Flash Report

11 primary ideas (WDC long, FAST long, HO=F long, ICE long, TXN long, TMUS long, ...) - 3 clear every desk rule | funnel: 1 trade today: VZ long · 9 on watch · data degraded · 11 book alerts

Feed date 2026-10-28 · built Oct 29, 2026 07:15. Every idea from every research feed, run through the house procedures.

1Can I trust today's data?

Data degraded. Read with care: Nightly build: started 10-27 18:30, finished 10-28 09:01 (14.5 h); 1 traceback(s) in the run; Feed contracts: dt_ideas_weekly.json: dated 2026-10-23, older than 2026-10-27; dt_ideas_weekly_spreads.json: dated 2026-10-23, older tha

★Today's primary ideas · Daily Testing + Range Reversals

Session 2026-10-29 · 11 ideas passed the house truth test on time-weighted history (2026 ×10 · Tariffs→2025 ×8 · Apr 2023→Tariffs ×6 · earlier ×1; edge ≥ +5 pts, p < 0.10 on the weighted sample) · 3 clear every desk rule (Blend > 0, no earnings in the window, no conflict with the book, clean data, one bet per correlated cluster: 60-session corr ≥ 0.7 or the same regime theme).

TRADEWDCLONGRange-Close reversal · Completed reversal from the low

Win vs base71% vs 65%+6 pts · p 0.06

Median move+6.69%in the trade's favour · 21 sessions

Analogs324effective n 138 (weighted)

Size (Blend)6.00% NAV = $44,412 (Kelly 7.5% · Brier 7.5%; options risk capped at 0.75% NAV)

Structure	Long stock
Levels	ref 396.42 · target 422.94 · thesis fails 392.85 · time exit 2026-11-27
Book	not held · ADDS TO held: STX (live) (memory theme, corr 0.81)

Name Report →Source page →

TRADEFASTLONGRange-Close reversal · Completed reversal from the low

Win vs base67% vs 60%+7 pts · p 0.03

Median move+3.18%in the trade's favour · 21 sessions

Analogs399effective n 173 (weighted)

Size (Blend)6.00% NAV = $44,412 (Kelly 7.5% · Brier 7.5%; options risk capped at 0.75% NAV)

Structure	Long stock
Levels	ref 51.45 · target 53.09 · thesis fails 50.37 · time exit 2026-11-27
Book	not held

Name Report →Source page →

TRADEHO=FLONGDaily Testing · commodities · 🔨 Hammerolder session 2026-10-28

Win vs base70% vs 54%+16 pts · p 0.00

Median move+3.59%in the trade's favour · 10 sessions

Analogs68effective n 67 (weighted)

Size (Blend)6.00% NAV = $44,412 (Kelly 7.5% · Brier 7.5%; options risk capped at 0.75% NAV)

Structure	Short put spread (bull put) Sell the 4.52 put, buy the 4.34 put (3–7% below spot). High-prob, modest move — win if it holds, rises, or dips slightly.
Levels	ref 4.66 · target 4.83 · thesis fails 4.61 · time exit 2026-11-11
Book	not held · leads one bet with: CL=F

Name Report →Source page →

WATCHICELONGRange-Close reversal · Completed reversal from the low

Win vs base65% vs 56%+9 pts · p 0.01

Median move+1.44%in the trade's favour · 10 sessions

Analogs391effective n 170 (weighted)

Size (Blend)0 · no trade (Kelly 0.0% · Brier 5.0%)

Structure	Long stock
Levels	ref 155.96 · target 158.21 · thesis fails 154.40 · time exit 2026-11-12
Book	not held
Why watch	desk Blend 0 (Kelly f -0.14: a +1.44% median payoff is too small for a 65% win); earnings in the window: ICE earnings 2026-10-29

Name Report →Source page →

WATCHTXNLONGRange-Close reversal · Completed reversal from the low

Win vs base63% vs 55%+8 pts · p 0.01

Median move+1.58%in the trade's favour · 10 sessions

Analogs371effective n 151 (weighted)

Size (Blend)0 · no trade (Kelly 0.0% · Brier 4.8%)

