Paper Trader

Bluechip — leveraged quality (Buffett/Power Corp) · one review each market day, every decision logged.

Portfolio Value
$97,736.38
as of 2026-09-10
Total Return
−2.26%
−$2,263.62 since inception
Latest Day
—
session P&L
Sessions
7
since 2026-08-25
Cash
$-70,399.89
uninvested buffer
Overviewmandate, equity, what it holds
Mandate

The AQR finding that Berkshire's record is quality plus cheap steady leverage, run live: a fixed 25-name universe of A+-or-better-rated mega-caps, a monthly quality screen (return on equity, net debt to EBITDA, free cash flow — missing data fails, never passes), and up to 12 names the LLM picks with a written moat-and-valuation thesis each, held at 1.7x gross exposure — Berkshire's own measured leverage — on a margin debit the risk desk caps in code. Daily sessions only monitor: exits need a broken thesis, and a falling price is not one. The catch is stated rather than buried: Buffett's leverage came from insurance float that cannot be recalled, this comes from margin that can, so the book publishes how far it can fall before a call — and accrues the ~6.5% interest Alpaca does not charge, because free leverage would flatter it by exactly the cost that separates the two.

How it was decided

Eric's mandate of 2026-08-24 — 'Warren Buffett / Power Corp: triple-A/blue-chip companies with 25% leverage built in' — researched and designed the same day by Claude Fable 5, then raised on 2026-08-25 to 1.7x to match Berkshire's measured float leverage exactly. The evaluation was pre-committed before the first trade: judge at 60+ sessions, net of financing, against the honest null of 1.7x the index — and count a margin call as a failed run, not a bad quarter.

Equity Curve · from $100,000
Current Holdings · 12
SymbolMarket ValueUnrealized
META $16,220 +14.22%
NVDA $14,717 +3.64%
MSFT $14,312 +0.79%
JNJ $14,081 -0.84%
PG $13,911 -2.03%
KO $13,719 -3.38%
V $13,638 -3.96%
AMZN $13,626 -4.04%
PEP $13,549 -4.58%
ADP $13,523 -4.77%
MA $13,432 -5.41%
GOOGL $13,407 -5.58%
Riskmeasured after the session, never shown to the agent
Attribution · market vs decisions

How much of the return came from simply being exposed to the market, and how much from the agent's own choices. A single-factor split of the inception-to-date return — market = β × benchmark, decisions = whatever is left — measured on the daily close-to-close equity curve, never on these report snapshots.

Not yet. 16 aligned daily observations, need 20 A beta fitted on fewer than 20 aligned days is noise, and reporting one anyway would be the flattering mistake.
FRTB Capital · standardised approach · 2026-09-10

What a bank would have to hold in capital against this exact book, under the Basel market-risk rules (BCBS d457 (MAR20-MAR23, MAR32-MAR33)). Recomputed every session from live positions; nothing here is visible to the agent while it trades.

Capital requirement
$61,195
SbM + DRC + RRAO
Of equity
62.6%
capital / portfolio value
Market-risk RWA
$764,937
capital × 12.5
Components, buckets, and why DRC is the interesting line
ComponentChargeShare
SbM delta · high correlations bind $35,975 58.8%
SbM vega · no optionality in the book $0 0.0%
SbM curvature · linear payoff $0 0.0%
DRC — jump to default · LGD 100% $25,220 41.2%
RRAO · no exotic underlyings $0 0.0%
Total capital requirement$61,195100%
Equity buckets · MAR21.72
BucketRisk weightNamesNet valueKb
8 50% ADP, MA, MSFT, NVDA, V $69,625 $23,358
5 30% AMZN, JNJ, KO, PEP, PG $68,883 $13,863
6 35% GOOGL, META $29,628 $8,420
Why the second-biggest line is the interesting one. The DRC is jump-to-default: the issuer disappears overnight. Equity has no recovery, so Basel sets loss-given-default at 100% and charges an unrated name 15% of its market value — 41% of the requirement here, and nothing about hedging, forecasting or diversification reduces it. That the high correlation scenario binds is itself a readout: a long-only book is dearest when correlations rise, a hedged one when they fall.
Value at Risk · historical simulation · 2026-09-10

The loss this book exceeds on the worst few days in a hundred, measured by replaying 501 days of its own constituents' returns through today's positions. Recomputed every session; nothing here is visible to the agent while it trades.

