DeribitBtcWeeklyDrawdownCashSecuredShortPut

BTC Deribit Weekly Cash-Secured Put Sell at Significant 3-Day Drawdown (Vol Risk Premium Capture)

Single-leg short-vol options strategy on Deribit BTC weekly puts. When BTCUSDT spot drops >=8% over a rolling 3-day window, sell a 1-week ~10%-OTM put on Deribit and hold to expiry. The edge stacks three documented effects: (1) BTC tends to mean-revert after sharp short-horizon drawdowns (historically ~62% positive 7-day return after >=8% 3-day drops, 2017-2024), (2) implied volatility expansion during selloffs makes puts richly priced vs subsequent realized vol, and (3) the volatility risk premium — IV consistently exceeds RV in crypto by a wide margin, especially at fear extremes when retail panic-buys protection. This is intentionally the simplest possible options edge: ONE entry threshold, ONE strike offset, ONE tenor, hold to expiry. Fills the portfolio's zero-coverage options bucket (15% quota gap) with an OHLCV-only signal that does NOT depend on supplementary data (funding/OI/taker_ratio), avoiding the data-coverage blocker that has stalled 13+ recent hypotheses. Avoids classical TA pattern recognition (no candlestick shape rules, no breakout, no trend follower). Avoids long-short multi-instrument P&L complexity (single position at a time).

Intentionally the simplest possible options edge: ONE entry threshold, ONE strike offset, ONE tenor, hold to expiry. The edge stacks three documented effects: (1) BTC mean-reverts after sharp short-horizon drawdowns (historically ~62% positive 7-day return after >=8% 3-day drops), (2) implied vol expands during selloffs so puts are richly priced versus subsequent realized vol, and (3) the persistent crypto volatility risk premium (IV >> RV), especially at fear extremes when retail panic-buys protection. Selling the OTM put after the drop monetizes all three. Fills the portfolio zero-coverage options bucket (15% quota gap) with an OHLCV-only signal that does NOT depend on supplementary data, avoiding the data-coverage blocker that stalled 13+ recent hypotheses. Avoids classical TA pattern recognition and long-short multi-instrument P&L complexity (single position at a time). Genuine options exposure with real option-leg trades in the analyzed backtest (the smoke path fires only in verification). Cash-secured defined risk keeps it walk-forward-clean. leverage=1.0 (options venue cap); cash-secured sizing via risk_frac, no unused-leverage gate.

Wrong-sided premise (catching a falling knife), empirically refuted by a realistic-tail backtest. Selling a cash-secured put RIGHT AFTER an 8% 3-day drawdown bets on mean-reversion, but the ~38% of drawdowns that continue drive the put deep ITM, and those tail losses overwhelm the premium: total_return -66%, Sharpe -0.92, profit_factor 0.84, negative expectancy, with a catastrophic 2025 (-211% annual) including single days of -153% (2025-02-28) and -69% (2025-04-07). Critically, the engine IS correctly capturing the short-vol left tail here (avg_loss $335 > avg_win $210, largest_loss -$1,552 > largest_win +$860, return_skew -1.40) — so this is a genuine losing edge, not a modeling artifact. The hypothesis explicitly claimed to avoid the falling-knife error but the drawdown TRIGGER is precisely that error; the documented 62%-mean-reversion stat does not compensate for the magnitude of the 38% tail. There is also a risk-bounding failure: a -153% single-day equity move and 251% max_drawdown on a 'cash-secured, max-loss-bounded' 20%-notional put mean the secured-loss cap is being violated (a single put controls >1.5x equity of downside) — a sizing/accounting defect. Not iterate: the only fix is to STOP selling into drawdowns and instead sell in a non-bear regime, which is exactly the sibling strategy 4a121630 (Deribit short-put with IV/RV + 50-SMA bull filter), so a corrected version would be redundant; and a strategy losing -66% with a catastrophic blowup has no robust region to optimize toward. FAILURE PATTERN: cash-secured short BTC weekly put TRIGGERED BY a recent sharp drawdown is the falling-knife short-vol error — the mean-reversion premium is real but smaller than the realized left-tail when the selloff continues (2025), so the strategy carries genuine negative skew and loses; short-vol put selling must AVOID selling into downtrends (regime filter), not seek them out. The engine models the short-put ITM tail correctly, so the loss is real, not an artifact.