Structure	Long stock
Levels	ref 309.83 · target 314.73 · thesis fails 303.94 · time exit 2026-11-12
Book	not held
Why watch	desk Blend 0 (Kelly f -0.13: a +1.58% median payoff is too small for a 63% win)

Name Report →Source page →

WATCHTMUSLONGRange-Close reversal · Completed reversal from the low

Win vs base60% vs 52%+8 pts · p 0.03

Median move+0.30%in the trade's favour · 5 sessions

Analogs288effective n 134 (weighted)

Size (Blend)0 · no trade (Kelly 0.0% · Brier 0.9%)

Structure	Long stock
Levels	ref 148.08 · target 148.52 · thesis fails 146.01 · time exit 2026-11-05
Book	not held
Why watch	desk Blend 0 (Kelly f -3.57: a +0.30% median payoff is too small for a 60% win)

Name Report →Source page →

WATCHPLTRLONGDaily Testing · 🔨 Hammerolder session 2026-10-28

Win vs base68% vs 62%+6 pts · p 0.07

Median move+7.91%in the trade's favour · 21 sessions

Analogs193effective n 128 (weighted)

Size (Blend)6.00% NAV = $44,412 (Kelly 7.5% · Brier 7.5%; options risk capped at 0.75% NAV)

Structure	Long calls Slightly-ITM / ATM calls (~196.81), 4–8 weeks out. High odds + a 8% expected move — convex upside, risk = premium.
Levels	ref 196.81 · target 212.38 · thesis fails 193.27 · time exit 2026-11-26
Book	not held
Why watch	earnings in the window: PLTR earnings 2026-11-02

Name Report →Source page →

WATCHCL=FLONGDaily Testing · commodities · 🔨 Hammerolder session 2026-10-28

Win vs base62% vs 54%+8 pts · p 0.10

Median move+4.16%in the trade's favour · 21 sessions

Analogs69effective n 67 (weighted)

Size (Blend)6.00% NAV = $44,412 (Kelly 7.2% · Brier 7.5%; options risk capped at 0.75% NAV)

Structure	Long call spread Buy ~ATM (102.78), sell ~107.05 (about the expected move wide). Strong edge, defined risk.
Levels	ref 102.78 · target 107.06 · thesis fails 99.70 · time exit 2026-11-26
Book	not held
Why watch	same bet as HO=F (60-session corr 0.70, energy theme)

Name Report →Source page →

WATCHPL=FSHORTDaily Testing · commodities · ⭐ Starolder session 2026-10-28

Win vs base54% vs 46%+8 pts · p 0.09

Median move+2.10%in the trade's favour · 21 sessions

Analogs69effective n 67 (weighted)

Size (Blend)0 · no trade (Kelly 0.0% · Brier 5.5%)

Structure	Long puts ~ATM puts 4–8 weeks out (or short the stock with a stop above the signal high). Moderate down-edge, risk = premium.
Levels	ref 1,540 · target 1,507 · thesis fails 1,624 · time exit 2026-11-26
Book	not held
Why watch	desk Blend 0 (Kelly f -0.14: a +2.10% median payoff is too small for a 54% win)

Name Report →Source page →

WATCHFDXLONGDaily Testing · 🔨 Hammerolder session 2026-10-28

Win vs base65% vs 56%+9 pts · p 0.01

Median move+1.80%in the trade's favour · 10 sessions

Analogs310effective n 135 (weighted)

Size (Blend)4.00% NAV = $29,608 (Kelly 0.2% · Brier 6.2%; options risk capped at 0.75% NAV)

Structure	Short put spread (bull put) Sell the 295.44 put, buy the 283.26 put (3–7% below spot). High-prob, modest move — win if it holds, rises, or dips slightly.
Levels	ref 304.58 · target 310.06 · thesis fails 299.71 · time exit 2026-11-11
Book	not held
Why watch	earnings in the window: FDX earnings 2026-10-28

Name Report →Source page →

WATCHLE=FSHORTDaily Testing · commodities · ⭐ Starolder session 2026-10-28

Win vs base54% vs 37%+17 pts · p 0.05

Median move+0.65%in the trade's favour · 21 sessions

Analogs25effective n 24 (weighted)

