INCOME BOOK: $100K/MO AT REALIZED RATES FROM AN UNCORRELATED, FULLY WRITTEN BASKET · 8 FORTRESSES · $2,950,000 ACCOUNT
Chain marks 2026-09-22 SGT pre-market (fortress_v28 CSV replay, no live IBKR; Jan-2029 LEAPS; run dir INCOME_BOOK) · horizon to LEAPS expiry ~2.3 yr · sibling: 3-account plan
SNAPSHOT
The question: what is the most uncorrelated basket that really pays $100k a month with every contract written? "Really" means priced at what your calls have actually earned, not at model premium: each name's income is scaled by its realized-to-model ratio from the trade log (next section). An optimizer then picked names and whole-contract counts to minimize max loss plus a penalty on the book's daily dollar P&L swing, with every name between 4% and 15% of max loss. The account is sized so ML = 0.93x cash, the no-liquidation calibration.
Income Book: Income / Month (CC + cash)
$103,041
CC $100,037 + cash int $3,004 · 221 written / 0 uncapped
Income Book: Total P&L at Expiry
$4,235,463
structure gains + net income · 62%/yr on cash
Income ROI on cash
42%/yr
on Invested 66% · on Max Loss 45%
Hedge coverage
7.3x
kept CC $1,389,528/yr vs rent $189,090/yr
Invested / Max Loss
$1,883,463
ML $2,738,463 = 0.93x cash · NLV $2,950,000
Notional Exposure
$5.66M
1.9x on the account
Structure P&L @ Expiry
$1,442,065
all legs at +10%/yr drift, ~2.3 yr
Idle cash
$1,066,537
$3,004/mo @ 3.38% (bench 3.88% −0.5%)
Hedge rent / yr
$189,090
billed on ALL contracts in both scenarios
Liquidation Crossover
none (survives >-95%)
ML ≤ cash · crash-vol, TIMS-race
Daily P&L swing
2.62%
of the account (1 sd, 180 days) · the previous book was 2.31%, but earns only $51,130/mo at realized rates
Avg pairwise correlation
+0.03
180 days · -0.01 over 5 years, +0.01 on crash days
Model would have said
$172,904/mo
the same book at model premium; realized basis $100,037/mo
WHY THE INCOME IS CALIBRATED: WHAT YOUR CALLS HAVE ACTUALLY EARNED
From the Trade Query snapshot: every call closed since 2026-06-01, realized yield on the notional it was written against (qty x strike), divided by the yield the model assumes for the same name. Across the list the book has realized 52% of model. A name with few trades is pulled toward that average (as if it had 10 more trades at the average), and no name is credited above its model. Names with no calls in the window (XOM, LLY, LMT here) use the average.
Name
Calls closed since Jun 1
Realized ÷ model (raw)
Ratio used
Model income / mo
Realized-basis income / mo
Basis
MU
30
x0.85
x0.77
$38,629
$29,427
own tape (30 trades)
META
8
x0.26
x0.40
$31,250
$11,914
thin tape (8), pulled toward the book average
HOOD
2
x1.09
x0.62
$29,453
$17,255
thin tape (2), pulled toward the book average
XOM
0
—
x0.52
$21,287
$10,126
no tape: book average 0.52
AMZN
9
x0.88
x0.69
$19,117
$11,758
thin tape (9), pulled toward the book average
SPCX
15
x0.95
x0.78
$16,421
$12,446
own tape (15 trades)
LLY
0
—
x0.52
$11,261
$4,639
no tape: book average 0.52
LMT
0
—
x0.52
$5,486
$2,471
no tape: book average 0.52
THIS BOOK
$172,904
$100,037
THE TRADE-OFF: MAX LOSS vs DIVERSIFICATION (all rows pay $100k/mo at realized rates)
Moving down the table buys a calmer book with more max loss. This book is the knee. The previous income book is shown for reference: at realized rates it pays about half its old headline.