Single-leg defined-risk SHORT-VOLATILITY options strategy on Deribit BTC weekly puts, triggered by a price drawdown and OHLCV-only. The backtest rolling-options expander resolves the weekly BTC chain from parameters[rolling_options] (underlying BTC, weekly roll, 7 DTE, 1-DAY leg, ~10% OTM put at put_offset). On 1-DAY bars the strategy tracks the rolling drawdown = close_now / close_{drawdown_days ago} - 1; when BTC has dropped >= drawdown_pct (8%) over the rolling drawdown_days (3) window it SELLS one ~10%-OTM weekly put (cash-secured) and HOLDS IT TO EXPIRY, closing only at the roll buffer just before expiry (or a loose catastrophe premium-loss stop). One open short put at a time. Sizing is cash-secured: it reserves risk_frac (20%) of equity as the secured notional (via the base equity*risk_frac/spot sizing), so worst case is BTC->0 with max loss = strike - premium, never a margin blowout. The base RollingOptionsStrategy owns schedule/entry/roll; this subclass adds the drawdown gate and the one-at-a-time constraint. 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 the Layer-3 sandbox (no schedule there) and is gated OFF whenever a real schedule exists, so the analyzed backtest always trades real option legs. NO supplementary data (funding/OI/taker). 3 economic parameters (drawdown_pct, drawdown_days, put_offset).

Real, greeks-correct short-put options exposure (28 puts, net_entry_theta +797, net_entry_vega -125, avg_moneyness 1.05) — the plumbing works and, importantly, the engine correctly captures the asymmetric short-vol tail here (losses > wins, skew -1.40).

Simple, OHLCV-only, low-parameter design that would fill the 0% options bucket if the premise worked.

Wrong-sided premise: selling puts INTO an 8% 3-day drawdown is catching the falling knife — total_return -66%, Sharpe -0.92, profit_factor 0.84, negative expectancy. The mean-reversion-after-drawdown claim does not survive the tail losses when drawdowns continue.

Catastrophic 2025 (-211% annual) with single days of -153% (2025-02-28) and -69% (2025-04-07) as post-trigger selloffs kept falling and the puts went deep ITM — the realistic short-vol left tail dominates the premium.

Risk-bounding violated: a -153% single day and 251% max_drawdown on a supposedly cash-secured 20%-notional put mean the max-loss cap is not holding (a single put controls >1.5x equity of downside) — a sizing/accounting failure.

The correct version of this idea (sell puts in a bull regime, not into a drawdown) already exists as sibling 4a121630, so a fixed version would be redundant.

single-day -153%, 2025 -211%

bounded cash-secured loss

positive

DeribitBtcWeeklyDrawdownCashSecuredShortPut sold weekly OTM BTC puts on Deribit triggered by an 8% 3-day drop, betting on mean reversion plus the volatility risk premium to fill the empty options bucket. The plumbing worked — 28 real, greeks-correct short puts — but the premise was backwards: selling into the drawdown caught falling knives, and when selloffs continued the puts went deep ITM, producing a -66% return, -0.92 Sharpe, negative skew, and a -211% 2025 with a -153% single day that also exposed a broken cash-secured loss cap. The analyst ruled it a genuine wrong-sided edge (not a modeling artifact, since the engine captured the short-vol tail correctly), un-iterable because the only fix — selling in a non-bear regime — already exists as a sibling strategy. It ended after one iteration as abandoned, never advancing to optimization or risk review.

Selling puts INTO a sharp drawdown is the classic falling-knife short-vol error: the 62% post-drop mean-reversion stat does not compensate for the magnitude of the 38% continuation tail, so the strategy carries genuine negative skew and loses — short-vol put selling must AVOID downtrends via a regime filter (as the existing bull-filtered sibling does), not seek them out, and a single put must be sized so its ITM tail truly stays cash-secured.

It was abandoned at the pre-optimization backtest-review gate (verdict: abandon) as a wrong-sided, falling-knife premise empirically refuted by a realistic-tail backtest — the ~38% of drawdowns that keep falling drive the puts deep ITM and overwhelm the mean-reversion premium — compounded by a sizing/accounting defect that breaks the cash-secured cap, so optimization and all later stages were never reached.

A single-leg short-vol options strategy on Deribit BTC weekly puts — when BTCUSDT spot dropped ≥8% over a rolling 3-day window, selling a 1-week ~10%-OTM cash-secured put (secured notional 20% of equity) and holding to expiry — an OHLCV-only, three-parameter design meant to capture the volatility risk premium and post-drawdown mean reversion while filling the portfolio's zero-coverage options bucket.

Across 28 real put trades (greeks-correct: net theta +797, net vega -125, avg IV 53%, avg moneyness 1.05) it had a genuine losing edge with a dominant left tail: total return -66%, Sharpe -0.92, profit factor 0.84, expectancy -$23.70/trade, avg_loss ($335) > avg_win ($210), return skew -1.40, and a catastrophic 2025 (-211% annual) with single days of -153% (2025-02-28) and -69%, plus a risk-bounding failure (251% max drawdown / -153% single day on a supposedly cash-secured 20%-notional put).