XTesterStrategy Passport
Check a specific strategy version: sources, scenarios, plan, criteria, retained evidence and order explanations.
Start a Passport
Strategy Passport for a selected strategy, completed calculation or result. For a saved check, use Open Strategy Passport in Calculations. A passport collects evidence for a specific code and configuration version.
- Open the intended strategy or its original result.
- Select data source and check conditions.
- Choose Preview Scenarios and inspect the plan.
- Explicitly run the selected scenarios after review.
- Wait for status and open that job’s saved passport.
Sources include an executed run, completed background backtest, selected portfolio participant or the current strategy. Executed source can differ from today’s edited file: inspect source identity and fingerprint before comparing.
An old result without captured inputs requires a new baseline. Unknown settings are not guessed. Current-strategy research on generated scenarios does not require a previous historical backtest.
History and Generated Scenarios
Historical data checks the strategy against declared history and selected episodes. The required instrument and period must be available. Source gaps and limitations still matter.
Synthetic scenarios create hypothetical OHLCV paths from selected families and seeds. A current-strategy source specifies instrument/contract, capital, execution, initial price and generation horizon. Generated markets are not reconstructed exchange history.
Combining history and synthetic inputs compares different evidence sources without making them interchangeable. Preserve the distinction between historical results, reused periods and newly generated scenarios in plans and reports.
Inspect the Plan
The plan identifies strategy/source, settings, proposed scenarios, input size and unavailable checks. Compare instrument, period, timeframe, initial funds, fees and execution model with your question.
Choose phase timeframe, historical episode count and cost changes. Inspect evaluation thresholds. Additional checks are separate: parameters, other inputs, local detail, Monte Carlo and reserved periods.
Select checks for a stated question and available inputs. More rows alone do not establish reliability. After changing conditions, preview and inspect a new plan before execution.
Preview Scenarios persists a plan; it does not establish completed execution. Run Approved Scenarios is a separate action. Progress and cancellation use the shared calculation queue.
Interpret Criteria
The current form suggests at least 20 closed trades, return of at least 0% and drawdown up to 30%. These are visible editable starting values, not a universal strategy-reliability standard.
Read execution status separately from criterion verdict. Completed means the work finished; it can still report insufficient data, not evaluated or failed criteria. Errors, cancellation and unsupported configurations cannot count as passing checks.
A learning example that sends no orders cannot supply 20 closed trades. Zero activity can validate code execution without establishing trading-strategy robustness.
Retain and Compare Evidence
For an executed historical source, turn off new episode, cost and extra scenarios, then inspect the plan to retain existing evidence. Save Existing Evidence records the existing report without another trading run.
Additional research retains code/configuration identity and its source context. Several records from the same period are not independent validation. Reused data, overlapping episodes and small trade samples retain their limits.
A portfolio participant retains its shared-account context and attribution. Replaying it separately is a different comparison; the whole portfolio’s return cannot be assigned to every participant.
Warmup, unknown phases and incomplete coverage remain limitations. A phase score is a classifier output, not a probability of future profitability. A passport does not automatically enable Live monitoring or change risk.
Explain an Order
Select a specific order in the simulation account and choose Explain Order. The request preserves its original context: check that the order, project and run are the ones you intend.
Example request: “Explain which code condition produced this order, which inputs were available then and where the log supports it. Do not change or run anything yet.”
- Selected order
- Answer to the follow-up
Compare the answer with source, timestamps, settings and logs. An explanation does not change a strategy or validate a hypothesis by itself. Code editing, optimization and reruns require separate action.
If a Check Is Unavailable
No captured snapshot: resolve input capture and run a new baseline. Wrong project: open the job’s originating folder. Stale context: refresh source and plan rather than substituting unrelated identifiers.
Insufficient inputs or horizon: read the exact requirement and explicitly choose suitable conditions. Longer timeframes require warmup; a short generated series cannot replace missing history silently.
Unsupported execution or detail settings require revising the plan. Silent substitution of modeled inputs for real data is not a valid repair. Retain version, source, planId/jobId, status and the specific cause for support.
Related commands
Command Palette reference
Find these actions by name in the Command Palette; availability depends on the current project and operation state.
xtester.accountExplainOrderxtester.labJobs.checkRobustnessxtester.labJobs.openRobustnessxtester.openStrategyPassportxtester.strategy.openPassport