been optimized automatically.

TCA and Best-Execution Review

TCA can support documented execution oversight; it does not by itself guarantee compliance. FINRA Rule 5310 addresses reasonable diligence by member firms in customer execution. MiFID II Article 27 addresses investment firms’ execution obligations and considers factors such as price, costs, speed, and likelihood of execution.

The applicable process depends on jurisdiction, entity, instrument, and client relationship. Preserve benchmark methodology, data lineage, exceptions, review decisions, and follow-up actions. Best-execution review and trade surveillance are related governance activities with different questions; a favorable VWAP result is not proof that every obligation was met.

Getting Started

Start with a reconciled set of fills and a clearly defined benchmark. Add intended-order and market data where the question requires them. Use LuxAlgo’s Journal to review the actual trading record, Quant to examine sensitivity to costs, and specialist TCA tools when you need detailed execution attribution.

The practical improvement comes from closing the loop: identify a repeatable cost pattern, make a controlled change, and verify that the later results improve after accounting for missed trades and risk.

References

• CFA Institute: Trade Strategy and Execution
• OneTick: Point-in-time TCA benchmark methodology
• Bloomberg: Transaction Cost Analysis
• FINRA: Rule 5310
• ESMA: MiFID II Article 27
• LuxAlgo: Journal overview
• LuxAlgo: Making strategies with Quant
• LuxAlgo: Strategy settings and backtest reports