Engineering note · Strategy risk evidence
A Risk Flag Is Evidence, Not a Safety Rating
Separate exposure escalation, duration outliers and equity masking into evidenced trading-strategy risk flags without claiming that an unflagged strategy is safe.
Buyer question
Which visible facts should make me investigate a strategy before trusting its headline results?
Evidence boundary
A flag points to a case that needs review. It does not prove fraud, danger or future loss.
Hard rule
No flag may claim that an unflagged strategy is safe.
A warning label is not an audit result
A strategy listing can show profit, drawdown and a smooth balance curve while leaving important account states out of view. The answer is not one broad risk score. It is a set of narrow flags that each point to a visible sequence and can be checked independently.
The source discussion asked how to detect grid-like or martingale behavior without pretending that a simple label proves what a system does. The useful boundary is evidence. If size grows after adverse movement, attach the trade sequence. If a position duration is unusual, name the baseline. If closed profit hides open loss, show balance and equity together.
Source and public evidence
Keep the three flags separate
Combining unrelated signals into one score hides the evidence and makes false confidence easier. Keep each flag attached to its own input, threshold and observed case.
Exposure escalation
Show the order sequence and how risk changed after adverse movement.
Trade IDs, side, size, entry price and timestamp
Duration outlier
Compare holding time with one declared rolling baseline.
Symbol, session, entry time, exit time and baseline window
Equity masking
Keep open P and L beside closed results at the same timestamp.
Balance, equity, floating P and L, deposits and withdrawals
Minimum evidence before a buyer decides
- complete order sequence for the flagged case
- symbol, timeframe, session and time zone
- balance and equity at matching timestamps
- spread, commission and swap treatment
- declared threshold and comparison window
- one negative case where the flag stays off
Three acceptance tests
These cases test exposure, time and reporting separately. Each result needs the exact evidence that caused the flag.
Test 1
Exposure escalation
Input: Position size grows after adverse movement while the strategy remains in the same trade idea.
Pass: The flag includes the exact order sequence, size changes and timestamps. It does not infer intent from one large trade.
Test 2
Duration outlier
Input: One position remains open far longer than the strategy's declared rolling baseline.
Pass: The flag names the baseline window, threshold and observed duration. A different session or symbol is not silently mixed into the comparison.
Test 3
Equity masking
Input: Daily closed profit is positive while open losses remain outside the published return figure.
Pass: The report keeps balance, equity and open P and L together so the same timestamp tells the complete account-state story.
What this checklist does not prove
These flags do not prove that a strategy is a grid, martingale or deceptive product. They identify specific evidence that deserves review. Strategy logic, data quality, broker execution and account ownership remain separate questions.
They also do not predict profit, loss or drawdown. A buyer still needs exact rules, costs, unseen-data testing and a complete handoff contract before treating a backtest as decision evidence.
Free next step
Check the evidence before trusting the label
Use the scope builder to record the strategy rules, sizing sequence, account-state fields and three pass cases. That turns a vague safety claim into facts a developer or buyer can verify.