VaR 99% · 1 day
$3,774
worst 1 day in 100
Of equity
3.86%
loss / portfolio value
ES 97.5%
$4,008
average of the tail beyond
VaR 99% · 10 day
$9,717
regulatory horizon
Quantiles, component VaR, the backtest, and what VaR will not tell you
ConfidenceHistoricalOf equity ParametricGapTail pts
95% $2,266 2.32% $2,513 −9.8% 25
99% $3,774 3.86% $3,554 +6.2% 5
ES 97.5% $4,0084.10% the average of the tail, not its edge
Why the 10-day figure is measured, not scaled. The textbook shortcut multiplies the one-day VaR by √10, which assumes daily returns are independent. Here that gives $11,936 against a measured $9,717 — 22.8% too high. Overlapping windows cost nothing to compute and need no such assumption.
Component VaR · at 99% · sums to the total
NameMarket valueWeightComponentShare
META $16,221 9.6% $620 17.4%
NVDA $14,724 8.8% $509 14.3%
AMZN $13,626 8.1% $482 13.6%
GOOGL $13,409 8.0% $364 10.3%
MSFT $14,322 8.5% $362 10.2%
MA $13,432 8.0% $298 8.4%
V $13,640 8.1% $291 8.2%
ADP $13,523 8.0% $249 7.0%
These are allocations, not a ranking. Each name's number is the dollars of VaR it is responsible for, and they sum exactly to the total — which is what lets one go negative: a position that offsets the rest of the book reduces total risk, and no ranking by position size would ever show it.
The part that makes it a measurement. Over 251 out-of-sample days the 99% threshold was breached 1 time(s) against 2.5 expected — Kupiec p = 0.276, verdict consistent with the model, Basel traffic light green. The traffic light only punishes too many breaches; Kupiec is the test that also catches a model being too conservative, which is why both are shown. Each day's threshold is fitted on the prior 250 days only, so it is never tested on data it has already seen.
Two different questions. Everything above shocks today's positions with market history. The account's own equity curve — 20 sessions, 14.5% annualised volatility — says what it has actually done: 95% VaR 1.84% · 99% VaR 1.84%. They diverge whenever the agent has been trading.
What VaR will not tell you. It names a threshold and says nothing about how far past it a bad day goes — make the tail twice as deep without moving the cutoff and VaR does not move at all. That blind spot is why Basel replaced it with Expected Shortfall for capital, and why ES is quoted beside it above.
Counterparty Risk · SA-CCR · 2026-09-10

Market risk is symmetric — the price moves, you win or you lose. Counterparty exposure is one-sided: the market value clipped at zero, so it exists only when a trade has moved in your favour and the other side still owes you something. A fully-paid share bought on an exchange has none of that. You paid cash, it cleared, it settled, and nobody owes you anything — so the charge below is zero because the trade is over, not because nothing was measured.

SA-CCR EAD
$0
α × (RC + PFE) · 0 derivative(s)
CVA capital
$0
spread risk, not default
Settlement in flight
$0
T+1
Custodial concentration
100%
one broker, everything
The derivation: SA-CCR terms, what was excluded, settlement, leverage and concentration
SA-CCR termValue
Derivative trades in the netting set 0
Replacement cost · max(V − C, 0) $0
Add-on aggregate · supervisory factors $0
PFE multiplier · floored at 5% 1.0000
PFE · multiplier × add-on $0
α · CRE52.1 1.4
Exposure at default$0
CVA capital · CRE55, BA-CVA reduced $0
Why · what never reaches SA-CCR
KindPositionsMarket valueReason
cash equity long 12 $168,165 fully-paid long share: bought for cash, centrally cleared and settled — no open obligation, so no counterparty exposure
Settlement is the one that is genuinely live. US equities settle T+1, so a fill from this session is a commitment against a leg that has not been delivered yet: $0 in flight, $0 past due, capital charge $0. Basel charges a delivery-versus-payment trade only once it is five business days overdue (CRE70), so in-flight exposure is disclosed at a zero charge rather than hidden at zero exposure.
Leverage · capability vs utilisation
ProvisionedMargin debitShort valueGrossNet
4× · shorting on · options L3 $70,400 $0 1.72× 1.72×
Leveraged. This is what makes the zeros above readable as a fact about today's positions rather than about the account's permissions: the book is provisioned for 4× margin, shorting and options, and is currently using 72.0% of the first and 0.0% of the second. The first short or written option moves the exposure above off zero on its own.
The counterparty that does exist. Alpaca (paper) holds $-70,400 of cash and $168,165 of securities — 100% of the book, by construction. Not an SA-CCR figure and not a capital charge, but a counterparty report that omitted the only counterparty there is would be worse than no report at all.
What this is really measuring. Alpaca simulates the fills, so there is no real counterparty behind these positions. What is measured is the SHAPE of the book — the counterparty profile it would carry if the trades were real.
[ LOCKED ]Risk metrics — Sharpe, max drawdown, and performance vs the S&P 500 — stay hidden until 20 sessions (13 to go). Below that, a handful of days can't say anything meaningful, so we don't pretend otherwise.
Sessionsevery decision, in the agent's own words
Latest Review · 2026-09-10
Trades

No trades this session — held all positions.

