Daily snapshot · allocator targets · 2026-09-25 9:30 PM ET
Portfolio Risk Metrics LIVE PAPER
Sharpe
-5.14
total vol risk
Sortino
-5.37
downside only
Calmar
-7.10
CAGR / MaxDD
Max DD
-7.7%
worst drawdown
Total Return
-6.96%
trailing 23 trading days (Alpaca 1M history window)
Trade Win Rate
39%
305 wins / 775 closed
Alpaca account history · through 2026-09-25
Cumulative P&L LIVE PAPER
$-19,693.47
775 closed round-trips · net realized P&L, FIFO-matched fills
$5.5k−$8.0k−$21.6k
Jun 23 '26Aug 1Sep 1Sep 25 '26
Alpaca fills · FIFO realized · excludes BIL/SGOV/SHV & <$25 dust · not a backtest
Daily P&L Calendar (vs $VOO)
Window total$-6,042.70
Green days10/23
Avg win$331.66
Avg loss$-719.95
Best day
$886.67
Worst day$-2,641.02
Agreement13/22 (59%)
Up-days agreement5/9
Down-days agreement8/12
Correlation r0.326
Mean |delta|0.757%
August 2026
VOO vs P&L · % day move
MonTueWedThuFri
3—+1.46%
4—+1.78%
5—-0.20%
6—-0.11%
7—+0.59%
10—-0.05%
11—-0.31%
12—+0.27%
13—+0.66%
14—-0.18%
17—-0.46%
18—-0.68%
19—+0.21%
20—-0.85%
21—+0.41%
24—-0.28%
25-138.29+0.29%
26-128.66+0.02%
27+327.96+0.68%
28-907.09-0.23%
31+886.67-0.33%
September 2026
VOO vs P&L · % day move
MonTueWedThuFri
1-1208.39-0.67%
2+743.89+0.46%
3+712.99+1.05%
4-1126.22-0.40%
7—·
8-99.68-0.52%
9-1658.61-0.48%
10+147.25-0.59%
11-58.57+0.84%
14+40.47-0.45%
15-462.79-0.44%
16-550.47-0.44%
17+226.31+1.12%
18-2641.02+0.13%
21-69.93+1.55%
22-309.61-0.01%
23+11.32-0.76%
24+14.92+0.00%
25+204.85+0.48%
28
29
30
Alpaca get_portfolio_history · rolling 1M broker window (2026-08-25 to 2026-09-25) · VOO from Yahoo Finance daily closes (auto-refreshed each site build) · "·" = no VOO data · correlation on daily % returns · dollar P&L vs index %-move is agreement context, not attribution · Mon–Fri only
Strategy Fleet Inventory
Capital = allocated book value · Growth = since start date, transfers excluded · Data as of 2026-09-16 · 2026-09-17 4:15pm ET · Capital ▼ default sort
Paper sleeves marked "Live · Paper" trade a shared Alpaca paper account (PA3HAIILY8S9) via the executor; Kalshi fleet account is shared with an external actor. "Growth" = equity change since start date excluding transfers (paper sleeves: unit-NAV mark-to-market on fixed notional; live: account P&L on static baseline). "—" = not yet measurable (n too small). Explorations without capital show "—" for capital-derived cells.
Strategy Summary 80/20 HOLDOUT
Exit-date split · holdout excluded from optimization.
Full Sample775 closed round-trips
Net Profit
$-19,693.47
all closed trades
Win Rate
39.35%
305 wins / 775 closed
Profit Factor
0.43
gross profit / gross loss
# of Trades
775
closed round-trips
Max Drawdown
$23,309.64
peak-to-trough decline
Avg Return
$-25.41
mean P&L per trade
Out-of-Sample · Held Back155 of 775 (20.0% of sample)
Avg S/Trade
$-5.77▲ $24.56
OOS mean P&L
Win Rate
29.03%▼ 12.9pp
45 wins / 155 OOS
Profit Factor
0.29▼ 0.14
OOS gross profit / loss
Drawdown/Profit
—— N/A
OOS drawdown / net profit
Total Trades
155▼ 465
held-back round-trips
Edge Decay
▼ 32.42%
profit factor vs in-sample
small sample, deltas may be noisy
Cumulative Equity
In-sampleOut-of-sample
$5.5k−$8.0k−$21.6k
train / test
Jun 23 '26Aug 1Sep 1Sep 25 '26
Current Holdings ● LIVE — Alpaca
Symbol
Strategy
Qty
Avg Entry
Current
Market Value
Unrealized P&L
BIL
Memory Rotation
626.6013
$91.61
$91.62
$57,409.22
$4.14 (+0.01%)
MSFT
Bilbo Paper
19
$504.69
$516.17
$9,807.23
$218.12 (+2.27%)
DIG
—
120.8418
$71.64
$65.81
$7,952.60
$-704.84 (-8.14%)
ALAB
—
1
$364.97
$364.62
$364.62
$-0.35 (-0.10%)
CSCO
—
2.9999
$106.63
$106.70
$320.09
$0.21 (+0.07%)
SBLK
Benchmark Core
1
$30.92
$29.54
$29.54
$-1.38 (-4.46%)
Alpaca paper · updated 2026-09-27 7:33 PM ET
Position History
Date (ET)
Symbol
Side
Qty
Price
Strategy
2026-09-25 5:54 PM
BIL
BUY
69.6037
$91.62
Memory Rotation
2026-09-25 5:53 PM
BIL
BUY
531.7988
$91.62
Memory Rotation
2026-09-25 12:05 PM
STNG
SELL
6.2002
$81.52
Hormuz Carry
2026-09-25 9:34 AM
MRVL
SELL
1
$263.50
Ndx Momentum
2026-09-25 9:33 AM
NBIS
SELL
1
$242.64
Ndx Momentum
2026-09-25 9:32 AM
LRCX
SELL
1
$309.00
Ndx Momentum
2026-09-24 4:11 PM
ARM
SELL
1
$305.75
Jump Mirror
2026-09-24 3:46 PM
CSCO
BUY
2.9999
$106.63
—
2026-09-24 3:46 PM
ALAB
BUY
1
$364.97
—
2026-09-24 3:46 PM
ARM
BUY
1
$308.36
Jump Mirror
Active Strategies
Allocated Sleeves · LIVE PAPER places paper-account orders; PAPER SIM places no orders
Memory RotationESTLIVE PAPER
Leveraged momentum — SNDK / MU / TQQQ / DIG · Re-risk cap 25%→30% on/after 2026-08-29 only when 30Y 5-day yield change ≤ 0 bp
1-min Truth Social poller — NOTIFICATION ONLY: alerts sent, trading disabled 2026-08-20, no orders placed
Pending: BUY F ($750.00), BUY GM ($750.00) · WEEKEND
Last run 2026-09-27 6:28 PM ET
Inverse CryptoESTACTIVESIGNAL-ONLY · PAPER SIM
7 strategies tested — no alpha after costs
Last run 2026-09-26 10:20 AM ET
DRAMESTPAUSEDSIGNAL-ONLY · PAPER SIM
PAUSED TRADING — signal-only paper simulation continues; Dynamic Regime-Adaptive Momentum with vol-scaled sizing + VIX-adaptive ATR trailing stop
Last run 2026-09-25 9:31 PM ET
VNQ ReversionESTACTIVESIGNAL-ONLY · PAPER SIM
REIT mean-reversion sleeve — VNQ vs BIL
Last run 2026-09-26 10:20 AM ET
Earnings MomentumESTACTIVESIGNAL-ONLY · PAPER SIM
Post-earnings momentum — scanner 9:00 PM ET, GTC entries at next bridge pass, 6.5% target / 5.5% stop / 5-session time exit
Last run 2026-09-26 10:20 AM ET
Braided-DeltaESTACTIVESIGNAL-ONLY · PAPER SIM
LIVE SPY timing sleeve — sector residual-dispersion braid plus breadth and basket confirmation; close signal executes next open. Flipped live 2026-08-30 (user directive; early flip).
Last run 2026-09-27 8:00 AM ET
WaterholeESTACTIVESIGNAL-ONLY · PAPER SIM
LIVE SPY stress ladder — robust SPY volatility/drawdown and TLT-return composite; 0/40/70/100% exposure. Flipped live 2026-08-30 (user directive; early flip).
