Vol-Structure Deep-Dive — what we tested and why none beat the book

After the Grand Book was assembled, we went deeper into the option-structure space looking for a better single sleeve or a cheaper hedge. Four candidates, each tested properly (BTC+ETH, M2 2024-26, by-year, cost-vs-premium). All four are negatives. That is the finding: the BTC/ETH weekly vol book is mined out — the Grand Book (decorrelated sleeves) is the endpoint of vol-structure optimization, and further gains need a different lever (funding/basis carry, new instruments) or execution. This page documents the verdicts so they are not re-tested.

1. Jade Lizard — sell put + bear-call-spread (no upside risk if credit ≥ width)

Idea: we measured that the crypto weekly tail is TWO-SIDED and even UP-skewed (BTC: 13 of the 20 worst strangle weeks were up-moves). The jade lizard removes the upside tail by capping the call side with a long call — keep put-side premium, no up-crash risk.

BTC variant	credit%	%no-up-risk	r/DD	worst	2024/25/26
best (−3%/+2%/3w)	2.48	18%	2.17	−16.19	+36/+23/+3
plain strangle (ref)	—	—	2.41	−16.78	+28/+31/+14
futures-stopped strangle (our book)	—	—	7.2	−14.98	+61/+54/+16

Verdict ❌: worse than the plain strangle (r/DD 2.17 vs 2.41). "No upside risk" only holds 18% of weeks for the best variant — to guarantee it you need a tight call spread that gives back the call premium, and the worst weeks are PUT-driven anyway (down-moves blowing the naked put) which the call structure can't touch. The futures-stop@breakeven dominates everything (r/DD 7.2) because it is two-sided and doesn't pay premium for the hedge.

2. ETH calendar pendant — does CAL_BTC_MA replicate on ETH?

Idea: the BTC 28/14 double-calendar on MA_CND_CROSS signals was the sprint's one real diversifier (r/DD 18.3, corr ~0). Does an ETH version add a SECOND orthogonal sleeve?

ETH 28/14 calendar	r/DD	MaxDD	corr to book	2024/25/26
BLIND weekly	1.9	−23.6	—	+3/+15/+25
MA_CND_CROSS signals	1.3	−54.0	+0.36	+11/+21/+37
EXPANSION / STOCH signals	2.1 / 1.1	−26 / −45	+0.13 / +0.10	—
BTC CAL_BTC_MA (for comparison)	18.3	−9.3	~0	+74/+65/+30

Verdict ❌ (but informative): no second orthogonal sleeve — ETH calendars are r/DD ~1-2 with −26 to −54% drawdowns. The decorrelated ones (EXPANSION/STOCH, corr ~0.1) are far too weak; the one with return (MA) has DD −54 and corr +0.36. The BTC calendar's edge is the BTC-specific 2-week pinning — it does NOT replicate on ETH. This confirms CAL_BTC_MA is a real asset-specific find, not a generic calendar artifact.

3. Broken-Wing Butterfly — put-side credit BWB

Idea: a net-credit, defined-risk, cheap-convex put structure (long Kh / short 2 Km / long Kl, asymmetric wings).

Verdict ❌: only one config was a positive credit (BTC −2%/−4%/−7%: r/DD 1.66, +11.2 total) — but 2025 and 2026 were both negative, and it is worse than every sleeve already in the book. All other configs were net debit and lost money. A debit fly needs price to pin at the body; a credit BWB carries the tail. No edge over what we have.

4. Two-sided Standing Tail — the conceptual result

Idea: buy a small far-OTM BTC put + call every week as a permanent book-level crash hedge (the two-sided version of the 3a put overlay, fixing the "crypto tail is two-sided" problem).

Overlay on Grand Book	Total	MaxDD	worst	Δ vs no-hedge
BASE (no tail)	+75.6	−2.0	−1.93	—
+ 10% OTM tail @0.1%/wk	+74.9	−2.0	−2.03	dTot −0.7, dDD 0, dWorst −0.10
+ 10% OTM tail @0.3%/wk	+73.5	−2.2	−2.23	dTot −2.1, dDD −0.2, dWorst −0.30

Verdict ❌ — and this one teaches something: the tail hedge makes everything slightly WORSE. The book is already too smooth (worst week −1.93%) for a crash hedge to pay — there is no fat tail left to insure, so the hedge is pure bleed. The decorrelation already did the tail-removal job. This is the same conclusion as our 3a finding (one-sided put hedge bled) and the AQR literature (mechanical long-option floors are poor DD reducers), now shown at the portfolio level: a well-diversified short-vol book does not want a standing option hedge — it wants more decorrelated sleeves.

Bottom line

Four structures, four negatives — none beats the futures-stopped strangle or adds an orthogonal sleeve. The honest conclusion: the BTC/ETH weekly vol-structure space is exhausted for our purposes. The Grand Book (decorrelated sleeves + the BTC pinning calendar) is the endpoint of vol optimization. Real further upside needs a different return source (perp funding / basis carry — orthogonal to vol), new instruments (SOL/alt options for richer IV), or moving from research to execution (the live weekly + overnight paper book). All numbers M2 2024-26, indicative levels; the verdicts (relative ranking) are the robust part.

Sources: run_sprint_jade.py, run_sprint_eth_cal.py, run_sprint_bwb_tail.py, full log SPRINT_LOG.md.