Agent Reasoning
No thesis breaks found across any holding — news flow is routine (Alphabet earnings beat, Meta AI agent rally, MA/V payment-network partnership news, PG/KO consumer commentary, macro inflation/rate jitters). None of it rises to fraud, delisting, guidance collapse, or a screen fail. Price weakness across several names (GOOGL, MA, V, PEP, KO, AMZN) reflects the broader market pullback (S&P -1.9% week, VIX up to 17.7 but still well under the 22 defensive trigger), not business deterioration. Session — Daily monitoring check (12/25 PASS from last review, 2026-08-25 cache). No screen re-run today. Actions — No trades. All 12 holdings reviewed via news; no broken thesis (no fraud, delisting, guidance collapse, or screen fail) identified. Weakness in GOOGL/MA/V/PEP/KO/AMZN is price-only, consistent with today's broad market pullback (SPX -0.5%, VIX +7.5% to 17.7) amid inflation-data jitters — not a reason to sell under the strategy's rules. Book — 12 holdings, roughly equal-weighted 13.7%–16.6% (META largest at 16.6% on AI-driven strength, +14.5% unrolled gain). Gross exposure 1.72x vs 1.70x target; margin debit $70,400; cash -$70,400. Maintenance cushion: holdings can fall ~40% before a margin call. Cumulative synthetic financing: $200.59 (today's accrual $12.54 at 6.5% APR) — tracked, not paid, but counted against this book's net result. Watching — MSFT (bull case cites 75% upside on Azure/Copilot momentum despite flat 12-month price — monitoring for confirmation, not action); PG (Cramer/street flagging "no growth" concern — watch for a real screen deterioration at next review, not acting on sentiment alone); GOOGL (down ~5.5% unrealized despite strong earnings — pure valuation/price move, thesis intact). No exits, no adds. Next deep review within 28 days of 2026-08-25.
History · 6 earlier sessions
2026-09-09 $98,438.72 — 0 trades

No trades this session — held all positions.

No thesis breaks found across all 12 holdings. JNJ's dip, PG's "no growth" commentary, GOOGL's AI-competition chatter, and NVDA/AMD valuation talk are all normal market noise/valuation debate — not fraud, delisting, or guidance-collapse-grade deterioration. Macro backdrop shows rate-hike jitters and oil above $100, pressuring the tape modestly, but nothing holding-specific breaks a thesis. Session — Daily monitoring check; screen breadth 12/12 PASS (as of last review 2026-08-25). No buys today per contract. Actions — No trades. Reviewed news for all 12 holdings: no fraud, delisting, or guidance-collapse signals. JNJ (-2.2% day) and broad softness reflect macro rate-hike/oil concerns, not company-specific deterioration — holding through drawdown per strategy. Book — 12 positions, equity $98,375, long $168,747 (est.), gross exposure ~1.72x vs 1.70x target (within tolerance, no rebalancing action today since buys are prohibited and no forced sell trigger). Margin debit ~$70,400. Maintenance cushion: holdings can fall ~40% before a margin call. Cumulative synthetic financing: $188.05 (+$12.54 today) — tracked, not paid, but counted against net performance. Weights roughly even (13-16% each): META largest at ~16.5% (+14.6% unrealized, AI agent launch tailwind), NVDA ~15.4% (+6.8%), MSFT ~14.6% (+1.4%). Laggards: GOOGL (-5.5%, AI competition noise), ADP (-5.3%), MA (-4.9%), KO (-4.3%), PEP (-4.3%) — all price drawdown, no thesis breaks. Watching — GOOGL: BofA flagged AI-rival pressure despite traffic growth — monitor cloud/search metrics into next review. PG: persistent "low growth" narrative from Cramer/street — watch for margin or volume deterioration in next earnings. No new entry candidates evaluated today (no-buy session). Macro: rate-hike odds and oil >$100 bear watching for regime shift (VIX 16.2, still below defensive threshold of 22).
2026-09-08 $99,114.25 — 0 trades

No trades this session — held all positions.