Last run 2026-09-27 8:00 AM ET
BTC Regime GateESTACTIVESIGNAL-ONLY · PAPER SIM
EMA20d>EMA200d long/cash timing of BTC — OFFENSIVE=IBIT (spot BTC ETF), DEFENSIVE=BIL. Backtest 4.66y (USDT+USDC, 25bp costs): +22.6% CAGR / -32.6% maxDD vs B&H +11.6% / -67.6%; ~1.8 flips/yr. Flipped live 2026-08-30 (user directive; early flip).
Last run 2026-09-27 8:00 AM ET
Retired · no allocation, kept for historical context
Maple SleeveESTRETIRED0% ALLOCATION
Canadian blue-chip dividend diversification sleeve
Pre-market gap momentum scanner — allocation removed after live evaluation
Retired · last run 2026-09-27 1:30 AM ET
Meme Sentiment SIGNAL-ONLY / NOT VALIDATED / NO LIVE CAPITAL
Meme Mania
89.2
Feeds Used
reddit, polymarket, kalshi, twitter
Feeds Missing
kalshi: OK · polymarket: OK · reddit: STALE CACHE 4h old · twitter: OK
Ticker
Meme Score
Suggested Weight
No candidates above threshold
Meme signal · 2026-09-25 9:32 PM ET
Earnings Calendar SCANNER
Ticker
Earnings Date
Days Until (at scan)
Decision
Reason
Monday, 2026-09-28
CCL EST
2026-09-28 (Monday)
3 days
NO SETUP
price -12.12% below 50-day MA (trend gate)
IDT EST
2026-09-28 (Monday)
3 days
NO SETUP
momentum fading not stable/accelerating
JFR EST
2026-09-28 (Monday)
3 days
NO SETUP
momentum fading not stable/accelerating; price -2.53% below 50-day MA (trend gate)
MSS EST
2026-09-28 (Monday)
3 days
NO SETUP
price -3.7% below 50-day MA (trend gate)
MTN EST
2026-09-28 (Monday)
3 days
NO SETUP
momentum mixed not stable/accelerating; price -5.66% below 50-day MA (trend gate)
NKE EST
2026-09-28 (Monday)
3 days
NO SETUP
price -9.17% below 50-day MA (trend gate)
Tuesday, 2026-09-29
CAG EST
2026-09-29 (Tuesday)
4 days
NO SETUP
momentum mixed not stable/accelerating; price -5.91% below 50-day MA (trend gate)
CNXC EST
2026-09-29 (Tuesday)
4 days
NO SETUP
momentum fading not stable/accelerating
KMX EST
2026-09-29 (Tuesday)
4 days
NO SETUP
momentum mixed not stable/accelerating; price -3.67% below 50-day MA (trend gate)
MUC EST
2026-09-29 (Tuesday)
4 days
NO SETUP
price -6.08% below 50-day MA (trend gate)
UEC EST
2026-09-29 (Tuesday)
4 days
NO SETUP
momentum mixed not stable/accelerating; price -12.83% below 50-day MA (trend gate)
Wednesday, 2026-09-30
CALM EST
2026-09-30 (Wednesday)
5 days
NO SETUP
momentum mixed not stable/accelerating; price -16.59% below 50-day MA (trend gate)
FDS EST
2026-09-30 (Wednesday)
5 days
NO SETUP
momentum mixed not stable/accelerating; price -2.52% below 50-day MA (trend gate)
JBL EST
2026-09-30 (Wednesday)
5 days
NO SETUP
price -0.26% below 50-day MA (trend gate)
JEF EST
2026-09-30 (Wednesday)
5 days
NO SETUP
momentum mixed not stable/accelerating; price -10.09% below 50-day MA (trend gate)
MU EST
2026-09-30 (Wednesday)
5 days
NO SETUP
Pullback 1.78% > 1.5% threshold (price above MA entry zone); Runup 10.55% >= 10% cap (already run up too much); momentum mixed not stable/accelerating
PRGS EST
2026-09-30 (Wednesday)
5 days
NO SETUP
momentum mixed not stable/accelerating; price -5.16% below 50-day MA (trend gate)
Scanner: pullback ≤ 1.5%, runup ≤ 10%, RSI ≤ 70, entry window 5 days · Generated 2026-09-25 9:10 PM ET
Earnings Momentum LIVE PAPER
Trading Days Since Signal EST
14 / 5
Open Legs
0 / 5
Signal
2026-09-07 9:02 PM ET
Symbol
Status
Entry
Target
Stop
Current
Unrealized P&L
CHWY EST
— · no live position confirmed
$23.66 EST
$25.20
$22.36
—
—
BAH EST
— · no live position confirmed
—
—
—
—
—
CMG EST
— · no live position confirmed
—
—
—
—
—
EXLS EST
— · no live position confirmed
—
—
—
—
—
LMT EST
EXITED · time exit
—
—
—
$526.63
—
6.5% target · 5.5% stop · 5-session time exit · exit monitor 9:40 AM / 11:40 AM / 1:40 PM / 3:40 PM ET · last check —
TimesFM-3 Shadow Research GLD · 21dPAPER / RESEARCH
Zero-shot 330M foundation model (Google, non-commercial checkpoint) vs the
forecast_v2 hybrid Monte-Carlo baseline. Walk-forward: 504d context → 21d horizon.
Directional Brier = mean((p_up − actual)²); 0.25 = coin flip. Shadow
only — no positions. Last run 2026-09-27 5:31 PM ET.
TimesFM-3 (GEPA champion config)
MAE (21d)
$10.67 EST
Dir Hit
54.3% EST
Brier (prob.)
0.2729 EST
CRPS
4.1983 EST
Hybrid v2 (baseline)
MAE (21d)
$10.48 EST
Dir Hit
65.7% EST
Brier (prob.)
0.2317 EST
GEPA champion
(holdout-locked):EST selection Brier 0.2233
over 21 windows · holdout Brier
0.2043
over 14 locked windows ·
dir 71.4% ·
ctx 504 · cov QQQ+DXY+GDX · method median_binary
promising — passed holdout bar on a small adaptive sample; confirm on rolling basis before any weight
Real-money automated strategies on the Public.com brokerage accounts, one
block per account. Data below comes from the stored Public.com portfolio/transaction snapshot and local strategy state,
as of 7:10pm ET, Sep 27. These are last observed values; a failed refresh can leave an older snapshot.
Each account and strategy carries its own as-of stamp.
Sizing: 15% of account equity per position, floor $400, cap 40% of equity, minimum entry $400; each buy = min(target, spendable funds − 0.65% fee reserve). De-risk ladder: after 2 consecutive losing exits the fraction is halved (f=0.075) until a winning trade. Replaced the $500→$3,000 Tier-2 ratchet on Sep 4, 2026.
Management: Exits: 6-night hold timer (HOLD_DAYS=6; re-trigger resets, higher-sig coin wins) and BTC > EMA100 regime gate (flip → cash). Kill gates ENFORCED in code (pre-registered SCALP_PROMPT.md): −35% drawdown from the ledger high-water mark, trailing-12-month return < −15%, avg execution shortfall > 15bp/side (20-side window, from 10 sides) — a trip refuses new entries, alerts once, and sticks until reviewed. Exactly-once orders (durable intent + zero SDK retries), 8dp raw-POST sells, process lock, same-signal-date guard, data-glitch gate, crypto-only open-order guard. Schedule: 8:01 pm ET trade, 9:15 pm watchdog, 10:31 pm catch-up, 9:30 am liveness, 8:30 am status line. 100% LLM-free.
Position
BTC (crypto)
Days Left
1
Sizing Tier
normal (f=0.15)
Target Notional
$567.67
Next Buy
BLOCKED — spendable $-345.34 funds < $400 floor
Loss Streak
0 losing exits in a row
Ledger P&L
$0.00 realized · DD $0.00 from HWM
Fee Tier
0.60%/side (120bp RT; Public API tier for < $10k monthly crypto volume). 0.50% applies only in months with >= $10k volume — the backtest reaches that in 4 of 14 OOS months; NOT assumed by the bot (reserve 0.65%).
