Trade review software should preserve what was known and planned at the time of a decision, then help compare similar records without rewriting the story around the outcome. It is an evidence tool, not an explanation engine.
Start With Data Quality
Validate symbols, timestamps, side, quantity, fills, fees, and position matching before calculating results. Broker exports can use different formats and time zones. Duplicates, partial fills, corporate actions, and missing fees can materially change a summary.
Preserve the Original Plan
Record setup label, market and sector context, catalyst, planned entry, invalidation, size, maximum loss, and skip conditions before or at entry. After the trade, record actual fills and any deviation separately. Do not overwrite the original plan with hindsight.
Use Metrics as Descriptions
- P&L and fees: what the recorded trades produced after included costs.
- R-multiple: outcome relative to the recorded initial risk assumption.
- Setup and time groups: comparisons that require consistent labels and adequate samples.
- Plan adherence: observable differences between the written plan and execution.
These metrics can identify patterns to investigate, but they do not establish why a result occurred or predict the next one.
Run a Weekly Evidence Loop
Review data issues first, then group comparable trades and examine repeated process deviations. Choose one small change, define it precisely, and observe it over a preselected period. Keep the old and new samples separate rather than changing several rules at once.
Treat AI Summaries as Optional Drafts
Some trade-review products use AI to summarize notes or suggest tags. Those outputs require verification against the underlying records and clear privacy controls. Deterministic statistics and editable rule-based flags can often answer the first review questions without generated language.
What MAC Terminal Reviews
MAC Terminal’s Journal imports supported CSV or pasted fills and calculates P&L, win rate, R-multiples, time-of-day breakdowns, and rule-based behavior flags. It does not generate personalized AI feedback from Journal notes. Shared market analysis and themes are generated in a separate scheduled workspace, not as an explanation of a member’s trades.
Educational only. Trade-review statistics depend on complete, accurate records and do not predict future returns or prove that a process change will improve results.