Daily Check — 2026-09-08 Session — Daily monitoring only (no buys). Last screen 2026-08-25: 12/12 holdings PASS. Actions — No trades. Reviewed news across all 12 holdings; nothing rises to thesis-breaking (no fraud, delisting risk, or guidance-collapse-grade deterioration): - PG: margin-pressure headlines ($1.4B FY27 FX/cost headwind) and "no growth" commentary — a valuation/margin watch item, not a broken moat or balance-sheet issue. - PEP: mixed press (cheap-vs-FCF bull case alongside "what's wrong" bear case) — normal noise, thesis intact. - MA/V: AXP relative underperformance noted, but no negative read on MA/V fundamentals themselves. - NVDA/META/GOOGL: routine AI-capex and competitive headlines, no fundamental deterioration. - ADP/AMZN/JNJ/KO/MSFT: routine market commentary, nothing thesis-altering. Market backdrop: S&P -0.36%, NASDAQ -0.39%, VIX 15.3 (calm regime, no freeze/defensive trigger). No name near a screen FAIL. Book — 12 holdings, weights ~13.7–15.7% each. Gross exposure 1.71x vs 1.70x target (essentially at target, buy budget to target is negative $1,025 — confirms no room to add, consistent with no-buy day). Margin debit $70,400. Maintenance cushion: holdings can fall 41% before a margin call. Cumulative synthetic financing: $175.52 (today's accrual $75.22 — note: this appears elevated vs prior days' typical ~$13-15/day pace; flagging for review-day reconciliation, not actionable today). Watching — PG margin trajectory (FY27 cost headwind) and PEP volume/growth trends for the next deep review; no changes warranted today. No new entry candidates evaluated (buy-day only). No exits triggered.
2026-09-02 $99,463.36 — 0 trades

No trades this session — held all positions.

Session — Daily monitoring check, 2026-09-02. Screen breadth: 12/12 holdings PASS (last review 2026-08-25). No buys today per contract. Actions — No trades. Reviewed news across all 12 holdings; nothing rises to broken-thesis level: - JNJ: Louisiana jury talc verdict — ongoing litigation overhang, already priced/known risk, not new fraud/delisting evidence. - META: $10B legal charge to be booked in Q3 — one-time charge against a company generating >$60B annual FCF; capex plans intact. Not guidance-collapse. - V: DOJ debit-card case document production — discovery phase, trial not until 2028, no adverse ruling. - NVDA: guidance/estimate noise post-earnings, Mediatek/Hugging Face deal speculation — normal news flow, no fundamental deterioration. - ADP: soft August payrolls print (38k vs 47k consensus) — macro data point, not company-specific. No screen FAILs, no fraud, no delisting risk anywhere. Holding through normal drawdown/news noise per strategy design. Book — 12 positions, roughly equal-weighted 13.7–14.9% each (ADP 14.2, AMZN 13.8, GOOGL 13.7, JNJ 14.7, KO 13.7, MA 14.0, META 14.9, MSFT 14.6, NVDA 14.9, PEP 13.9, PG 14.3, V 14.1). Equity $99,447; long $169,842; margin debit $70,400. Gross exposure 1.71x (target 1.70x, essentially on target, buy budget -$781 so no room even if buys were permitted). Maintenance cushion: holdings can fall 41% before margin call. Cumulative synthetic financing: $100.30 (today's accrual $12.54). Watching — No near-entry or near-exit names today; all 12 holdings remain PASS. Continuing to monitor JNJ talc litigation trajectory and META's Q3 charge disclosure for next review; no action warranted now.
2026-09-01 $99,386.34 — 0 trades

No trades this session — held all positions.

Session — Daily monitoring check (2026-09-01). No new screen run; last review 2026-08-25, 12/12 PASS. Actions — No trades. Reviewed news on all 12 holdings; nothing rises to thesis-broken. Market-wide soft patch (SPX -0.64%, NASDAQ -1.13%) driven by rising yields/oil, not company-specific deterioration. Notable but non-actionable items: Alphabet "brain drain" narrative (soft, no fundamental data); PepsiCo Ukraine plant damage (immaterial, no injuries); routine dividend/governance chatter on PG/JNJ; NVDA/MSFT modest weakness tied to broad tech pullback. None constitute fraud, delisting risk, or guidance collapse — holding through. Book — 12 positions, each ~13.8-14.9% of equity (within 20% cap). Gross exposure 1.71x vs 1.70x target (buy budget to target: -$833, essentially at ceiling — no room to add even if buys were allowed today). Margin debit $70,400. Maintenance cushion: holdings can fall 41% before a margin call. Cumulative synthetic financing: $87.76 (today's accrual $12.54 @ 6.5% APR) — tracked cost this book is judged against. Watching — No entry candidates evaluated today (daily session, no buys permitted). On watch for next review: GOOGL (recent "brain drain"/talent-retention headlines — monitor for any concrete guidance impact), PEP (soft relative performance vs KO, plus minor Ukraine plant incident — immaterial but tracking), NVDA (elevated volatility around AI capex headlines, screen still PASS on fundamentals). No exits triggered.
2026-08-31 $99,707.48 — 0 trades

No trades this session — held all positions.