HOOD 45, XOM 61, SPCX 39, MU 5, META 8, AMD 7, AMZN 22, SNDK 1
Lower-cost diversified
less max loss, more daily swing
$2,516,477
$2,710,000
$100,003
9
+0.09
3.05%
-3.91%
META 9, XOM 66, HOOD 47, SPCX 40, MU 5, AMZN 24, LLY 4, AMD 3, SNDK 1
True $100k, uncorrelated (this book) ✓
the knee: the big step in diversification
$2,738,463
$2,950,000
$100,037
8
+0.03
2.62%
-3.61%
XOM 72, MU 6, AMZN 37, HOOD 51, META 9, LLY 6, SPCX 31, LMT 9
More diversified
more max loss for a smaller further gain
$2,818,214
$3,040,000
$100,047
8
+0.03
2.51%
-3.54%
XOM 74, META 10, AMZN 38, MU 6, LLY 7, HOOD 45, SPCX 28, LMT 11
Most diversified
the next step: more max loss for a small further gain
$2,890,620
$3,110,000
$100,018
9
+0.06
2.40%
-3.64%
XOM 76, META 10, MU 6, AMZN 37, LLY 7, HOOD 40, LMT 16, SPCX 16, AMD 3
Previous income book
the book this page showed before, at realized rates
$1,677,900
$1,810,000
$51,130
8
+0.02
2.31%
-3.53%
META 6, XOM 43, GOOG 17, PLTR 21, LLY 4, MU 3, DELL 5, LMT 8
WHY IT IS UNCORRELATED
Each cell answers one question for a pair of stocks: when one has a good or bad day, does the other tend to do the same? Measured on daily returns over the last 180 trading days, from −1 (always opposite) through 0 (unrelated) to +1 (always together). Each pair appears once. The second table shows how much of the book's daily P&L risk each position actually carries; a name well below its size is doing diversification work.
XOM
MU
AMZN
HOOD
META
LLY
SPCX
MU
-0.22
AMZN
-0.30
+0.09
HOOD
-0.26
+0.19
+0.24
META
-0.28
+0.07
+0.42
+0.29
LLY
-0.03
-0.11
+0.02
+0.12
-0.04
SPCX
-0.21
+0.15
+0.24
+0.20
+0.09
-0.04
LMT
+0.12
-0.08
-0.05
+0.01
-0.04
+0.05
+0.25
−0.3 move opposite0 unrelated+0.3 move together (±1 would be perfectly opposite / identical)
Move together: AMZN and META (+0.42), HOOD and META (+0.29). Even these are moderate; closely related stocks often sit at +0.7 or higher.
Move opposite: XOM against AMZN (-0.30), META (-0.28), HOOD (-0.26), MU (-0.22), SPCX (-0.21). On days the tech names sell off, XOM has tended to hold up, so it partly cancels the rest of the book.
Independent: LLY: no correlation stronger than ±0.15 with anything in the book.
Every other pair sits between -0.11 and +0.25: largely unrelated day to day. Average across all 28 pairs: +0.03.
Name
Share of max loss
Share of daily P&L risk
Avg correlation to the other 7 names
Reading
XOM
15.0%
-4%
-0.17
offsets the rest of the book (negative risk share)
MU
14.7%
24%
+0.01
carries more risk than its size
AMZN
14.7%
16%
+0.10
roughly in proportion
HOOD
14.6%
25%
+0.11
carries more risk than its size
META
13.8%
12%
+0.07
roughly in proportion
LLY
11.3%
5%
-0.00
diversifier: far less risk than its size
SPCX
9.8%
20%
+0.10
carries more risk than its size
LMT
6.1%
3%
+0.04
diversifier: far less risk than its size
DOES IT HOLD UP IN A REAL SELLOFF? (5 years of daily history)
The correlations above come from 180 mostly ordinary days. This replays the book over the selloffs that matter: correlation measured only on the market's worst 10% of days, the market's 20 worst single days, and three named episodes. Diversification shows up in the slow grind (2022). On sharp one-day crashes every book loses about the same share of its account, because everything falls at once; there the protection is the hedge puts and max loss below cash, not diversification.
Book
Avg correlation all days (5y)
Avg correlation on crash days
SPY's 20 worst days (avg day, % of account)
2022 bear Jan 3 – Oct 12
Aug 2024 unwind Jul 31 – Aug 5
Apr 2025 tariff crash Apr 2 – Apr 8
This book
-0.01
+0.01
-7.8%
-11%*
-12%*
-26%*
Previous income book
+0.18
+0.14
-8.1%
-20%
-8%
-25%
Cheapest true $100k
+0.21
+0.16
-7.2%
-48%*
-11%*
-27%*
SPY itself
-3.6%
-24%
-5%
-11%
* SPCX (listed Jun 2026) and SNDK (Feb 2025) had no prices for that episode and count as zero, so the cell is understated. Losses are LINEAR (each position's dollar delta x the stock's move): the hedge puts' extra payoff in a big drop is not credited, so real losses would be smaller. Read it as a comparison between books, not a forecast.
