DeribitBtcWeeklyLongStraddleRvCompression

Deribit BTC Weekly Long Straddle on Realized-Volatility Compression (Vega-Long Vol-Expansion Bet)

A delta-neutral, multi-instrument options strategy on Deribit BTC weekly ATM options that buys the next-expiring weekly ATM straddle (1 ATM call + 1 ATM put, same strike, same expiry) when realized volatility has been compressed and reverts upward. Addresses the SINGLE LARGEST quota gap in the 658-strategy portfolio: OPTIONS venue at 0.2% vs ≥15% target (14.8% gap, essentially zero coverage). Deribit is the most liquid crypto options venue with 1610+ active BTC/ETH contracts and full IV/greeks data in catalog — this is the lowest-hanging fruit for diversifying the portfolio off Binance USD-M perp dominance (87.8%). The mechanism is volatility mean-reversion: BTC realized volatility exhibits documented clustering AND mean-reversion (GARCH dynamics), and weekly periods following compression (low rolling 14-day realized vol) are statistically more likely to see vol expansion than periods following high realized vol. The strategy is delta-neutral at entry (call delta ≈ +0.5, put delta ≈ -0.5, net ≈ 0), so it's not a directional bet — it's a pure vega/gamma play. Profits if BTC moves >breakeven in EITHER direction by expiry. Conceptually orthogonal to every other strategy in the portfolio: not breakout, not reversion, not pairs, not microstructure, not carry — purely volatility-of-volatility. AVOIDS the failure patterns documented this session: (1) NOT a 4H bar-pattern strategy (no 4H regime dependency); (2) NOT dependent on lottery-win single trades (~30-40 trades/year, fairly uniform PnL distribution); (3) NOT a transplanted equity pattern (vol-clustering is universal in financial markets); (4) NOT a breakout (works equally in trending and ranging regimes — only needs volatility expansion).

Addresses the single largest portfolio quota gap (Options ~0.2% vs >=15% target) with a mechanism conceptually orthogonal to every existing strategy: not breakout, not reversion, not pairs, not microstructure, not carry -- purely volatility-of-volatility. The edge is volatility mean-reversion: BTC realized vol exhibits documented GARCH clustering AND mean-reversion, so weekly periods following compression (low rolling RV) are statistically more likely to see vol expansion than periods following high RV; a long ATM straddle is pure long-vega/gamma convexity that monetizes that expansion in either direction. Delta-neutral entry makes it non-directional. Avoids this session failure patterns: not a 4H bar-pattern strategy, not dependent on lottery single trades (~30-40 trades/yr, fairly uniform PnL), not a transplanted equity quirk (vol-clustering is universal), and not a breakout (works in trending or ranging regimes -- only needs vol expansion). Defined risk: long options cap loss at premium paid. Deribit is the most liquid crypto options venue with full IV/greeks history. Real options exposure with real option-leg trades in the analyzed backtest (smoke fires only in verification). leverage=1.0 (options venue cap); premium-bounded sizing, no unused-leverage gate.

Structurally wrong-sided premise — buying volatility loses the variance risk premium, confirmed by a correctly-implemented backtest. The long ATM straddle is genuinely long vol (net vega +5,309) and pays theta (-20,378) betting that RV compression predicts vol expansion (GARCH mean-reversion). But crypto implied vol persistently exceeds realized vol, so buying the straddle means PAYING the very premium that the short-vol VRP siblings (e.g. the bull-put-spread bc8310ca, which reached optimize-class) correctly HARVEST: you buy at elevated IV, bleed theta every week, and the realized expansion after compression rarely exceeds the IV-implied breakeven (the expected expansion is already priced into the IV). Result: loses EVERY year (2022 -11%, 2023 -4%, 2024 -9%, 2025 -26%, 2026 -11%; total -48.6%, Sharpe -7.55, PF 0.62, expectancy -$225/trade, omega 0.30). The positive skew (+1.23) confirms the engine correctly models the long-straddle payoff (occasional wins, many small premium losses) — this is a real wrong-sided edge, not an artifact. Not iterate: the loss is pervasive across ALL regimes (bear 2022 AND bull 2024/2025), so there is no robust region, and the 3 timing parameters (rv_lookback, rv_rank_max, rv_rank_lookback) only change WHEN to buy — no entry timing makes systematically long BTC vol profitable against a persistent risk premium. Not optimize: a strategy negative every year has nothing to tune toward. The correct expression of this volatility edge is to SELL the rich vol (the VRP-harvesting siblings), not buy it. FAILURE PATTERN: long ATM straddle / long-vega vol-expansion bets on crypto systematically lose the variance risk premium (IV > RV) — paying theta for overpriced optionality and waiting for RV expansion that rarely beats the IV-implied breakeven; the volatility edge on Deribit is on the SHORT-vol (premium-selling) side, defined-risk, not the long-vol side.

