Crypto Convexity | Recommendations

The decision, reframed. The old call was "restructure at $4M to escape a ~−20% crypto forced-liquidation." The crash analysis this session changed that premise: your crypto tail is not the liquidation threat (it survives a crypto-only −50%), and the real force-liquidation risk is a broad correlated crash on Main alone. So the sharper move is a sequence, not a single flip: de-risk Main's margin drivers now, give the existing book time to 31 Dec to keep its front-loaded upside, and complete the switch to the 2b Fortress at $4M / year-end, sized to clear the PM-floor.

WHAT THE CRASH ANALYSIS CHANGED

The crypto tail is NOT the liquidation threat. The exact per-account TIMS-race model (fresh live run) shows a crypto-only −50% crash — every crypto name halved, S&P and the rest of the book flat — force-liquidates NOBODY. Main −33% NLV, Joint −34%, Retired −33%, Neville −5%, all safe past −50%. The non-crypto half of each account props up NLV and margin. (See the Crash Liquidation tab.)
The real liquidation risk is a BROAD correlated crash, and it lives on Main. Under a market-wide selloff (S&P −30%, which drags the crypto/miner names to −85%+ AND the semis to −50% AND the index to −30% together), Main force-liquidates (~S&P −30%, PM floor ~−37%). Joint / Neville / Retired survive that too. So the exposure is a leverage/concentration question on Main, not a crypto-tail question.
This lowers the urgency the old page assumed. The prior "restructure now to escape a ~−20% crypto forced-liquidation" premise is superseded: that −20% was the broad-market number, and crypto in isolation is survivable to −50%. There is no crypto-crash emergency forcing an immediate switch.

THE THREE PATHS TO DEC 31

Modeled at 31 Dec 2026 (BS, time value on; moves beta-mapped per name; crash gets a 1.6× vol spike). P&L is on the crypto sleeves; the liquidation verdict is from the exact per-account model. "Hold all" = keep the current book intact; "De-risk" = close the margin-driving short puts now but keep shares + LEAPS + owned hedges; "Switch" = move to the 2b Fortress now.

Scenario to Dec 31	Hold all	De-risk now	Switch now	Read
Crypto crash (−50%, S&P flat)	−$657k mark · survives	−$220k · survives	−$338k · survives	All three SURVIVE (per the exact per-account model). Hold-all takes the biggest crypto-sleeve mark; de-risk the smallest.
Flat (0%)	−$96k	−$138k	−$146k	Near a wash. Holding bleeds the least theta; waiting costs almost nothing.
Crypto rally (+50%, S&P flat)	+$677k	+$357k	+$444k	Waiting wins: hold-all captures ~$233k more than switching now (shares participate from $0).
Broad crash (S&P −30%)	Main FORCE-LIQUIDATES	removes Main's driver	survives (defined-risk)	The ONE place hold-all fails. Closing Main's margin-driving short puts pushes its liquidation point out; switching removes it entirely.

Waiting wins the rally (~$233k over switching) and costs almost nothing flat. In a crypto-only crash all three survive — the only scenario where holding fails is a broad correlated crash on Main, which de-risking (or switching) removes. So the case for waiting is real, and the case for de-risking Main is separable from the crypto switch.

RECOMMENDATION — sequence it, don't flip it

De-risk Main NOW. Close the margin-driving short puts on Main (the legs that create its broad-crash liquidation), keeping the shares, LEAPS and owned hedges. This is the single move that attacks the only real force-liquidation exposure, and it has the best crash mark (−$220k vs −$657k hold-all). It is independent of the crypto-switch decision.
Give the rest of the book time to 31 Dec. The crypto tail survives a −50% crypto-only crash, so there is no emergency to switch. Holding keeps the front-loaded participation (shares from $0) that out-earns the deep-OTM 2b through the mid-range — worth ~$233k in a +50% crypto rally, near-free if flat.
Keep every hedge rolled. The crypto-crash survivability and the plan's 0.94× safety both depend on OWNED, live hedges. A late-window crash after the near-term (2026) hedges lapse is worse — the margin guide's rented-hedge warning. Roll them.
Complete the switch at $4M OR 31 Dec, whichever comes first. Redeploy into the income core + 2b Fortress. Fund Neville to $680,000 so ML/cash = 0.93× (cushion $47,441) — NOT $650k. At $650k the terminal floor is only ~$17k, below the ~$100k PM cliff; 0.93× clears it (see the margin guide).