Regime Gate
ON (BTC > EMA100)
Sig (last close)
BTC +0.1489 · ETH +0.0538 · SOL -0.1436
Recent runs
No runs logged yet.
Entry >1.75σ / 14d · EMA100 gate · 6-day hold / 1 position · backtest — (no artifact)
Executor: Public.com AI Agent 'Conditional SMH Day Trade' (public-side, autonomous) ·
Public-side gate not observed locally · STALE — flag >36h old
Account: Agent Cash Brok 1 - 5OI23198 cash
Sizing: 2.5% of account equity per trade, clamped to $100–$400 (prompt v2, Sep 4, 2026; PASTED into the Public-side agent 2026-09-04 — live).
Management: Entry: weekdays 3:45 pm ET gate — skip if SMH < 200-day SMA or QQQ 21-day return ≥ +12%; else 3:55 pm ET limit buy at ask + $0.05 (cancel unfilled at close). Exit: sell ALL at 9:30 am ET next day, market. Never short, no stop-loss, never hold past the next open, at most one buy per day. Kill rules (user-enforced): after 60 traded nights must beat the ungated and MA200-only books; stop if drawdown from peak > 35%.
Executor gate
not observed — decided on the Public-side agent
Book
—
Legacy model flag (research cross-check only, not the SMH executor): no-skip · — · as of 6:50pm ET, Sep 27 · 6:50pm ET, Sep 27
First live trade:
3:55 PM ET, Sep 1 — SMH limit buy $100.00 @ $545.47 limit;
FILLED 0.18334 sh @ $545.43, $0.00 fees (valid until Nov 30, 2026)
Weekdays 3:45 PM ET: skip if SMH < 200-day SMA or QQQ 4-week return >= +12%; else buy 2.5% of equity ($100–$400; prompt v2 PASTED 2026-09-04) of SMH at 3:55 PM ET (limit at ask +$0.05, cancel unfilled at close); sell ALL next morning at 9:30 AM ET market. Never short, no stop-loss, never hold past next open.
Asia-gate London Goldweekdays · 3:00 AM buy → 8:00 AM sell (ET)
Executor: Hermes cron executor (gold_bot.py, 6 phases) — LLM-free ·
Flag as of 5:32pm ET, Sep 27 · last observed
Account: Agent Cash Brok 1 - 5OI23198 cash
Sizing: 2.5% of account equity per session, clamped to $100–$400 (gold_bot.py target_notional, Sep 4, 2026); no partial entry below $90; buy also clamped by the daily funds allocator's headroom on this shared account.
Management: 6-phase weekday schedule (Hermes cron, LLM-free): 8:00 pm ET evening anchor → 2:00 am mark (GREEN if GLD > the 8:00 pm price, else RED) → 3:00 am buy (GREEN + BULL regime flag only; extended-hours LIMIT at ask + 1 tick) → 3:05 am cancel unfilled → 8:00 am sell ALL (LIMIT at bid − 1 tick) → 8:05 am sell fallback (market / marketable limit). Never short, never hold past 8:00 am, no stop-loss. Kill gates in code: after 60 traded sessions must beat always-long GLD; 120-session signal agreement must stay ≥ 51%.
Regime flag (weekly)
BULL
Flag detail
p_up 0.7, 21d median $407.59 vs $393.41 (+3.6%)
Kill gates
not tripped (60-trade vs GLD; 120-trade agreement ≥ 51%)
Weekly BULL/BEAR regime flag (trade only on BULL). Weekdays: 2:00 AM ET — GREEN if GLD now > 8:00 PM ET price; GREEN → buy 2.5% of equity ($100–$400) of GLD at 3:00 AM ET (extended-hours limit near ask; cancel unfilled at 3:05 AM); sell ALL at 8:00 AM ET (limit near bid; aggressive fallback if unfilled). RED/flat/incomplete → do nothing. Never short, never hold overnight, no stop-loss — the 8:00 AM exit is the exit.
2y net @0.6bp (GC=F basis): +13.8% CAGR / Sharpe 1.98 / −4.4% maxDD / 59.8% win / 291 sessions (long 4.5%/0.61). Live 2bp cost: ≈low-teens CAGR. Decaying edge: first ½ +28.4%/4.62 vs second ½ +4.0%/0.55 (London 17.1%→7.5%). Stop: last 60 trades < long GLD; 120-trade agreement <51%.
Alpaca Paper Sleeves Alpaca paper (separate broker)● PAPER — Alpaca
Paper-trading sleeves on the Alpaca paper account — a different broker from
Public.com, tracked separately. Hermes cron bilbo_paper — hourly 9:55am–3:55pm ET + 4:10pm ET reconcile
· State as of 4:10pm ET, Sep 8 (ledger.updated_et)
EST
Alpaca Paper SleevesBilbo v2 hourly · 9:55am–3:55pm ET
Executor: Hermes cron bilbo_paper — hourly 9:55am–3:55pm ET + 4:10pm ET reconcile ·
State as of 4:10pm ET, Sep 8 (ledger.updated_et) · STALE — no state update in 36h+
QQQ Sleeve
FLAT
Bilbo Long
3 open (MSFT, TSLA, SMCI)
Bilbo Dual Ledger
primary author-exits + shadow cap-only
Book
Alpaca paper — ~25bp equity risk/trade
Armed compression boxes
Symbol
Box Low
Box High
Grey Bars
Entry
MSFT
$493.81
$500.73
13
$511.58
SMCI
$36.61
$37.40
21
$37.97
TSLA
$350.36
$356.65
11
$371.97
Recent events
Time (ET)
Sleeve
Event
Detail
4:11pm ET, Sep 25
bilbo
Equity mark
4:11pm ET, Sep 25
bilbo
Order error
GOOGL APIError: {"code":40010001,"message":"client_order_id must be unique"}
4:10pm ET, Sep 25
bilbo
Carveout reservation
2:56pm ET, Sep 25
bilbo
Order error
GOOGL APIError: {"code":40010001,"message":"client_order_id must be unique"}
2:56pm ET, Sep 25
bilbo
Skipped
AMD insufficient_cash
Bilbo hourly (Megacap 20, v2 faithful): confirmed hourly close above the 5-grey-candle compression box, RTH closes only, trend gate vs prior-day daily EMA21; buy shares at next 5-min bar open; exits keyed to the stock — 5-min close below box low, or after a +1.0 daily-ATR run a trail giving back 75% of peak, or 10-trading-day cap. Dual ledger: primary author rules + shadow 10-day-cap-only on the same entries. Risk ~25bp equity/trade at the box-low stop, max 8 concurrent. QQQ EMA21 reversion: buy $500 QQQ at next open when QQQ close <= 0.95×EMA21, sell at the open 5 sessions later.
Bilbo v2 replication 2021–2026 (fit 86% vs author live; 0 spurious): +64bp/trade primary / +118bp/trade cap-only shadow, PF 1.40, day-cluster t=3.37 (n=1,758), net 1bp/side. Alpaca paper-only, 4-week forward gate.
8-Bot Options Desk ADVISORY · NO ORDERS
Snapshot Sep 25, 10:53 AM ET
Bots live · profiles
8/8EST
Last desk brief · bot08 run
Sep 25, 10:54 AM ET
Scheduler · execution only
8/8 last-run OKEST
Desk equity snapshot
$3,000.00
Owner cap
$3,000.00
Cap headroom · cap minus equity
$0.00EST
Open reservations · ledger
0EST
Active vetoes / flags · symbols
20EST
Audit findings · fixed / total
12/12 fixed · 0 openEST
DAG E2E · S1–S6
PASSSep 07, 07:39 PM ET
bot01Sep 25, 08:32 AM ET● OKscan
bot02Sep 21, 08:46 AM ET● OKscan
bot03Sep 25, 10:26 AM ET● OKidle
bot04Sep 25, 08:01 AM ET● OKscan
bot05Sep 25, 10:12 AM ET● OKflag
bot06Sep 25, 10:26 AM ET● OKidle
bot07Sep 25, 10:46 AM ET● OKrisk
bot08Sep 25, 10:54 AM ET● OKbrief
EST counts / headroom · audit A1–A10, C2, bot08 budget
ETF CSP Sleeve · 4-Lot Diversified PAPER · NO ORDERS
Candidates sit GATE_PENDING until a locked out-of-sample window exists, then scripts/candidate_gate.py (deflated-Sharpe, novelty, spanning) rules on paper admission. Weekly cycle: Mondays 8:00 AM ET.
