Date: 2026-09-08 JST. Research only; no exchange execution or schedules.

Primary instrument: xyz:NVDA trade-price candles from official Hyperliquid candleSnapshot. Compare prior completed cash NVDA RTH daily OHLC (Yahoo independent feed, not TradingView-certified parity) and xyz:NVDA UTC 24-hour OHLC as pivot sources. Cash NVDA 5-minute RTH candles are an additional transferability check, NOT a substitute for XYZ backtesting. The original uploaded Pine is an indicator, not a strategy. This study tests its Traditional pivot formulas and price reactions, not a claim to replicate its full RSI/divergence/alert state machine or TradingView runtime. Cash and XYZ prices are never silently interchanged as execution prices.

• D/W/M Traditional formulas exactly as v2.16; intraday default hierarchy excludes Y.
• Visible-default level set: D P/R1/R2/S1/S2; W P/R1/S1; M closest above/below previous completed close, selected before the next candle. R short, S long, P direction from approach. R/S role reversal is not in this first test.
• Use only a previous completed pivot period; first partial W/M source period discarded. Cash daily OHLC dates are mapped to the execution chart UTC trading-day key; outside cash RTH, this is an explicit assumed XYZ chart boundary, not certified TradingView bar-by-bar parity.
• ATR14 Wilder RMA, 50 execution bars warm-up. A new touch requires previous close outside 0.25 current closed-bar ATR on the valid approach side and a signal candle whose price range intersects level +/-0.25 ATR. Floor buffer: 2 ticks, NVDA test tick=0.01.
• Raw touch variant accepts that event. Rejection variant additionally requires the signal close back on the approach side beyond max(0.10 ATR,2 ticks). These are study definitions, not native entry instructions from Pine.
• If multiple candidates qualify in a candle, take nearest level to the preceding close; one simulated position at a time in each independently tested variant. No reentry on the same exit candle. This suppresses repeated-contact inflation.
• Enter next candle open. TP=1 ATR, SL=1 ATR frozen at signal close. 5m horizon=12 bars (1h),15m=16 (4h),1h=6 (6h). Cash-only simulations force exit by regular session close and never enter across a session gap. XYZ permits around-the-clock holding. Reject non-contiguous XYZ paths if price gaps are missing data, not fabricate fills.
• Both TP and SL within one OHLC bar => stop-first. Gap beyond SL => open-price loss; favorable gaps exit at target, no beneficial fill improvement. Time expiry exits at close.
• Base round-trip all-in trading friction=6bps (0.06% of entry notional), plus historical funding when available. Sensitivities:0,12,20bps. These are assumptions, not verified historical spreads/fees. Historical Mark, L2 queue, latency, partial fill, large-size capacity and exact real exchange stop triggers are NOT recreated.
• Report count, TP/SL/timeouts, net-positive percentage, average net R, average net bps, profit factor, max cumulative-R drawdown, break-even friction and ambiguous bars. Never present target payoff as expectancy.
• Split chronological execution coverage 70/30 by timestamp; exclude trades spanning split. No tuning to the last segment. Weekly-cluster bootstrap intervals (1000 draws) are exploratory; few weeks and retrospective NVDA selection limit inference.
• Compare 20 deterministic shifted pivot grids per scenario. Each D/W/M period grid shifts by a seeded random +/-0.2..0.8 of its previous-period range. Same level count, formulas, entry rules and execution assumptions. This is a descriptive placebo benchmark, NOT a formal significance test or perfectly exchangeable control.
• Break down D/W/M, long/short, cash RTH versus other times, early/late period. These are diagnostics, not independently validated strategies. All tested variants must remain in results.

Immutable raw/normalized data + acquisition SHA256 under project-scoped blob prefix research:pivot216:NVDA:20260908. Keep research results separate from live Strategy ledger, project goals and authoritative trading handoff. Future symbols are configured by exact execution symbol, pivot source, tick, trading-day/session/calendar and fees; never inherit NVDA's session conventions for FX, metals, Korea equities etc.