Direct answer
Strategy Review involves examining how a trading strategy performed and what factors contributed to that performance. The main risks are not only about the market itself, but also about how results are measured, how trades were actually executed, and how conclusions are interpreted.
Because outcomes vary with costs, execution quality, and jurisdiction, a review that looks correct on paper can still be misleading in practice. Historical relationships are also not a guarantee of future results, so a strategy that “worked” during one period may fail under other conditions.
Mechanism or definition
A practical way to think about Strategy Review is: you define the strategy rules, compile the relevant outcomes, and then analyze which inputs, costs, and execution details drove those outcomes.
Several mechanics can introduce risk:
- Measurement assumptions: The review depends on which data is used (e.g., planned versus actual fills) and what time window and filters are chosen.
- Attribution steps: You try to explain “why” performance happened, which requires assumptions about causality.
- Repetition logic: If you repeatedly adjust the strategy based on the same dataset, you can start fitting to noise.
A key limitation is that the review process often assumes that what was intended is what occurred. In real trading, execution and costs can differ from what a backtest or spreadsheet implies.
Evidence or example
Consider a simplified example with explicit assumptions.
- Assumption A: A strategy targets entries and exits based on a rule set.
- Assumption B: The review uses executed trade outcomes.
- Assumption C: Costs include spreads and commissions, and slippage can occur.
If, during the review period, slippage was unusually low (for example, because liquidity was higher than usual), then performance may look better than it would in a different market regime. If a later period has higher slippage and different volatility structure, the same strategy rules can produce meaningfully different results.
Another example is operational mismatch. If the review uses timestamps or pricing that do not reflect the actual order lifecycle (placement, partial fills, or delayed execution), the computed performance can be biased. Even if the strategy rules are unchanged, the “review results” may reflect process details rather than strategy quality.
Limitations and risks
Below are the most material risk categories.
1) Operational risks (process and data)
- Incorrect or incomplete records: Missing trades, wrong time alignment, or inconsistent position sizing assumptions can distort results.
- Planned vs. executed difference: Real fills, partial fills, and latency can change outcomes relative to any idealized model.
- Implementation variance: A strategy can behave differently depending on order types, execution venues, and operational handling.
A common failure mode is concluding that the strategy is strong when the review actually measured execution conditions that will not repeat.
2) Market risks (changing conditions)
- Regime shifts: Volatility, liquidity, spreads, and correlation patterns can change.
- Non-stationarity: Relationships that appeared stable in a historical window can weaken or reverse.
- Tail events: Rare but impactful moves can dominate outcomes, so a review can be sensitive to which sample period was chosen.
This is why historical results do not establish future results.
3) Counterparty and cost risks (friction and dependencies)
Even without focusing on any specific provider, reviews often ignore how trading costs and execution friction affect net performance.
- Spreads and commissions: These directly reduce net returns.
- Slippage and market impact: Execution quality affects the realized entry/exit prices.
- Different execution environments: If the review period’s conditions differ from the next period, cost drag may change.
4) Interpretation risks (human judgment and statistical traps)
- Overfitting: Re-optimizing many variations against one dataset can produce a model that performs well in-sample but poorly out-of-sample.
- Survivorship and selection bias: Focusing only on strategies or time windows that already “look good” can bias conclusions.
- Causal overreach: A review may attribute performance to a factor that is correlated but not causal.
Verification or next question
To verify Strategy Review conclusions independently, focus on whether the analysis can be repeated under clearly stated assumptions:
- **Which data was used? ** Prefer outcomes based on actual executions, not idealized targets. - **What costs were included? ** Ensure you included realistic friction assumptions. - **What assumptions are time-dependent? ** If costs or execution quality are likely to change, treat conclusions as conditional. - **How sensitive are results to the sample?