Market-Neutral Lab PAPER · READ-ONLY · 5OI24098
short 2 MSTR @ $142.50 · long 9 IBIT @ $45.11 · opened 2026-09-07 EST
marks: MSTR $159.44 · IBIT $47.56 · gross $747 EST
Pair paper P&L
−$12 EST
Net delta drift
+6.3% EST
MSTR vs IBIT (premium residual)
+6.1% EST
Margin est
$480 EST
C4 SPAC carry
IBAC $11.10 · P&L −$10 · gap vs trust -0.37% · extension vote 2026-09-24 EST
NHIC $7.45 · P&L −$319 · gap vs trust -30.24% · redemption 2026-09-15 EST
Put debits (Oct-16) ETHA 2026-10-16P 20.5 · IV +44% · mid $0.930 EST · nearest strike to underlying last $20.325 (quote Sep 25, 2026 3:55 PM ET) UNG 2026-10-16P 11 · IV +46% · mid $0.415 EST · nearest strike to underlying last $11.12 (quote Sep 25, 2026 3:55 PM ET)
HTB probe · Sep 25, 2026 8:05 PM ET · 17/28 OK
Tracker as of Sep 25, 2026 3:35 PM ET · terms/chains snapshot Sep 25, 2026 3:55 PM ET
EST paper book · no broker positions · missing —
GPU Tilt ESTLIVE PAPER
H100 daily $/GPU-hour (Ornn OCPI) leads semiconductor ETFs by ~1 month (Granger p=0.0017 for SOXX). Z-score gated daily signal — OFFENSIVE=60% SOXX/40% SMH, DEFENSIVE=BIL.
Pricing source: Ornn Data OCPI.
Mode
NEUTRAL
Target Book
SOXX 30% / SMH 20% / BIL 50%
Signal
2026-09-25 5:30 PM ET
GPU tilt NEUTRAL: H100 mom=+0.9% Z=-0.91 5d_avg=+0.64% → half position
H100 Signal $/GPU-hr
20d Ago
20d Momentum
20d Z-Score
5d Avg Return
$2.9175
$2.8925
+0.9%
-0.91
+0.64%
Z >+0.5 + rising: 60% SOXX / 40% SMH · <−0.5 + falling: 100% BIL · else ½ book · Last signal: Z -0.91 → NEUTRAL
⚡ VIX for Compute
Ornn daily GPU price observations. Returns and annualized volatility are calculated estimates over the dated history shown; unavailable statistics: —.
Compute VIX (avg)
67.0% EST
H100 last daily index
$2.49
Sep 27, 04:00 PM ET
GPU
Daily $/hr
Observed (ET)
History return
Ann. vol
20-session momentum
History window
A100 SXM4
$0.91
Sep 27, 04:00 PM ET
-27.8% EST
42.0% EST
-13.3% EST
2026-05-25 → 2026-09-27
B200
$8.08
Sep 27, 04:00 PM ET
+51.3% EST
54.2% EST
+19.7% EST
2026-05-25 → 2026-09-27
H100 SXM
$2.49
Sep 27, 04:00 PM ET
-12.6% EST
72.1% EST
-21.3% EST
2026-05-25 → 2026-09-27
H200
$4.62
Sep 27, 04:00 PM ET
+12.4% EST
93.7% EST
+1.4% EST
2026-05-25 → 2026-09-27
RTX 5090
$0.64
Sep 27, 04:00 PM ET
-24.7% EST
73.2% EST
+2.9% EST
2026-05-25 → 2026-09-27
LLM Token Index
$/Mtok (blended)
Observed
Anthropic
$1.600
2026-09-26
Deepseek
$0.052
2026-09-26
Google
$0.586
2026-09-26
Openai
$0.359
2026-09-26
Ornn gpu_history.json / otpi_history.json · historical, not spot
VIX > 40 Episode Alarm ARMED
Advisory only — no orders. Hourly watch: a VIX close above 40 opens an
episode (alert fires once); a close back below 35 re-arms. Signal replicated 2010→2026:
the 17 close-episode starts (2010/2011/2015/2020/2025) marked the best forward-entry windows
of the period. 2018/2024 spikes crossed 40 intraday only — closes are the rule.
VIX last close
14.5
Status
Quiet — no open episode
1-yr forward after episode start (median / worst)
+31% / +18%
Episodes since 2010
17
Alarm cron 65bfb9709462 · health blind_hours=0 (data OK at last hourly check) ·
alert posts to the trading chat; never a trade order. Updated 2026-09-09 6:55 PM ET
ATR SPY Put Credit Spread ESTLIVE PAPER
Weekly −1·ATR SPY 10-wide put credit spread, 1 lot, same-week Friday expiry.
Sell short strike at prior weekly close − 1.0×ATR14; skip unless natural credit ≥ $0.20.
Fill-confirmed entries only; one re-quote budget per week. Account PA3CS30BOBWC.
Realized P&L (void-adjusted)
$2.58
Max Risk / Lot
$1,000
Swan Breaches (13wk)
0/13w
Funding (sells at open)
OBP $100,135
Week
Strikes
Expiry
Credit
Status
2026-W36
750/740
2026-09-04
$0.00
skipped_credit natural credit 0.00 < gate 0.20
2026-W37
755/745
2026-09-11
$0.69
settled
2026-W38
750/740
2026-09-18
$1.35
settled
2026-W39
745/735
2026-09-25
$0.09
skipped_credit natural credit 0.09 < gate 0.20
EST: sleeve-book estimate from state ledger, void-unfilled corrections applied; broker-verified fills only count. Updated 2026-09-24 12:49 AM ET
Kalshi Perps — Live Account & Shadow Books LIVE
Real Kalshi Prime margin account (dedicated FCM subaccount, $2,500 funded 9/15).
Real book is positioned:
metals-carry gate is 0/3 positive funding prints, and the NDX daily up/down study
has not seen a QQQ EMA21 signal day yet. Shadow books below are paper verdicts, not trades.
Account Equity (real)
$2,733
Available / Resting
$0 · 0 orders
Dailies Wallet (event_contract)
$$0.00
Carry Gate (3× positive prints)
0/3
Carry Shadow P&L (always-short / follow)
$0.00 /
$0.00
Perp
Implied Spot
Funding /8h
Funding %/yr
KXAAVEPERP
154.03
0.0000%
+0.0%
KXADAPERP
0.2552
0.0000%
+0.0%
KXBCHPERP
332.75
0.0000%
+0.0%
KXBNBPERP
776.2
0.0000%
+0.0%
KXBTCPERP
84574.0
0.0110%
+12.1%
KXDOGEPERP
0.0974
0.0000%
+0.0%
KXETHPERP
2692.4
0.0000%
+0.0%
KXGOLDPERP
4280.7
0.0000%
+0.0%
KXHYPEPERP
91.599
0.0000%
+0.0%
KXKSHIBPERP
0.0059
0.0000%
+0.0%
KXLINKPERP
14.1505
0.0000%
+0.0%
KXLTCPERP
71.318
0.0000%
+0.0%
KXNEARPERP
5.2087
0.0000%
+0.0%
KXPALLADIUMPERP
1273.3
0.0000%
+0.0%
KXPLATINUMPERP
1782.8
0.0000%
+0.0%
KXSILVERPERP
64.123
0.0000%
+0.0%
KXSOLPERP
122.185
0.0000%
+0.0%
KXSUIPERP
1.243
0.0000%
+0.0%
KXVVVPERP
30.146
0.0000%
+0.0%
KXWLDPERP
0.5561
0.0000%
+0.0%
KXXRPPERP
1.5228
0.0174%
+19.1%
KXZECPERP
1591.91
0.0000%
+0.0%
NDX Daily Up/Down (real book — dailies trade on the event_contract wallet; perps on margined)
No dailies position
NDX Daily Up/Down Shadow (parallel-arm study for the QQQ EMA21 sleeve — equity route stays primary)
QQQ vs EMA21 (signal at −5.0%)
+2.9%
Signal day?
no
Today's market
active
Baseline up-rate / break-even taker price
63.2% / 0.62
Real account last order activity: KXGOLDPERP ask @ 4.2922 · · 2026-09-25 6:00 PM ET.