Left: market exposure (leveraged notional by theme). Right: the same dollars un-leveraged: deployed invested capital by theme plus idle cash, summing to the account NLV, so the grey Cash slice is its TRUE share of capital (36% of the account, earning ~$3,004/mo at 3.38%).
THE BOOK, ALL CAPPED (8 fortresses)
Net income = gross CC on every contract x keep rate x the name's realized ratio, minus the hedge rent. Contract counts are the calibrated frontier's knee, pinned.
Ticker (marks)
Type
Contracts (inc / unc)
Spot
Spot @ Expiry (+10%/yr)
LC / SP / HP
Invested
Max Loss (share)
Notional
Lev
Net Δ/sh (ex-CC)
Structure P&L @ Expiry
Net Income / Mo
Net Income to Expiry
Total P&L
ROI/yr IC
ROI/yr ML
ROI/yr Notl
XOM
09-22
BALANCED
rented hedge
72
all capped
157.50
196.64
120 / 145 / 125
$266,688
$410,688
15.0% of book
$1,134,000
4.25x
1.04
$285,129
$10,126
keep 68% · realized x0.52
$282,754
$567,883
91%
59%
22%
HOOD
09-22
BALANCED
rented hedge
51
all capped
121.96
152.27
55 / 90 / 75
$323,799
$400,299
14.6% of book
$621,996
1.92x
1.04
$172,273
$17,255
keep 68% · realized x0.62
$481,836
$654,109
87%
70%
45%
AMZN
09-22
BALANCED
rented hedge
37
all capped
260.30
324.99
195 / 260 / 215
$236,800
$403,300
14.7% of book
$963,110
4.07x
1.08
$244,158
$11,758
keep 73% · realized x0.69
$328,331
$572,489
104%
61%
26%
SPCX
09-22
BALANCED
rented hedge
31
all capped
152.49
190.39
95 / 130 / 100
$174,716
$267,716
9.8% of book
$472,719
2.71x
1.05
$120,981
$12,446
keep 74% · realized x0.78
$347,541
$468,522
115%
75%
43%
LMT
09-22
BALANCED
rented hedge
9
all capped
537.24
670.75
440 / 520 / 430
$85,545
$166,545
6.1% of book
$483,516
5.65x
1.08
$122,132
$2,471
keep 68% · realized x0.52
$68,989
$191,121
96%
49%
17%
META
09-22
BALANCED
rented hedge
9
all capped
743.08
927.75
400 / 590 / 510
$305,145
$377,145
13.8% of book
$668,772
2.19x
1.08
$169,827
$11,914
keep 76% · realized x0.40
$332,685
$502,512
71%
57%
32%
MU
09-22
BALANCED
rented hedge
6
all capped
1033.72
1,290.62
500 / 730 / 590
$319,590
$403,590
14.7% of book
$620,235
1.94x
1.06
$154,783
$29,427
keep 77% · realized x0.77
$821,722
$976,505
131%
104%
68%
LLY
09-22
BALANCED
rented hedge
6
all capped
1155.99
1,443.27
870 / 1150 / 920
$171,180
$309,180
11.3% of book
$693,594
4.05x
1.09
$172,783
$4,639
keep 68% · realized x0.52
$129,540
$302,322
76%
42%
19%
INCOME BOOK (8 fortresses)
221
221 inc / 0 unc
$1,883,463
$2,738,463
0.93x cash
$5,657,942
—
—
$1,442,065
$100,037
$2,793,398
$4,235,463
97%
66%
32%
MARK-TO-MARKET vs DRAWDOWN
Unrealized P&L if every underlying is haircut the same % (crash-vol adjusted). A uniform shock cannot show diversification (every name falls together); it measures structure. Not force-liquidated anywhere in this range (crossover none (survives >-95%)).
