{
  "purpose": "Four explicitly selected positive illustrations, not a representative success-rate or profitability sample. All 57 close/reclaim candidates from the four fixed assets remain in the download.",
  "assets": [
    "BTC",
    "ETH",
    "SPY",
    "AAPL"
  ],
  "chart": "Cached 4H/session-block candles; weekly pivots from the previous completed calendar week. Source timestamps denote bar starts; values become known at the completed bar. Session blocks are not four hours of continuously traded time.",
  "window": "2025-01-01 through cached complete weeks before 2026-10-25",
  "contact": "First candle range intersects full band; this candle closes inside it, with week opening inward of its inner edge. A recorded close is used rather than a hypothetical fill at a wick.",
  "priceResponse": "First later close back inward of the inner edge, within six subsequent native bars. Lower: close above upper edge; upper: close below lower edge.",
  "positiveIllustrationFilter": "No outer-edge breach through that response candle; next up-to-six native bars have favourable high/low excursion at least 0.15 of prior weekly range; at least two of RSI50, MACD-zero and EMA9/21 conditions remain unconfirmed on response close. Future outcomes are used for this explanatory selection, never represented as information available at the response.",
  "selection": "Four editorially selected cases after viewing outcomes; no optimized strategy, no universal indicator replacement claim. Positive-case filter and all screened close/reclaim rows are retained.",
  "rules": {
    "RSI30/70": "Lower: RSI14 >=30; upper: <=70",
    "RSI50": "Lower: RSI14 >50; upper: <50",
    "MACD-signal": "Lower: EMA12 minus EMA26 exceeds its EMA9 signal; upper: below signal",
    "MACD-zero": "Lower: MACD line >0; upper: <0",
    "EMA9/21": "Lower: EMA9 > EMA21; upper: <",
    "SMA50-side": "Lower: close > SMA50; upper: <"
  },
  "calculation": "Wilder RSI14 with SMA seed and gap restart; EMA adjust=False with history warm-up; MACD 12/26/9; EMA9/21; SMA50. EMA formulas are computed from preceding cached history, not seeded at each chart window. Crypto gaps must be absent within the displayed week; equities use observed regular-session blocks. Source corrections and vendor bar conventions can change results.",
  "metric": "Confirmation price change is signed close-to-close movement from the response, not an executable entry or net trade return. No intrabar high/low ordering is assumed. Every rule already satisfied at the response is shown with zero additional bars. Absence is censored at the end of the shown week."
}