Verdicts stay shadow until gates pass (carry: persistent positive funding + paper beat vs baseline; dailies: ≤0.62 pricing with depth on a signal day). Updated 2026-09-27 12:10 PM ET
Kalshi Balance — Total, Perpetuals, Predictions LIVE
As of Sep 27, 7:00pm ET
Every dollar the desk has on Kalshi, one total line plus the two wallets
that make it up: the perpetuals margin (FCM) subaccount and the predictions
event_contract wallet. The hourly pull appends one row per hour and the total is recomputed
from the two legs — the widget asserts that they still sum, and a wallet that fails to read
is plotted as a gap, never as $0.
Total on Kalshi
$2,737.96
reconciles ✓
Perpetuals (margin)
$2,737.96
Sep 27, 7:00pm ET
Predictions (event_contract)
$0.00
$0.00 open exposure
Hourly points
3
100% coverage
Balance history — 24h
TotalPerpetualsPredictions
$2.90k$1.45k$0
9/27 5pm9/27 6pm9/27 7pm
Window start Sep 27, 5:53pm ET · $2,731.55End $2,737.96Change $6.41 (+0.23%) — history covers 1h of the 24h window
Account
Equity
Cash
Components
Total on Kalshi
$2,737.96
—
Perpetuals + predictions (sum asserted)
Perpetuals (margin / FCM)
$2,737.96
$2,673.01
open perp position value -$1,240.10 ·
margin in use $76.50 ·
unrealized $9.34
Predictions (event_contract)
$0.00
$0.00
open event exposure $0.00 · no open position
History begins Sep 27, 2026 5:53pm ET (hourly pulls, 3 rows so far) · newest Sep 27 7:00pm ET · coverage 100% of the 1h shown · values: Kalshi margin/balance (perps) + portfolio/balance (event_contract) — measured account values
SMH Overnight API ESTLIVE
API-native migration of the paused SMH in-app agent. Weekdays 3:55pm ET:
gated BUY 1 SMH LIMIT ask+$0.05 as a BRACKET with a broker-side disaster stop
(limit − 1.5×ATR14); DAY/CORE dies unexecuted at the close. Exit: market sell 9:31am ET;
stop child as backstop. Gates: SMH > 200-session SMA and QQQ 21-session return < +12%.
Account 5OI24098.
Status
GATE ON
Round Trips
0/60
Sleeve P&L (realized)
$+0.00
Ungated O/N Benchmark
—
QQQ 21-Session
-3.6%
Equity DD (halt −35%)
+0.0%
Holding 0 sh · last entry 2026-09-15 · $542.70 · benchmark compounds
the same per-trade capital every weekday, gated or not. Kill gate at 60 round trips vs
benchmark — lose to it and the sleeve halts. EST: sleeve ledger from Public API state;
updated 2026-09-15 04:53 PM EDT.
MN Carry (Kalshi×IBIT) ESTLIVE
Delta-neutral funding-carry: SHORT KXBTCPERP + LONG IBIT (Public
5OI23198 — hedge account per 9/16 directive). Gate M5 T0.5 X0.5 C2 (Astra-frozen round-3):
enter when the trailing 5-print funding mean ≥ 0.5bp, exit on 2 consecutive cold means;
every settled print re-evaluates it (12am/8am/4pm ET). A 1-minute deterministic watchdog
flattens on kill (−$50/8d or 3 negative days) and auto re-enters on HOT — no agent in the
trade path. Optimization rounds: R4 (Pareto 450 configs / TimesFM-3 forecast layer / GEPA
121-call evolution) = NO-GO — none beat the frozen gate on holdout; round-3 retained.
Review passes 8:05am/4:05pm ET.
Status
FLAT
Sleeve P&L
$+0.00
Perp leg
$+0.00
Hedge leg
—
Gate (5-print mean)
+0.24bp COLD
Open incidents
0
Perp 0 ct @ — (notional
≈ $0) · hedge 0 IBIT · est. carry at hot pace ≈ $0.00/day
while gate stays HOT · watchdog 1-min, kills armed · snapshot 9/27 7:32pm ET · mark-based
P&L is an estimate (EST) — broker ledger reconciles at exits.
MN Strategy — Reconciled Scorecard LIVE
Combined P&L for the market-neutral perp/hedge strategy across
both venues, from venue ledgers only: Kalshi margin fills (realized_pnl)
+ venue fees + funding credits replayed against the fill-derived position timeline, and
Public hedge legs in Agent Cash Brok 1 5OI23198 / Agent Options Brok 5OI24098
(closed-lot realized from get_history; other sleeves' legs on the same
accounts excluded). Window: 9/16 1:39am ET open → now. Kalshi perps are $1/point per
contract — verified fill-by-fill against venue realized P&L.
Net all-in (incl open + experiments)
$-156.70
Closed P&L (fees + funding in)
$-135.56
Kalshi realized
$-603.72
fees $49.50
Public hedge legs
$+417.23
Funding collected
$+1.43
VOO same window
$-1.34
-0.05% on $2,643.32 deployed
Pair (perp × hedge)
Kalshi leg
Fees
Hedge leg
Funding
Open MTM
Net
RT (W/L)
BTCKXBTCPERP × IBIT 5OI23198OPEN
$-37.88
$36.56
$-38.91
$+0.74
$+3.90 EST
$-35.59
20 (10W/10L)
SOLKXSOLPERP × FSOL 5OI24098
$-74.33
$3.58
$+89.57
$+0.69
$+0.00 EST
$+19.51
2 (1W/1L)
BNBKXBNBPERP × VBNB 5OI24098
$-25.46
$1.78
$-13.70
$+0.00
$+0.00 EST
$-37.38
1 (0W/1L)
ZECKXZECPERP × ZCSH 198+098
$-466.05
$7.58
$+405.31
$+0.00
$+0.00 EST
$-53.15
4 (0W/4L)
Day (ET)
Kalshi realized
Fees
Funding
Public realized
Net day
Cumulative
09/02
$+0.00
$0.00
$+0.00
$-0.49
$-0.49
$-0.49
09/03
$+0.00
$0.00
$+0.00
$+0.00
$+0.00
$-0.49
09/04
$+0.00
$0.00
$+0.00
$-75.71
$-75.71
$-76.20
09/16
$+5.92
$12.18
$+0.30
$+16.30
$+34.69
$-41.51
09/17
$-501.79
$33.92
$+0.80
$+370.36
$-96.70
$-138.21
09/18
$-107.85
$3.40
$+0.33
$+106.77
$+2.65
$-135.56
Open now — Kalshi: BTC short 164 ct @8.083 mark 8.0592 ·
Public: 0.0164 BTC 198 · experiments: MSTZ $-25.04 · capital deployed $2,643.32.
Reconciliation: Kalshi equity $3,116.18
= entry ref $2,489.73 + implied deposits $1,175.34
(deposits are not exposed by the Kalshi API — derived) + strategy P&L. Open positions
marked at venue marks (Kalshi mark_price; Public position_value = portfolio total −
cash, SMH overnight share excluded). Day-of rows are partial until the 8pm ET funding print
settles. Built 2026-09-18 13:33:39 ET · source:
mn-perp/scripts/mn_scorecard_build.py (read-only venue GETs).
Swinging Balls LIVE
Leg-by-leg profit taking on top of the carry
(mn_skew_take.py, live in the 1-min watchdog). Arm when the winning leg's
unrealized P&L ≥ $6 (48bp of one-side notional) — arms on the LEG, not the
pair (near-balanced notional means a rip moves both equally; a pair trigger
would never fire, which is why XRP sat flat through a 10% move). Once armed:
a ≥$2 giveback from peak skew takes the RAIL long (locks the rip), the naked
perp short keeps earning carry as a carry_lock with two stops — a
$1.50 trail off high-water and a floor at pair skew ≤ $0 that closes the
whole pair, then a 20h cooldown. v1 covers the rip case; the dump case
stays with the P&L kill (taking the winning short first would strand the
rail long with no carry book).