Size (Blend)0 · no trade (Kelly 0.0% · Brier 2.4%)

Structure	Long puts ~ATM puts 4–8 weeks out (or short the stock with a stop above the signal high). Moderate down-edge, risk = premium.
Levels	ref 227.32 · target 225.84 · thesis fails 235.05 · time exit 2026-11-26
Book	HELD same side: LEZ6 (live)
Why watch	desk Blend 0 (Kelly f -1.71: a +0.65% median payoff is too small for a 54% win)

Name Report →Source page →

Second opinion · the full funnel (every research feed, independent re-test)

2What we trade

VZ · long

L2 exhaustion · 2 independent research families: Daily Candidates · trade, Trigger Portal

Why: re-test on its own history (10 sessions)66.7% vs 51.8% base · n=63 · p=0.0225
before 2023 62.5% · since 2023 100.0%

Signal50-day MA -7.6% (5th percentile) · typical move +2.08%

Size (Kelly + Brier Blend)4.0% of NAV = $29,608 (Kelly 1.77%, Brier 7.5%)

Expressionlong VZ

Option alternativebull put spread ~8% wide: sell ~41.45 put / buy ~38.0 put · max risk $5,552

Exits (set before entry)ladder 20/30/30/20 at +1.8% / +3.7% / +4.8% / +5.9% · hard stop -4.0% · time exit 2026-11-11

Events in the windowFOMC decision 2026-10-28, Payrolls 2026-11-06, CPI 2026-11-10

Book check
no overlap with the open book

Name Report →Research Cockpit →

3What almost made it

Name	Side	Re-test	Blend	Why it is not a trade	Book
GDX/GDXJ	short	58.4% vs 49.4% n=101 · 10d	0.0%	Blend 0 (Kelly or Brier sees no edge)	—
^GSPC/RSP	long	58.1% vs 46.4% n=93 · 10d	0.0%	Blend 0 (Kelly or Brier sees no edge); index ratio - not directly tradeable	adds to held QQQ long (60-session corr +0.65)
^OVX/^VIX	short	60.7% vs 50.4% n=150 · 10d	6.0%	index ratio - not directly tradeable; called both ways by the feeds	—
RSP/^GSPC	short	57.0% vs 46.4% n=93 · 10d	0.0%	Blend 0 (Kelly or Brier sees no edge); index ratio - not directly tradeable	adds to held QQQ long (60-session corr +0.65)
SPY/RSP	long	59.8% vs 49.9% n=92 · 10d	0.0%	Blend 0 (Kelly or Brier sees no edge)	adds to held QQQ long (60-session corr +0.65)
SPYV/SPY	short	66.7% vs 55.3% n=78 · 10d	0.0%	Blend 0 (Kelly or Brier sees no edge)	adds to held QQQ long (60-session corr +0.85)
XLK/XLP	long	61.9% vs 54.7% n=139 · 10d	0.0%	Blend 0 (Kelly or Brier sees no edge)	adds to held GEV long (60-session corr +0.60); adds to held QQQ long (60-session corr +0.82); adds to held SKHY long (60-session corr +0.66)
SMH/IWD	long	68.3% vs 58.0% n=82 · 10d	6.0%	same bet as ^GSPC/RSP (60-session corr 0.73)	adds to held STX long (60-session corr +0.64); adds to held GEV long (60-session corr +0.66); adds to held QQQ long (60-session corr +0.77); adds to held SKHY long (60-session corr +0.74)
JNJ	long	70.5% vs 56.8% n=61 · 21d	0.0%	TEF: momentum (off-mandate); Blend 0 (Kelly or Brier sees no edge); called both ways by the feeds	—

164 names entered from every research feed; 154 were cut (fails the independent re-test (147), no usable price history (synthetic basket or too short) (7)). The independent re-test passed 10 of 314 tests - a few passes happen by chance alone, which is why a name needs every procedure, not just one.

4What this does to the book

Each card and watch line above carries its book check: same name held (agrees or conflicts) and 60-session correlation with every open position. Held names reporting, and macro releases, in the next 10 sessions:

CPI 2026-11-10
Payrolls 2026-11-06

5Are our procedures working?