WHY WAITING CAPTURES UPSIDE

From the value ladders: the existing book is front-loaded (shares + shallower/ITM LEAPS participate from the first dollar) while the 2b Fortress is back-loaded (deep-OTM, most value in the top half of the move). Both reproduce ~$26M at the end-2027 targets; the existing book's lead peaks at ~+$4M around the 10–20% rungs and fades to ~zero at the targets. So switching early forfeits the mid-range participation for no gain at the extreme targets — the reason to hold through the near term and only switch once you are near a completion trigger.

THE EXIT PLAN — CLOSE MARA AT $40 / BMNR AT $70, RE-ENTER ON THE JAN-2029 CHAIN

The MARA and BMNR sleeves are legacy BBC-format constructions that pre-date the current fortress architecture; the plan retires them at one decisive exit each and re-enters through the modern stack. At MARA $40 and BMNR $70, close each sleeve ENTIRELY (shares, LEAPS — sold, never exercised — short puts, CCs). By then the Jan-2029 LEAPS should be listed (January LEAPS list ~Sep–Nov two years ahead), so the OTM strikes AND the targets get revisited FRESH at live marks — nothing pre-committed from today's mental-math projections. Redeploy a small slice (~10–15%) into the new moonshots; move the bulk to the safer lanes.

Name	Exit price	Sleeve value there	Profit vs today	Redeploy (~10–15%)	To safer lanes (~85–90%)
MARA	$40 (3.6x spot)	$1,480,000	$1,367,400	$148,000–$222,000	$1,258,000–$1,332,000
BMNR	$70 (3.2x spot)	$2,511,000	$2,023,659	$251,100–$376,650	$2,134,350–$2,259,900
BOTH		$3,991,000	$3,391,059	$399,100–$598,650	$3,392,350–$3,591,900

THE EXIT, STEP BY STEP

Exit: the name touches its level (MARA $40 / BMNR $70; each name independent) → close the ENTIRE legacy sleeve. Sell the LEAPS (capture time value), buy back the short puts/CCs. Singapore: tax-free. Stage over days if spreads are wide.
Re-derive, don't re-use: on the Jan-2029 (or longest liquid) chain, re-run leap_builder.py --forecast --tickers {TICKER}:{budget}:target={fresh target}@{fresh date} --structure pure --port 7496 at live marks. New spot, new IV, new tenor → new OTM strikes and honestly revisited targets — the current architecture's discipline, not the BBC-era's.
Size the moonshot slice small: ~10–15% of proceeds, calls-only (premium = entire ML, un-liquidatable; a put-spread re-check via leap_builder is fine if IV is spiked). The slice is money you can lose entirely without touching the harvested gains.
Bank the bulk (~85–90%) into the safer lanes, working: the restructure plan's cash targets first (Neville 2b reserve, Retired index core), then owned-hedge income fortresses / the SPY–QQQ DIAMOND ladder — at the margin guide's rates the banked ~$3.6–3.9M earns roughly $25–65k/mo while it sits.
Re-run the risk stack after each exit (risk_cockpit --dump-legs + the Crash Liquidation tab): each exit should visibly push Main's broad-crash liquidation point further out.

The one tradeoff owned, eyes open: a single exit banks nothing if the name retraces BEFORE its level (MARA from $35 back down, say) — that path risk is the price of simplicity vs the analyzed-and-set-aside $1M-increment ratchet (which banked earlier but re-struck repeatedly and gave up most of the tail; it remains an option if the ride to $40/$70 gets uncomfortable). Exits at 3.6x/3.2x spot also sit where the front-loaded legacy book has captured most of its edge over any deep-OTM re-entry — selling the front-load high, re-buying convexity cheap and long.