Pair
Stage
Peaks
Armed
No open pairs — engine idle
Arm $6 · giveback $2 · trail $1.50 · floor $0 (breakeven)
· cooldown 20h — read-only view of live watchdog state; execution belongs to
the watchdog, never to this page.
Market Context — S&P 500 Heatmap EMBED
Live TradingView heatmap (SPX500 universe, size = market cap,
shade = change, grouped by sector) for regime context next to the fleet cards.
Rendered client-side by TradingView; no desk data and no scorecard numbers flow
through it. Config mirrors OpenStock's public widget source via the 8bot-desk
market terminal.
Embed is market context only — it is not desk state, not a
scorecard input, and not investment advice. Widget: TradingView stock-heatmap,
dataSource SPX500, grouping sector, dark theme.
Hormuz carry EVENT-ARMED (freight confirm): ShippingWatch alert 37h old (Thu Sep 24, 5:22am ET); Jump short Brent $1.4M as of Fri Sep 25, 5:33pm ET — freight premium confirms the strait risk (BWET 20d mom 54.66% > 20% floor) without Jump long-Brent; tranche-limited book
2 alert(s) demoted as de-escalation pleas
Daily signal · 5:30pm ET cron · 2026-09-25 6:06 PM ET · EST estimate: armed sleeve signal targets, not confirmed holdings.
X Strategy Review — MrMilkTrading NQ "Morning Dip Limit" scalp EST
Post 2103973130108596686
(2026-09-26, @MrMilkTrading) claims an out-of-sample September 2026 run of
Out-of-sample September 2026: 67% win rate, 1.75 profit factor, ~2 trades/day, +$4,650 from 1 NQ. Screenshot label: 'Morning Dip Limit'; backtest on 1-second NQ data with modeled fills and fees, averaged over 10 candle start times..
No rules are published — no level, no stop, no target. The Kaufman-efficiency-ratio "trade only when choppy"
gate is not in the post; it is the forwarded hypothesis and was tested here on its own merits.
Reconstruction: 600-config grid of the canonical morning-dip-limit family (buy-limit at prev close or open minus
k×ATR, stop = sl×ATR, target = rr×sl, flat 15:59 ET, max 2/day) against 441 NQ 1-minute sessions
(Jan 2025 → Sep 25 2026), $4.00 round-turn + 0.5pt stop slippage.
Window (600 configs)
Sessions
Configs net>0
Median PF
Median win%
Trades/day
Median net
discovery 2025-01..2026-06
378
32.8%
0.94
51.6%
1.85
−$9,353
OOS 2026-07..2026-09-25
62
9.5%
0.81
49.3%
1.87
−$7,133
claim month 2026-09-01..25
18
46.3%
0.98
54.2%
1.86
−$102
Claim arithmetic checks out: 18 sessions × 2 trades = 36 fills → $129/trade = 6.5 pts;
at 67% wins / PF 1.75 that implies a 22.5pt average win against a 26.1pt average loss (reward:risk 0.86) —
a small-target wide-stop scalp, no metric incoherence. Measured trade frequency 1.85/day matches the claim.
Kaufman efficiency ratio: "only trade when choppy" — tested
Regime at the fill (pooled, 236k disc trades)
Disc win%
Disc PF
Disc avg
OOS win%
OOS PF
OOS avg
low KER (chop) + last 1h net down
56%
1.17
+$67
52%
0.93
−$40
low KER (chop) + last 1h net up
53%
0.98
−$10
50%
0.79
−$127
high KER (trend) + last 1h net up
51%
0.82
−$78
52%
0.91
−$51
high KER (trend) + last 1h net down
47%
0.86
−$61
46%
0.74
−$175
Signed KER quartile (efficiency × direction of last hour)
Disc PF
OOS PF
trend-DOWN (efficient sell-off)
0.83
0.67
mild-DOWN
0.97
0.95
mild-UP
1.17
0.82
trend-UP
0.96
0.86
The framing is half-right. "Low KER" is really "don't buy a dip into a persistent one-hour
sell-off": high-KER + last-hour-down is the worst cell in both windows (knife-catching), low-KER + last-hour-down
the best in both. But no cell clears PF 1.00 out-of-sample — it is a risk filter, not an edge.
The pre-computable version fails outright: a walk-forward gate on the prior session's KER60
(low quintile of its own trailing 60 sessions) is worse than ungated on 3 of 4 anchor configs OOS
(PF 0.44 / 0.52 / 0.69 vs ungated 0.71 / 0.99 / 0.88).
September 2026 was simply a good month for the family (46% of the grid profitable vs 9.5%
across the surrounding OOS window) — 31 of the
600 configs hit the claimed profile (WR≥60%, PF≥1.5, ≥$4,000) in September, and those same
configs sit at median PF 1.00 outside it. The eight best discovery configs are all OOS-negative (PF 0.71–0.88).
Pooled monthly net per config swings ±$5k on 1 NQ (May 2025 +$5.1k, Jul 2026 −$9.6k, Sep 2026 +$0.3k).
Useful?
YES (regime info)
Generalizable?
NO (as a strategy)
Build?
NONE
Watchlist
KER as risk gate (paper)
YES as a regime-filter data point - 'high KER + recent down' is the reliable loser for any mean-reversion entry, confirmed on 236k discovery / 40k OOS trades. NO - the family itself is negative (OOS median PF 0.81; only 9.5% of 600 configs net-positive; the discovery-top-8 are all OOS-negative). NONE - nothing to deploy. No futures rail on our venues anyway (Alpaca: stocks/options/crypto; Public: stocks/ETFs/options/crypto/prediction markets), and the intraday 2-trades/day scalp needs live intraday execution.
KER-as-risk-gate for existing mean-reversion sleeves (paper study) - test whether gating entries when the last 60 x 1m bars show high efficiency + a down net move improves the ATH-dip / Bilbo family.
Method: Canonical reconstruction of the 'morning dip limit' family: buy-limit at prev-RTH-close (or session open) minus k x prior-day ATR(14) of RTH ranges, valid 09:30 to 11:00/12:00 ET, then stop = sl x ATR and target = rr x sl (time-flat 15:59 ET), max 2 trades/day. 600-config grid x 441 NQ sessions (1-minute, Databento GLBX continuous Jan 2025 - Sep 11 2026 spliced to yfinance NQ=F Sep 14-25; the Databento 2026-09-14 session excluded, continuous-contract roll verified against the yfinance cross-check). Costs: $4.00 round-turn + 0.5pt slippage on stop exits. Discovery window 2025-01-02 to 2026-06-30 (378 sessions), OOS 2026-07-01 to 2026-09-25 (62 sessions), claim month Sep 1-25 2026 (18 sessions). Reviewed 2026-09-27.
All reconstruction numbers are estimates (EST) from 1-minute futures data — his actual fills, sizing and
account P&L are not public.
X Strategy Review — shentrades "$SPY 0DTE Setup" (1-min RSI scalp) EST
Post 2103894939075739988
(2026-09-26, @shentrades): $SPY 0DTE Setup: 0DTE, 1-minute timeframe, S/R levels; RSI<30 = calls, RSI>70 = puts; 15% TP/SL with OCO; max 1-10 minute hold.No stats are given — no win rate, no P&L, no screenshot, no strikes, no costs. The author sells a
Playbook + indicator + Discord.
Test: SPY 1-minute RTH bars (Alpaca IEX), 206,369 bars / 539 sessions (Aug 2024 → Sep 25 2026), RSI(14) cross of
30/70 on 1-minute closes, then the actual ATM 0DTE option priced minute-by-minute with Black-Scholes
(same-day 16:00 ET expiry, fixed IV) so theta and gamma are inside the number; exit at ±15% of entry premium
or the 10-minute stop; spread charged as a round-trip premium haircut.
