# IFVG Desk — bug report template

Copy this template and include the smallest example that shows the discrepancy. Instructions for collecting the header and event JSON are in [the audit guide](audit-trail-v1.2.md). Report what happened at the relevant confirmed candle; today's checklist can legitimately differ from an older saved decision.

## Problem

- Short description:
- Expected behavior and the rule you expected to apply:
- Actual behavior:
- Steps to reproduce:
- Does it repeat after recalculation with the same data/settings? If it changes, describe both runs:

## Chart and script

- IFVG Desk version and source file (indicator or strategy); note any local changes:
- Full symbol/feed ID:
- Chart timeframe and standard/synthetic candle type:
- Regular/extended chart session and data access differences, if relevant:
- Loaded date range, replay/live/historical mode and observed time:
- Problem candle's full date/time **with timezone**, plus `bar_open_ms` / `bar_close_ms` from JSON when available:
- Full Setup ID (`symbol|frame|birth-open-ms|bull/bear`), not the shortened table ID:
- Event ID(s) and event kind(s):

## Retention and settings

- Record setup history on/off:
- Retained audit capacity, retained count and older records removed:
- Exact Setup ID filter, or blank/all:
- Earliest event still present for this setup; identify any known missing events:
- Changes to settings, chart/session, loaded history or script since the issue occurred:
- Attach the exported header/settings JSON. If unavailable, include screenshots or values of all engine, context, gate, risk and audit inputs:

```text
Paste the complete report header/settings JSON here.
```

The cap is global (300 events by default, up to 500), so filtering one Setup ID may leave only part of its lifecycle. Non-overlapping retest bars and patterns that failed gap-birth filters are not setup events. Indicate missing coverage rather than assuming every omitted bar is an error.

## Saved event records

Include the relevant retained events in chronological order. Keep the gap-detection/inversion event and the accepted/skipped/rejected decision when available, followed by any invalidation, expiry, eviction or paper resolution that helps explain the issue.

```text
Paste the original event JSON lines here, with full Setup IDs and timestamps.
```

If an export is unavailable, include an unabridged row tooltip and a chart screenshot showing the relevant candle and gap. Note the script version shown in that screenshot. Keep the JSON's event timestamps; Pine Logs prefixes on historical batch output mark the export execution instead.

## Strategy Properties — complete separately for broker-result issues

The audit reports the shared paper engine and excludes costs. It cannot read user overrides to TradingView's broker emulator Properties. Attach a Properties screenshot or fill in these values:

- Initial capital and account/base currency:
- Commission type and value:
- Slippage in ticks:
- Order-fill settings, including processing on bar close and limit-order verification:
- Recalculation options, including every tick and after order fills:
- Bar Magnifier setting and any relevant lower-timeframe data availability:
- Margin settings and other Properties changed from the source defaults:
- Planned cash risk and quantity step from Inputs (these determine the explicit submitted quantity):
- Relevant Strategy Tester trade/order rows, entry/exit timestamps, quantities, prices and costs:

State whether the discrepancy concerns signal selection, the fixed paper plan, or a broker-emulator fill/result. For `paper_resolved`, the original plan and result are in `detail`; its `candidate_plan` is only a hypothetical plan at the resolution candle's close.

## Collection note

Pine Logs export requires an editable personal copy, with **Record setup history** and **Write bug report to Pine Logs** enabled. Published studies cannot generate Pine Logs. [TradingView documents the restriction here](https://www.tradingview.com/pine-script-docs/writing/debugging/). Save the copied records before recalculating. The export option does not send this report or create an external alert; submit your completed report through your chosen feedback channel.