Delta-neutral LONG-VOLATILITY options strategy on Deribit BTC weekly ATM options: each weekly cycle BUY the next-expiring ATM straddle (1 ATM call + 1 ATM put, same strike/expiry) when BTC realized volatility has been COMPRESSED, betting on mean-reverting vol expansion. The backtest rolling-options expander resolves the weekly BTC chain from parameters[rolling_options] (underlying BTC, weekly roll, 7 DTE, two 1-DAY ATM legs at offset 0.0, kinds C and P) and the strategy BUYS both legs (leg_sides C:BUY, P:BUY) only when the realized-vol gate passes: RV-RANK <= rv_rank_max (0.30), where RV is the rolling rv_lookback (14) day realized vol of BTC (std of daily log-returns annualized) and RV-rank is its position within a trailing rv_rank_lookback (180) sample window. Delta-neutral at entry (call ~ +0.5, put ~ -0.5), so it is a pure vega/gamma bet that profits if BTC moves beyond breakeven in EITHER direction. Exit closes the straddle at the roll buffer before expiry, or early on a profit-take (combined straddle mark >= entry premium x (1 + profit_take)). Max loss is the premium paid (defined risk). The base RollingOptionsStrategy owns schedule/roll; this subclass adds the RV-compression gate, atomic two-leg entry and profit-take. Subclasses BOTH RollingOptionsStrategy and FactoryStrategy with the four abstract hooks re-declared (Layer-1 AST). A verification-only long-perp smoke trade on the Deribit primary proves order plumbing in Layer-3 (no schedule there), gated OFF whenever a real schedule exists so the analyzed backtest always trades real option legs. ~3 parameters (rv_lookback, rv_rank_max, rv_rank_lookback); pure OHLCV signal.

Correctly implemented long-vol exposure: net_entry_vega +5,309 (long), net_entry_theta -20,378 (paying decay), delta-neutral (-0.93), positive return_skew +1.23 — the genuine long-straddle profile, real greeks, 110 straddles, no accounting artifact.

Legitimate, distinct mechanism concept (vol-of-vol / GARCH mean-reversion) that would fill the options and neutrality gaps IF it were on the right side of the premium.

Loses every single year (2022 -11%, 2023 -4%, 2024 -9%, 2025 -26%, 2026 -11%; total -48.6%), Sharpe -7.55, profit_factor 0.62, expectancy -$225/trade, omega 0.30, PSR 0.0001.

Structurally on the WRONG SIDE of the variance risk premium: crypto IV persistently exceeds RV, so buying the straddle pays the premium the short-vol siblings harvest — theta bleed (paying $20k) overwhelms the occasional vol-expansion wins.

RV compression does not predict enough realized expansion to beat the IV-implied breakeven (the expected expansion is already priced into the IV paid); positive skew confirms wins exist but they don't cover the cost.

Loses across ALL regimes (bear 2022 and bull 2024/2025), so there is no robust parameter region; the 3 timing params (rv_lookback/rv_rank_max/rv_rank_lookback) cannot flip a persistent-premium loser to a winner.

positive

negative every year 2022-2026

positive across regimes

DeribitBtcWeeklyLongStraddleRvCompression bought weekly ATM BTC straddles after realized-vol compression, a clean delta-neutral vol-of-vol bet meant to fill the empty options bucket with a long-gamma vol-expansion edge. The implementation was correct — 110 real straddles with genuine long-vega/short-theta greeks and a positive payoff skew — but it sat on the wrong side of the variance risk premium: buying elevated IV and bleeding theta, it lost every year for a -48.6% total and -7.55 Sharpe. The analyst ruled it a structural wrong-sided premise, not an artifact, with no robust region for the three timing parameters to reach, and noted the correct expression is to sell the rich vol (as the short-vol siblings do). It ended after one iteration as abandoned, never advancing to optimization or risk review.

Long ATM straddles / long-vega vol-expansion bets on crypto systematically lose the variance risk premium: IV persistently exceeds RV, so paying theta for overpriced optionality and waiting for expansion that rarely beats breakeven loses in every regime, and entry timing cannot flip a persistent-premium loser — the Deribit volatility edge is on the SHORT-vol, defined-risk premium-selling side (as the VRP-harvesting siblings show), not the long-vol side.

It was abandoned at the pre-optimization backtest-review gate (verdict: abandon) as a structurally wrong-sided premise empirically confirmed by a correctly-implemented backtest — buying crypto volatility pays the persistent variance risk premium (IV > RV), so weekly theta bleed overwhelms the occasional expansion and RV compression doesn't predict enough realized expansion to beat the IV-implied breakeven — a loser in all regimes with no tunable region, so optimization and later stages were never reached.

A delta-neutral long-volatility options strategy on Deribit BTC weekly ATM options — each weekly cycle buying the next-expiring ATM straddle (1 ATM call + 1 ATM put, same strike/expiry) when realized volatility was compressed (RV-rank ≤ 0.30), betting on mean-reverting vol expansion (GARCH dynamics) as a pure vega/gamma play to fill the portfolio's near-zero options bucket, with an OHLCV-only RV signal and three timing parameters.

Across 110 straddles (220 option legs, correctly long-vol: net vega +5,309, net theta -20,378, delta -0.93, positive skew +1.23) it lost decisively in every year (2022 -11%, 2023 -4%, 2024 -9%, 2025 -26%, 2026 -11%): total return -48.6%, Sharpe -7.55, profit factor 0.62, expectancy -$225/trade, omega 0.30, PSR 0.0001, max drawdown 51%.