Window
n
Trades/day
Win%
Mean/trade
Median/trade
signal RSI(14), 10m hold
5,492
10.2
48.2%
−1.35%
−4.31%
signal RSI(7), 10m hold
16,816
31.2
47.9%
−1.39%
−4.50%
signal RSI(14), 5m hold
5,492
10.2
48.0%
−1.22%
−2.10%
BASE RATE any minute, 10m hold
41,018
76.1
48.2%
−1.35%
−3.82%
BASE RATE any minute, 5m hold
41,019
76.1
47.8%
−1.27%
−1.94%
INVERTED (fade the signal), 10m
5,491
10.2
47.4%
−1.60%
—
The RSI filter is decoration. Signal minutes lose exactly what random minutes lose:
48.2% wins / — of premium on the signal vs
48.2% / −1.35% entering at any minute with the same bracket.
Fading the signal is also negative (−1.60%), so the 30/70 cross carries no directional information at a
1–10 minute horizon — what is left is cost. Measured frequency: 10.2
signals/session at RSI(14), 31.2 at RSI(7).
Cost decomposition, regime stability and the caveat
Round-trip spread
Win%
Mean/trade
Per session
0.0%
48.5%
−0.35%
−3.6%
0.5%
48.3%
−0.85%
−8.7%
1.0%
48.2%
−1.35%
−13.8%
2.5%
47.5%
−2.85%
−29.1%
Year (RSI14, 10m, IV 15%, 1% spread)
n
Win%
Mean/trade
2024
1,059
48.2%
−1.32%
2025
2,451
48.8%
−1.18%
2026
1,982
47.5%
−1.58%
Signal = base rate. RSI(14) minutes: 48.2% wins, -1.35% of premium. Random minutes, same option, same bracket: 48.2% wins, -1.35%. The filter selects a subset of minutes and adds no predictive content.
Both directions lose. Fading the signal is also negative (-1.60%), so the 30/70 cross carries no directional information at a 1-10 minute horizon; what is left is cost.
Not a cost artefact alone. With ZERO spread the bracket still loses -0.35% per trade (win 48.5%): a symmetric +/-15% bracket on a convex, decaying option is mildly negative even before you pay anybody.
Stable across regimes: -1.32% (2024), -1.18% (2025), -1.58% (2026). No year, no period, no tweak dimension tested turns it positive.
Frequency is the multiplier: 10.2 signals a session at RSI(14) (31.2 at RSI(7)) at -1.35% of premium each is -13.8% of premium bled per session at a 1% all-in cost.
Caveat, stated: the model holds IV fixed, so it cannot capture an intraday vol pop that would flatter a long-gamma scalp on the winners. That cannot rescue the result, because the comparison that matters - signal vs base rate - uses the identical pricing model.
Useful?
NO (as a trade)
Generalizable?
NOTHING TO GENERALIZE
Build?
NONE
Read
buyer side of our premium sleeves
As a trade: no. As evidence: yes - it prices the buyer's side of the 0DTE premium trade we run on the sell side. NONE. No sleeve, no shadow, no paper lab. Counter-party read only: our premium-selling sleeves (ATR credit spreads, XSP cash-secured puts) are the other side of this bleed.
Method: SPY 1-minute RTH bars from Alpaca IEX (206,369 bars, 539 sessions, 2024-08-01 to 2026-09-25). Signal: 1-minute RSI cross of 30/70 (Wilder, computed on the RTH close series). Trade: buy the ATM 0DTE option (SPY $1 strikes, same-day 16:00 ET expiry), priced minute-by-minute with Black-Scholes at fixed IV so theta and gamma are inside the number; exit at +/-15% of entry premium or the 10-minute time stop; spread charged as a round-trip premium haircut. Fixed IV cannot capture an intraday vol pop. Reviewed 2026-09-27.
Numbers are estimates (EST) from 1-minute IEX bars and a Black-Scholes option model — the author's own fills,
strikes, sizing and account P&L are not public.
X Strategy Review — thetriggertrade "Resting OrderFlow" indicator EST
Post 2103971387161346082
(2026-09-26, @thetriggertrade) is a free open-source TradingView indicator, not a strategy:
Turns the volume footprint into a map of resting liquidity: price rows where buyers or sellers dominated and price has not traded back through, drawn as buy (support) / sell (resistance) nodes plus a period POC node and a live pressure column. No win rate, no P&L, no signal — the guide itself says 'these are usage frames, not signals'.
What it never claims: Pine has no order book access (resting size is INFERRED from executed footprint volume); needs a Premium/Ultimate TradingView plan for footprint data; footprint granularity varies by timeframe and history depth; repaints by design on recent bars; proxy mode is an approximation.
Test: the one falsifiable piece — does a fresh, untouched node hold on first touch? Footprint rows rebuilt in
Python with the script's own documented proxy rule, on NQ 1-minute RTH bars
(165,641 RTH 1m bars, 432 sessions, Jan 2 2025 -> Sep 14 2026) and an independent ES holdout (165,643 RTH 1m bars, same window (holdout instrument)).
Memory 500-bar and 2000-bar rolling footprint memory; freshness node must be untouched for 20 or 60 bars to count as resting; outcome fixed 15-point bracket resolved over 30 minutes, plus forward return at 15/30/60 min;
baseline identical bracket + forward returns at every 7th bar, no level at all.
NQ (calibration)
rows / memory / fresh
n
Up
Down
Baseline up
Edge
Fwd 15m
Fwd 30m
buy
2pt / 500 / 20
963
48.6%
49.1%
47.5%
+1.1pp
+0.78
+1.43
sell
2pt / 500 / 20
896
49.9%
46.3%
47.5%
+2.4pp
+1.65
+0.35
buy
2pt / 2000 / 60
293
50.5%
49.1%
47.9%
+2.6pp
-0.30
+10.90
buy
5pt / 2000 / 60
323
53.6%
45.5%
47.9%
+5.7pp
+3.38
+16.20
sell
2pt / 2000 / 60
288
43.8%
51.7%
47.9%
-4.1pp
+0.10
+3.17
sell
5pt / 2000 / 60
313
45.0%
51.8%
47.9%
-2.9pp
-0.07
+1.85
ES (holdout)
rows / memory / fresh
n
Up
Down
Baseline up
Edge
Fwd 15m
Fwd 30m
buy
5pt / 2000 / 60
302
20.9%
18.9%
14.9%
+6.0pp
+2.45
+2.47
sell
5pt / 2000 / 60
298
18.8%
21.1%
14.9%
+3.9pp
+1.33
+0.95
buy
2pt / 2000 / 60
359
19.2%
24.8%
14.9%
+4.3pp
+1.12
+1.32
sell
2pt / 2000 / 60
321
17.1%
20.2%
14.9%
+2.2pp
+0.65
+0.40
Shallow memory (500 bars) shows nothing on either side: NQ buy nodes resolved up 48.6% vs a 47.5% baseline, and the sell nodes resolved up MORE often than the baseline (49.9%) - the opposite of resistance.
Deep memory (2000 bars) is the only place the claim shows up, and it shows up on both NQ configs: buy nodes 50.5% / 53.6% up vs a 47.9% baseline, sell nodes 43.8% / 45.0% up.
The ES holdout (pre-registered direction: buy-side up-rate at least 3pp above baseline, sell-side at least 3pp below) passes the BUY half in both configs (+6.0pp and +4.3pp, forward 15m +2.45 / +1.12 vs a +0.18 baseline) and FAILS the SELL half in both (18.8% and 17.1% up vs a 14.9% baseline - above, not below).
Cell sizes are ~300 events per side, which puts a 95% CI of about +/-4 to +/-6pp on every up-rate: the surviving buy-side tilt is inside the band where a single holdout can flip the sign.
Tick-footprint upgrade untested: true aggressor side and finer rows need real trade data (Databento NQ trades quoted at $2.53 per week for this window). Not purchased - the free-proxy result did not justify the spend.
Useful as a tool?
YES
Tradable edge?
NO
Build?
NONE
Watchlist
deep-memory buy nodes (tick data)
As a free TradingView tool for a discretionary chart, yes - the guide is unusually honest (no order book claimed, repainting and timeframe dependence both stated). NO tradable edge. The buy side half-replicates (+4 to +6pp up-rate vs baseline on a holdout instrument), the resistance mirror fails on that same holdout, and the effect vanishes at shallow memory. NONE. The recipe is fully specified in the guide and could be reimplemented in Python from Databento ticks, but the level map is descriptive, not predictive - nothing to port into the stack.
