Direct answer
“Strategy Review” is a process for evaluating how a strategy performed, using inputs that typically describe (1) the strategy’s intent and rules, (2) the data it was applied to, (3) how trades or actions were executed, and (4) the evaluation method used to judge performance. Because the phrase can mean different implementations, you should treat “inputs used by Strategy Review” as the inputs required by the specific review method you are examining, not as a universal fixed list.
Mechanism or definition
A practical way to define Strategy Review inputs is to split them into stable mechanics and variable conditions.
- Strategy definition inputs (stable mechanics)
- Strategy logic: the rules or decision criteria that define when actions occur.
- Parameter values: fixed settings used by the logic (for example, thresholds, lookback windows, or risk parameters as defined in the strategy rules).
- Action mapping: how a decision becomes an executed action (order type, timing assumptions).
- Data inputs (variable conditions)
- Market data inputs: the historical series used to evaluate signals and outcomes.
- Time horizon and sampling: the review window, bar frequency, and any alignment rules.
- Data quality fields: handling of missing values, corporate actions (if relevant), and cleaning steps.
- Execution and cost inputs (often a common failure point)
- Costs: commissions, fees, and any modeled spread or transaction cost assumption.
- Execution model: assumptions about fills (for example, whether orders fill at observed prices, at mid-price, or with slippage).
- Latency and slippage: whether deviations between expected and actual execution are included.
- Evaluation method inputs (how the review is “scoring”)
- Metrics: what is measured (such as return measures, drawdowns, volatility, and stability across periods).
- Benchmarking: whether comparisons use a reference baseline and how it is defined.
- Constraints and filters: whether the review excludes certain periods, instruments, or trade types.
Evidence or example
Without assuming any real-time prices, you can still illustrate the dependency chain. Suppose a Strategy Review aims to compute a performance metric from historical outcomes.
- You need strategy definition inputs (logic and parameter values) to reproduce which actions would have been taken.
- You need data inputs (the historical series over the review window) to know what those actions saw.
- You need execution and cost inputs to translate “would have traded” into “assumed realized outcome.” If costs are omitted, results can look materially different.
- You need evaluation method inputs (the metric formula and aggregation rules) to compute the number being reported.
If any one input class changes—especially execution and cost assumptions—the review output can change even when the strategy logic stays the same. That is the core operational dependency of Strategy Review.
Limitations and risks
At least three material limitations are common:
- Missing or inconsistent inputs: if the strategy logic, parameter definitions, or data preprocessing steps are incomplete, the review cannot be independently verified.
- Regime change and non-generalization: historical relationships may not hold in future market conditions; a strategy can appear stable during one regime and fail in another.
- Model error in execution and costs: fills, slippage, and fees are often simplified. If those inputs are optimistic or inconsistent with reality, performance measures can be misleading.
Other risks include inconsistent time horizons (making results incomparable), cherry-picking selection windows, and survivorship-like biases if instruments or strategies are chosen after seeing outcomes.
Verification or next question
To independently verify “which inputs Strategy Review uses,” check the method definition behind the review you are reading:
- What exact data series and preprocessing steps are required?
- Which execution assumptions are specified (fills, slippage, and costs)?
- What metric formulas and evaluation windows are used?
- What stable strategy definition and parameter values are assumed?
A useful next question is: “Does Strategy Review separate strategy logic performance from execution and cost modeling?” If it does not, reported outcomes may reflect the execution model more than the strategy itself.