Session — Daily monitoring check (2026-08-31). Last screen 2026-08-25: 12/12 holdings PASS, breadth unchanged. Actions — No trades. News scan across all 12 holdings shows routine coverage (earnings recaps, analyst notes, competitive positioning, dividend commentary). No fraud, delisting risk, guidance collapse, or screen failures found. One item worth logging: a MarketBeat recap mentioned a "$18B settlement" tied to Meta — no primary-source detail or guidance impact found elsewhere; treating as noise pending review-day confirmation, not a thesis break. Macro backdrop: Iran/US strikes lifted oil, VIX ticked up to 15.2 (still calm, well below defensive threshold), indices modestly red. None of this touches individual holding fundamentals. Book — 12 positions, roughly equal-weighted 13.8%–14.9%. Gross exposure 1.71x vs 1.70x target (essentially at target, buy budget to target is -$581, i.e., no room — consistent with no-buy rule today anyway). Margin debit $70,400. Maintenance cushion: holdings can fall 41% before a margin call. Cumulative synthetic financing: $75.22 (today's accrual $75.22, first day tracked). Drawdown from HWM: 0.3% — well inside all risk-desk limits. Watching — Nothing near exit; all theses intact. For next review: (1) confirm nature/materiality of the Meta legal settlement headline before it's stale news; (2) monitor AMZN/GOOGL capex-funding divergence commentary for any margin-durability signal; (3) no new entry candidates evaluated today — universe re-screen happens at next deep review.
2026-08-25 $99,967.82 — 12 trades
ActionSymbolAmount
BUY PG $14,200
BUY KO $14,200
BUY PEP $14,200
BUY JNJ $14,200
BUY V $14,200
BUY MA $14,200
BUY AMZN $14,200
BUY META $14,200
BUY GOOGL $14,200
BUY ADP $14,200
BUY NVDA $14,200
BUY MSFT $14,200
Session — Monthly deep review; 20/25 universe names PASS. Actions — Initiated full book from cash, 12 names at ~14.2% each (~$14,200 notional), targeting 1.70x gross: - MSFT (P/E21, ROE34%): cloud+enterprise software moat, network effects, fairly priced. - NVDA (P/E16, ROA-adj strong): AI GPU/CUDA ecosystem dominance, cheap for growth. - ADP (P/E21, ROE72%): payroll-processing switching costs, sticky enterprise relationships. - GOOGL (P/E24, ROE49%): search+YouTube+cloud moat, reasonable multiple. - META (P/E16, ROE30%): social network effects, scale advertising, cheapest mega-cap tech. - AMZN (P/E25, ROE31%): AWS+logistics scale moat. - MA (P/E26, ROA-adj strong): payment-network duopoly, high switching costs. - V (P/E25, ROE61%): payment-network duopoly, toll-road economics. - JNJ (P/E22, ROE26%): diversified pharma/med-device moat, defensive. - PEP (P/E16, ROE52%): brand/distribution scale, cheapest defensive. - KO (P/E26, ROE42%): global brand/distribution moat. - PG (P/E20, ROE30%): consumer-staples brand portfolio moat. Skipped EXPENSIVE-flagged AAPL, COST, WMT (P/E>30, no compelling edge over cheaper peers already held in-sector). Skipped NKE, CSCO, ACN, TXN, LLY to respect sector caps (Tech and Consumer Defensive both capped at 3) and prioritize best moat/valuation combination. Book — 12 positions, ~14.2% each, gross ≈170% (target 1.70x), sectors: Tech 3, Consumer Defensive 3, Communication Services 2, Financial Services 2, Consumer Cyclical 1, Healthcare 1 — all within 3-per-sector cap. Margin debit ≈$70,400. Maintenance cushion: fresh position, no prior debit history — monitor as fills settle. Cumulative synthetic financing: $0.00 (day one). Watching — LLY, TXN, CSCO, ACN, NKE (PASS, next in line if a slot opens); AAPL/COST/WMT for re-entry if valuations compress below 30x.