Only if a real use case appears: a pre-registered, larger-sample test of the deep-memory buy-node tilt on tick footprint data. Method: The testable core (a fresh node holds on first touch) rebuilt in Python from NQ and ES 1-minute RTH bars. Footprint rows use the script's own documented proxy rule (intrabar close>open = buy, close<open = sell, flat inherits prior direction; the bar's volume spread across the rows its high-low spans) on a fixed price grid, exactly what the indicator does for bars without tick footprint. The indicator is free; the footprint data it needs is not - TradingView Premium/Ultimate ($30-60/mo class). Our own rail already has the same inputs free at 1-minute resolution, which is what this test used.
Reviewed 2026-09-27. Numbers are estimates (EST) from 1-minute bars and the indicator's
own documented proxy - not the author's chart, timeframe, or row settings.
Bilbo Box Breakout LIVEEST
Public account
5OI31211
real capital · long calls only
Equity
$2,456.12
cap $1,474 (60%) · target 4%
Open positions
2
max 3 concurrent · cap 10 calendar days
Last signal scan
0 taken · 16 stale
2026-09-25 3:06 PM ET
vs VOO since start
-1.92pp ✗
scaling gate
Realized (journal)
$0.00
922 journal rows
Symbol
Contract
Qty
Entry
Opened
AMZN
260.0C 2026-10-23
1
$8.35
2026-09-22 10:04 AM ET
NVDA
230.0C 2026-10-23
1
$7.45
2026-09-22 10:04 AM ET
PLTR
187.5C 2026-10-23
1
$8.25
2026-09-22 10:05 AM ET
SMCI
44.0C 2026-10-23
1
$2.62
2026-09-23 11:05 AM ET
MSFT
520.0C 2026-10-23
1
$13.95
2026-09-25 12:04 PM ET
Rules: 5-grey-candle Saty compression box, first non-grey hourly close above box high (10:00–15:00 ET), above daily EMA21, spread ≤5%; exits keyed to the stock (5-min close below box low, +1 ATR trail, 10-day cap); no volume gate.
Entries hourly at :04 past 10:00–15:00 ET · exits every 5 min · hourly equity read from the broker.
Published ledger's live-only record is ~breakeven (PF 1.07) and engine parity is unmeasured — sized to a 60% cap on live equity, no added capital until it beats the index.
Snapshot 2026-09-25 4:15 PM ET
Live Paper Trading Performance LIVE PAPER
23 trading days · Alpaca 1M window · through 2026-09-25
Trailing Days
23
Total Return
-6.96%
Sharpe
-5.14
Sortino
-5.37
Calmar
-7.10
Max DD
-7.70%
Trade Win Rate
24%
Profit Factor
0.42
Expectancy
$-4
Closed Trades
1371
Strategy
Equity
Return
Sharpe
Sortino
Calmar
Max DD
Daily Up %
Best Day
Days
Ndx Momentum EST
$2,025
+9.45%
6.10
92.31
901.59
-0.3%
29%
+5.0%
18
Jump Mirror EST
$4,205
+2.46%
1.44
2.65
10.99
-5.0%
46%
+6.3%
14
Bilbo Paper EST
$7,666
+1.06%
0.78
1.26
3.92
-6.3%
45%
+4.7%
12
Bond Rotation EST
$12,844
+0.78%
2.37
9.44
21.84
-0.8%
36%
+1.4%
12
Benchmark Core EST
$4,120
+0.11%
0.20
0.33
0.65
-3.2%
33%
+3.5%
13
BTC Regime Gate EST
$4,483
-0.99%
-0.57
-1.40
-5.00
-4.9%
38%
+5.8%
9
Earnings Momentum EST
$6,305
-1.20%
-1.31
-1.60
-3.04
-3.9%
43%
+1.7%
24
GPU Tilt EST
$4,719
-4.93%
-2.76
-2.79
-7.32
-9.4%
50%
+3.6%
11
Hormuz Carry EST
$882
-7.81%
-12.04
-12.29
-11.58
-8.2%
17%
+0.4%
7
Unattributed EST
$11,686
-21.11%
-3.75
-3.45
-3.34
-27.5%
43%
+5.2%
24
Maple EST
$6,916
-23.95%
-6.63
-7.99
-4.16
-23.9%
27%
+5.8%
12
8-Bot Options Desk EST
$3,000
—
—
—
—
—
—
—
22
Trade Quality (FIFO round-trips)
Strategy
Closed Trades
Win Rate
Profit Factor
Expectancy
W/L Ratio
Trades/wk
Ndx Momentum
8
75%
18.81
$22
6.27
1.7
Jump Mirror
27
44%
1.34
$4
1.67
5.9
Bilbo Paper
3
33%
0.42
$-46
0.83
0.7
Bond Rotation
15
0%
0.00
$-2
—
3.3
Benchmark Core
13
38%
1.07
$0
1.71
2.8
BTC Regime Gate
20
40%
0.46
$-2
0.69
4.3
GPU Tilt
285
25%
0.67
$-1
1.98
62.0
Hormuz Carry
2
0%
0.00
$-37
—
0.4
Unattributed
753
20%
0.46
$-3
1.82
163.7
Maple
182
26%
0.11
$-10
0.30
39.6
FIFO-matched Alpaca fills · excludes cash venues & dust · what a quant checks before trusting a Sharpe
Alpaca get_portfolio_history · EST fill-timeline × closes · not a backtest
EST P&L residual $-2.65 (0.00% NAV) · BIL/SHV 0%
EST equity gap $10,096.07 (12.79% NAV) · desk allocation $3,000.00 · account NAV through 2026-09-25 · desk through 2026-09-27
Since First Live Session LIVE PAPER
Portfolio vs $VOO vs $GOOG — all indexed to 100 on 2026-08-10 (first live paper session, post-reset $100k account). Benchmarks: split-adjusted closes, same calendar dates.
Cumulative % Return
Portfolio$VOO$GOOG
+3%-3%-9%
Aug 10 '26Sep 1Sep 25 '26
Latest: Portfolio -6.3% · $VOO +0.0% · $GOOG -4.2% · touch or drag a finger across the chart to inspect any date · sources: Alpaca account history + Yahoo Finance daily closes
Capital Allocation · No Margin
Reserve Rotation
0%
$0.00 EST
Memory Rotation
0%
$0.00 EST
Bond Rotation
76%
$0.00 EST
DRAM
0%
$0.00 EST
VNQ Reversion
0%
$0.00 EST
Maple
0%
$0.00 EST
Earnings Momentum
0%
$0.00 EST
Hormuz Carry
0%
$0.00 EST
Ndx Momentum
0%
$0.00 EST
GPU Tilt
24%
$57,409.22 EST
BTC Regime Gate
0%
$0.00 EST
Braided-Delta
0%
$0.00 EST
Waterhole
0%
$0.00 EST
Benchmark Core
0%
$29.54 EST
Bilbo Paper
0%
$9,807.23 EST
Jump Mirror
0%
$0.00 EST
Available Cash: $3,064.43Total Deployed: $75,883.29 / $78,947.73
Zero-allocation/deallocated positions (pending liquidation): $18,474.08Execution-readiness reserve (releases to core when two consecutive clean sessions are proven): $0.00 (0%)
Trading Discipline ESTSCORE 26
score, last 28 session(s)
✗Max Daily Loss ESTworst day -3.23% vs -2.0% limit
✓Max Drawdown EST-7.70% vs -8.0% limit
✗Guard Hygiene EST4 active alert(s): LEVERAGED_NAME_OVER_5PCT_DIG_10.1pct, SINGLE_NAME_OVER_10PCT_BIL_72.7pct, SINGLE_NAME_OVER_10PCT_DIG_10.1pct…
✓Active Trading Days EST22 of last 23 sessions vs 16 minimum
✗Beat the Market ESTbook -6.96% vs SPY +0.96% same window; alpha -7.92pp
✓No Margin ESTcash balance $3,064.43; negative cash indicates a debit
✗Allocation Drift EST2 sleeve(s) beyond ±2.5pp of target
EST · Score covers 7/7 objectives · account through 2026-09-25 · risk 2026-09-25 9:30 PM ET