Many missing alerts at the same offset
Check timezone selection, bar-open versus bar-close behavior, and whether the TradingView alert was recreated after the Pine code changed.
Investigate Backtest vs Alerts results
Start with the rows, not a grade. The summary tells you how much of the selected scope matched; the comparison grid tells you which events produced that result.
Read the counts first
Section titled “Read the counts first”| Metric | Question it answers |
|---|---|
| Backtest events | How many entry and exit events were imported inside the scope? |
| Alert events | How many webhook or CSV alert events were included? |
| Matched events | How many pairs met the selected direction, time, and price rules? |
| Missing alerts | Which backtest events had no compatible alert? |
| False alerts | Which alerts had no compatible backtest event? |
The displayed match rate is coverage under the current configuration. It is not a probability that the strategy is correct, profitable, or ready for live capital. Always report it with the event counts and tolerances.
Investigate a discrepancy
Section titled “Investigate a discrepancy”-
Confirm the date range, symbol, strategy name, and alert source. A scope mistake can make a correct workflow look divergent.
-
Confirm how zone-less timestamps were interpreted. Re-run after changing CSV timestamp timezone if the drift is consistent with a timezone offset.
-
Turn on Show only problems and separate missing alerts from extra alerts.
-
Inspect direction first. An entry-long event cannot match an exit-long or entry-short event even when its timestamp is close.
-
Inspect time and price differences. Change a tolerance only when the strategy’s expected timing, bar interval, and market behavior justify the new rule.
-
Re-run and record what changed. A tolerance change explains a different match result; it does not repair the source workflow.
Common patterns
Section titled “Common patterns”Extra alerts cluster together
Check duplicate TradingView alerts, repeated webhook delivery, pyramiding, and intrabar conditions that can fire more than once.
Entries match but exits do not
Inspect exit-order intent, stop or limit behavior, and the alert payload’s position direction. Do not compensate with a wider time tolerance first.
Price drift changes by market regime
Segment the rows by volatility, session, and order type. Compare with your own prior baseline rather than a universal percentage label.
Timing and price metrics
Section titled “Timing and price metrics”- Time difference is alert time minus backtest-event time. Its sign shows whether the alert was earlier or later.
- Price difference is alert price minus backtest-event price. Interpret it with symbol price, tick size, volatility, and the payload field used as the alert price.
- Averages can hide clusters and outliers. Use the comparison rows and chart markers to inspect the distribution.
Save evidence deliberately
Section titled “Save evidence deliberately”Use Save Snapshot when the scope, input files, and tolerances are worth comparing later. Include those details in the title or notes. If you connect the saved report to a case study, treat it as supporting evidence—not as an automatic promotion decision.
Continue the investigation
Section titled “Continue the investigation”- Review matching configuration
- Troubleshoot import and matching problems
- Review broker-reported fills after fill data exists