Crash Stats 15m (ETH) — X% | prev RTH min(VWAP, Close)

# Crash Stats 15m (ETH) — X% Drawdown Event Analyzer

A 15-minute indicator that scans up to the last 5 years to find **crash events** where the close falls by at least **X%** relative to the **lower of** the prior day’s **RTH VWAP** and **RTH close**. It then measures recovery and follow-through behavior, tags the market regime around each event, and summarizes everything in a table.

---

## What the script detects

**Crash event (trigger):**

* On a 15-minute bar, `close <= refPrice * (1 - X%)`.
* `refPrice = min(previous RTH VWAP, previous RTH close)`.
* First touch only: subsequent bars below the threshold on the same trading day are ignored.
* Extended hours (ETH) are supported; if ETH is off, the script safely infers the previous RTH reference.

**Per-event measurements**

1. **Time to “turn up”** – first close **above the event-anchored AVWAP** (AVWAP cumulated from the trigger bar onward).
2. **Time to recover the reference price** – first close ≥ `refPrice`.
3. **Time to recover Y% above the crash-day average price** – first close ≥ `crashDayVWAP * (1+Y%)`.
4. **Post-crash lowest price & timing** – the lowest low and how long after the event it occurs, within a user-defined horizon (default 10 trading days, approximated in calendar days).
5. **Intraday RTH low timing** – on the crash day’s RTH session, when did the day’s intraday low occur, and **was it on the first 15-minute bar**?
6. **First 15-minute low of the RTH day** – recorded for context.

All durations are shown as **D days H hours M minutes**.

---

## Regime tagging (A / B)

For each event the script classifies the surrounding trend using daily closes:

* Let `r6m = (prevClose – close_6mAgo) / close_6mAgo`,
`r12m = (prevClose – close_12mAgo) / close_12mAgo`.
* **A**: both `r6m > 0` and `r12m > 0` (uptrend across 6m & 12m).
* **B**: one positive, one negative, and `r6m + r12m ≥ 0` (mixed but net non-negative).
* Otherwise: **—**.

This helps separate selloffs in strong uptrends (A) from mixed regimes (B) and others.

---

## Inputs

* **X — Crash threshold (%)**: default 5.
* **Y — Recovery above crash-day average (%)**: default 5.
* **Lookback years**: default 5 (bounded by data availability).
* **Horizon for post-crash lowest (trading days)**: default 10 (approximated as calendar days).
* **RTH session**: default `09:30–16:00` (exchange timezone).
* **Show markers**: plot labels on triggers.
* **Rows to display**: last N events in the table.

---

## Table columns

* Index, **Trigger time**, **Drop %**, **Ref price**, **Regime (A/B/—)**
* **Time to turn up** (above anchored AVWAP)
* **Time to ref price**, **Time to day VWAP + Y%**
* **Window lowest price**, **Time to window low**
* **RTH first-15m low**, **RTH lowest time**, **Was RTH low on first 15m?**
* **Crash-day VWAP**

---

## How to use

1. **Set chart to 15-minute** and **enable extended hours** for equities (recommended).
2. Keep defaults (**X=5%, Y=5%**) to start; tighten to 3–4% for more frequent events on less volatile symbols.
3. For non-US symbols or futures, adjust the **RTH session** if needed.
4. Read the table (top-right) for per-event diagnostics and aggregate averages (bottom row).

---

## Notes & implementation details

* Works whether ETH is on or off. If ETH is off, the script back-fills “previous RTH” references at the next RTH open and uses the prior daily close as a fallback.
* The “turn up” definition uses **event-anchored AVWAP**, a robust, price–volume anchor widely used for post-shock mean reversion analysis.
* Events are **de-duplicated**: only one event per trading day (per target RTH cycle).
* Lookback is limited by your plan and the data vendor. The script requests deep history (`max_bars_back=50000`), but availability varies by symbol.
* Durations use minute precision and are rendered as **days–hours–minutes** for readability.

---

## Quick troubleshooting

* **No events found**: lower **X%**, enable **ETH**, or ensure sufficient history is loaded (scroll back, or briefly switch to a higher timeframe to force deeper backfill, then return to 15m).
* **RTH boundaries off**: check the **RTH session** input matches the venue.
* **Few rows in table**: increase **Rows to display**.

---

## Typical use cases

* Back-test how fast different symbols tend to stabilize after a sharp gap-down or intraday shock.
* Compare recovery behavior across regimes **A / B** for sizing and risk timing.
* Build playbooks: e.g., if the RTH low occurs on the first 15m bar X% of the time, plan entries accordingly.

---

## Changelog

* **v1.0**: Initial public release with crash detection, anchored-AVWAP reversal, reference & VWAP+Y recovery timers, regime tagging, window-low timing, RTH low timing, and first-15m low capture.