Options-Market-Confirmed Daily Shock Continuation on Deribit's Own Linear Perps: trade WITH a >=1-sigma BTC or ETH UTC-day move only when that coin's Deribit DVOL repriced >=2 sigma UP on the same day, fixed 7-day clock (BTC_USDC- and ETH_USDC-PERPETUAL.DERIBIT, 1-HOUR bars with daily 00:00 UTC decisions, long-short, unlevered, flat ~85% of the time)

Original Research Lead hypothesis (not seeded). It is a two-coin, two-sided event panel on Deribit's own USDC-margined linear perps. Deribit is at 4.3% of the corpus against a combined options/Deribit target of 15% or more. The signal comes from the same venue's options market. Mechanism: on most big crypto days implied vol barely moves, because vol sellers absorb the move. On some days the options market reprices the 30-day DVOL index sharply UP alongside the price move. Then option buyers are paying for protection or upside in the move's direction and dealers are left short gamma. That has two effects: (1) dealers' delta hedging chases the move for days while the gamma imbalance is unwound; (2) the IV repricing shows that informed, hedging-motivated participants treat the move as information, not flow. Either way the move should CONTINUE over a multi-day horizon. RULE: once per UTC day at the 00:00 UTC hourly close, for each coin c in {BTC, ETH} compute (a) zr = the coin's UTC-day log return divided by the std of its daily log returns over the prior 60 days, from its own perp's hourly closes; (b) zv = the UTC-day log change of DVOL_c (supp key DVOL_BTC / DVOL_ETH, read point in time with supp_as_of) divided by the std of DVOL_c's daily log changes over the prior 60 days. If zv >= dvol_z_entry (baseline 2.0) and |zr| >= price_z_entry (baseline 1.0), go LONG coin c's perp when zr > 0 and SHORT it when zr < 0, then hold for a fixed hold_days (baseline 7) UTC days. Execution substitution, stated explicitly: next_bar_open_fill on this 24/7 venue is use_equivalent, so the order is a market order at the 00:00 UTC hourly bar's close, which on a continuous market is the next bar's open. PRE-STUDY (my own measurement on Deribit USDC perp hourly closes, 2022-03 to 2026-10, frictionless, no tuning): DOWN events (zv>=2, zr<=-1): BTC 30 and ETH 29 events, 59 in total on 43 distinct dates. Mean 7-day forward return -2.9% (BTC, t -1.8) and -3.9% (ETH, t -2.1); pooled mean -3.37%, median -1.55%, the short side wins 59%. UP events (zv>=2, zr>=+1): BTC 15 and ETH 19 events, 7-day forward +4.2% and +2.5%. At zv>=1.5 there are 36 and 32 events, at +2.1% and +0.9%. The unconditional 5-day drift is +0.22% (BTC) and -0.05% (ETH). The effect is weak at 3 days (-0.3% to -1.7%) and builds over 5-10 days; this is why the clock is 7 days and not intraday. The effect is NOT stable by year. Down events by year (BTC / ETH 7-day mean): 2022 -2.6% / -6.0%, 2023 -1.7% / -1.3%, 2024 +2.0% / +3.1% (reversed), 2025 -0.3% / -0.2%, 2026 -7.9% / -9.3%. Expected about 90 trades on about 70 distinct event dates. BTC and ETH events often share a date, so the effective sample is closer to 70 than 90. PRIOR ART, stated honestly: the closest relative, 'ETH Implied-Move Breach Classified by the Options Market's Reaction' (ETHUSDT.BYBIT 4H, 48h clock, continue when DVOL jumps / FADE when DVOL is flat), ended no_edge (117 trades, PF 0.84, IC about 0). This hypothesis differs in four ways. (i) It drops the fade arm entirely, because the pre-study shows no information in it. (ii) It uses a daily decision and a 7-day clock, where the pre-study locates the drift, instead of 48h, where the pre-study also finds almost nothing. (iii) It conditions on the DVOL change normalised by DVOL's own change volatility, not on an implied-move breach. (iv) It trades both coins on Deribit's own perps, where the options flow is. Other DVOL designs that died were regime gates, fades of IV peaks, or trend gates. None was a same-day IV-repricing continuation on a multi-day clock. RISK CONTROL route (b): signal/time exit with capped sizing, plus a fixed wide intrabar protective stop as a disaster bound. Each position is 25% of equity notional at leverage 1.0, at most one position per coin, maximum gross 50% of equity. COST ARITHMETIC: the Deribit perp round trip is about 0.07% taker plus about 0.02-0.04% half-spread and impact, about 0.10% in total, plus 7 days of funding. Funding has a typical magnitude of 0.01%/8h, about 0.2% per week, and is often favourable to a trend-chasing short after a down shock. The pre-study gross per event is about 2.5-3.5%; half of that, about 1.2-1.7%, is more than 10x the round trip. Risk profile (conservative): BTC/ETH daily std of about 3-4% gives about 60-75% annualised vol at 1x. Each 25% position therefore contributes about 15-19% annualised volatility, below the 120% cap. The book sits far from liquidation at 0.5x gross. The worst 7-day adverse move of about 30% at 25% notional costs about 7.5% of equity, below the 15% trade-loss cap, and the 15% disaster stop bounds it at 3.75%.