Underlying drawdown
0%
-10%
-20%
-30%
-40%
-50%
Income Book
$0
0% NLV
-$557,969
-19% NLV · 1.9x
-$1,007,507
-34% NLV · 1.7x
-$1,363,859
-46% NLV · 1.5x
-$1,646,262
-56% NLV · 1.4x
-$1,868,681
-63% NLV · 1.3x
The x-multiplier is the NLV move divided by the underlying move (effective downside leverage). It falls as the drawdown deepens (1.9x at −10% → 1.3x at −50%): the hedge puts gain and the structure self-de-levers, which is the convexity that bends the curve flat.
MARK-TO-MARKET vs RALLY
The cost of writing everything: each name's structure gain is capped near one CC cycle. The pale line is what the raw structures would make uncapped.
▬ All capped▬ Uncapped ceiling
Underlying rally
+10%
+25%
+50%
+75%
+100%
All capped
$480,456
+16% NLV · 1.6x
$480,456
+16% NLV · 0.7x
$480,456
+16% NLV · 0.3x
$480,456
+16% NLV · 0.2x
$480,456
+16% NLV · 0.2x
Uncapped ceiling
$601,843
+20% NLV · 2.0x
$1,514,895
+51% NLV · 2.1x
$3,018,952
+102% NLV · 2.0x
$4,495,375
+152% NLV · 2.0x
$5,952,050
+202% NLV · 2.0x
Every contract is written, so each name's structure gain is limited near one CC cycle (~+8%). That is the price of the income: at +50% this book keeps $480,456 of structure gain where the raw structures would make $3,018,952 (+100%: $480,456 vs $5,952,050). Rolling calls up-and-out on a melt-up moves you toward the ceiling at the cost of income.
STRESS BATTERY vs THE MARGIN & LIQUIDATION FIELD GUIDE (module_risk_cockpit, crash-vol + skew, 1% steps to −95%, run 2026-09-25)
~$100k PM-floor race (NLV below the portfolio-margin floor)
not within −60%
Verdict: on every price path to −95%, under crash vol, a tripled house requirement and twelve months of decay, the account cannot be margin-called (7/7 attacks: NONE).
Hedge rent: $189,090/yr against $1,389,528/yr of kept premium at realized rates = 7.3x coverage. The hedges must never lapse; if they do, the true max loss is the full notional $5,657,942.
WHAT THIS PAGE SAYS, IN PLAIN ENGLISH
The answer: a $2,950,000 account running 8 fully written fortresses pays $103,041/mo at the rates your calls have actually earned (CC $100,037 + cash interest $3,004) = 42%/yr on cash, with a defined worst case of $2,738,463 and a liquidation crossover of none (survives >-95%).
Why it is bigger than the old page: the old page priced income at model premium, which your trade log says you realize only about half of. At realized rates the old $1.68M book paid about $51,130/mo, so a real $100k needs roughly 1.6x the max loss.
What made it this mix: the names whose realized income comes closest to the model moved up (MU, AMZN, SPCX, HOOD); GOOG and DELL fell away because they realize under half of model, and PLTR, with no calls on the tape, is priced at the book average and no longer earns its seat. XOM is the largest position by contracts and still the book's main offset; LLY and LMT move independently of everything.
What it still is: an all-capped income book. Every name's upside stops near one cycle, and a sharp market-wide crash hits it about as hard as any other book of this size.
ASSUMPTIONS AND CAVEATS
Marks: fortress_v28 CSV replay (no live IBKR) of the 2026-09-22 17:53 SGT chain export, i.e. pre-market frozen quotes, Jan-2029 LEAPS. Re-price every leg live in RTH and price whatIf margin before ticketing.
Realized ratios come from calls closed since 2026-06-01, a few months of one regime and one execution style. HOOD rests on 2 trades and SPCX on 15; XOM, LLY and LMT have none and use the book average. If your writing changes, the ratios should be rebuilt.
SPCX has only a few months of price history: part of why the optimizer likes it is that its short record shows low correlation, so treat its seat as provisional.
Correlations and the daily swing come from 180 trading days of Yahoo closes; the 5-year replay uses the same source. The replay is linear and does not credit the hedge puts.
The risk engine's beta-weighted crash test uses a 1.5 placeholder beta for LLY and LMT (no measured value); the liquidation battery keeps it, which makes it conservative here.
Read-only hypothetical; nothing here places orders.