TRIGGERS & GUARDRAILS

Main's broad-crash distance (Crash Liquidation tab): today Main liquidates around a market-wide S&P −30%. If that tightens (leverage creeps up, or a selloff starts), de-risk Main immediately — do not wait for Dec 31.

Crypto-only cushion: the book is safe to a crypto-only −50%. If you ever see the crypto-only crossover appear inside −50% on the tab, that is a new, real signal to act.

$4M is a proxy, Dec 31 is a checkpoint — complete on whichever hits first; if crypto rips past $4M early, still redeploy (you can lean slightly more OTM entering higher).

Exit levels (section above): MARA $40 / BMNR $70 → close that legacy sleeve entirely, redeploy ~10–15% into fresh Jan-2029 moonshots via leap_builder, bank the rest. Per name, independent of the Dec-31 / $4M completion.

Slippage budget: price the exit on the illiquid crypto legs before committing; stage the close over days if spreads are wide.

Zero tax cost (Singapore: no CGT, no wash-sale) — the switch is nearly frictionless whenever you choose to complete it, so the timing is a pure upside-vs-safety call, not a tax call.

THE EVENTUAL RESTRUCTURE — $4M SPLIT & CASH MOVEMENTS

Illustrative (not exact), for when you complete the switch. The current book (~$3,652,210 today) rises to $4.0M mainly on crypto uplift, then redeploys into the target splits. Neville funded to $680,000 so the 2b Fortress ML ($632,559) sits at 0.93× cash — below the ~$100k PM-floor cliff with room (cushion $47,441).

Step 1 — getting to $4M (crypto uplift)

Step 2 — redeploy to target splits (cash movements)

Account	NLV today	Crypto uplift	$4M-scenario NLV
Main	$1,743,947	+$173,895	$1,917,842
Neville	$702,537	—	$702,537
Joint	$804,974	+$121,727	$926,701
Retired	$400,752	+$52,169	$452,920
TOTAL	$3,652,210	+$347,790	$4,000,000

Account	$4M-scenario NLV	Target	Cash movement	Then holds
Main	$1,917,842	$2,000,000	+$82,158 in	$2.0M → 18-name income core
Neville	$702,537	$680,000	-$22,537 out	$680k cash → 2b Fortress (0.93× ML/cash)
Joint	$926,701	$320,000	-$606,701 out	$320k USD cash reserve
Retired	$452,920	$1,000,000	+$547,080 in	$1.0M → index Book-B
TOTAL	$4,000,000	$4,000,000	$629,238 in / $629,238 out	net zero

MECHANICS OF THE MOVE

Close the current book to cash across accounts (Singapore: tax-free), and repay Joint's ~$332,853 USD margin loan as part of unwinding Joint.

Joint sends the most out (~$606,701); it is the over-funded account and ends as a $320,000 USD cash reserve. Retired receives the most (~$547,080) to stand up the $1.0M index core.

Neville: hold $680,000 cash, deploy ~$136,509 net into the 2b Fortress; the rest is the defined-risk reserve keeping ML ($632,559) at 0.93× cash. (The moonshot version would need only ~$439,609; you chose the fortress.)

Then deploy Main $2.0M into the 18-name income core and Retired $1.0M into index Book-B — both at 0.94× ML/cash with owned hedges (stress crossover "none").

Disclaimer. This is a structured synthesis of your own analytical model for your own decision-making, not personalized investment advice, and I am not a licensed financial advisor. Numbers are BS const-IV estimates off specific live snapshots (accounts 2026-08-25, crypto legs 2026-08-22); re-price in RTH and validate against the broker (Risk Navigator) before acting.

RECOMMENDATIONS — DE-RISK NOW, GIVE IT TIME TO DEC 31, THEN SWITCH