The decision scorecard

126 decisions on record (10 overrides of the house rules), 38 closed trades scored. Each step of the decision function is scored against a version without it; under 10 decisions a step says "too few". The scorecard never changes a rule - FLAGs are for Brad.

Step	n	Measured	Verdict
N3 Ideas claimed win vs realized hit, per protocol	193	idea ledger outcomes (first ones mature ~Oct 16) - full table in the Research Cockpit	measured - no counterfactual yet
N4 Filter TRADE tier vs WATCH tier vs every primary idea (forward return at horizon)	67	primary ideas logged since go-live 2026-10-22: trade 13, watch 54; scored once their horizons pass	logging - scored once horizons pass
N5 Sizing P&L of the sizes used vs the same trades at equal capital, within each instrument class (credit spreads excluded)	0	too few closed trades per class	too few (0 < 10)
N7 Overrides (entries) override entries vs rule entries - win rate and mean return on capital	0	overrides n=0 win None% · +0 total (+0/decision) \| rule n=38 win 42.1% · -6,808 total (-179/decision)	too few (0 < 10)
N7 Overrides (exits) discretionary exits vs rule exits (realized at the exit; what-happened-next comparison pending)	61	brad: 61 exits, -959	measured - counterfactual (price after the exit) pending
N8 Execution fill vs IBKR mid at decision time	0	the model ledger fills at the mid by construction - measurable once broker fills exist	not measurable on a model ledger
N9 Exits trades ever green (on daily marks) that finished flat or red; capture = realized / best daily-mark P&L	26	median capture 51% · 10 of 26 trades that were ever green finished flat or red	behind its counterfactual (watch)

By research source (closed trades, n ≥ 5)

Source	n	Win	Median return	Realized
(none)	38	42.1%	-0.08%	-6,808

Latest decisions

When	Kind	Who	Name	Rule said	Decided
2026-10-28 18:25	exit	brad	NQZ6	Brad's exit in TWS (live account)	closed 3.0 for -464
2026-10-28 17:53	exit	brad	GOOGL	Brad's exit in TWS (live account)	closed 630.0 for -186
2026-10-28 17:51	exit	brad	MSFT	Brad's exit in TWS (live account)	closed 300.0 for -2,299
2026-10-28 17:50	enter	brad	GOOGL	trade	long_equity 700.0 @ 290.888906
2026-10-28 17:43	enter	brad	MSFT	trade	short_index 300.0 @ 544.0
2026-10-28 14:42	exit	brad	BX	Brad's exit in TWS (live account)	closed 100.0 for -51
2026-10-28 14:32	exit	brad	CLZ6	Brad's exit in TWS (live account)	closed 4.0 for -1,453
2026-10-28 14:32	exit	brad	CLZ7	Brad's exit in TWS (live account)	closed 4.0 for +272
2026-10-28 14:06	exit	brad	XLP	Brad's exit in TWS (live account)	closed 500.0 for +633
2026-10-28 14:03	enter	brad	LEZ6	trade	short_index 1.0 @ 2.269922
2026-10-28 13:05	exit	brad	RTX	Brad's exit in TWS (live account)	closed 10.0 for -408
2026-10-28 13:03	exit	brad	ZMZ6	Brad's exit in TWS (live account)	closed 5.0 for -1,172

Idea ledger: claimed vs actual

Protocol	Scored ideas	Claimed win	Actual hit	Gap	Mean return
Best Trades (trade-type engine)	64	61.5%	60.9%	-0.5 pts	+0.40%
Testing 15	42	62.4%	52.4%	-10.0 pts	+0.21%
Daily Candidates · watch	34	62.0%	50.0%	-12.0 pts	-0.03%
Weekly Ideas	17	64.4%	47.1%	-17.4 pts	+0.25%
Daily Testing	14	62.1%	35.7%	-26.4 pts	-3.15%
MA Stretch Lab	10	66.8%	40.0%	-26.8 pts	-1.73%
Weekly Spreads	10	55.9%	60.0%	+4.1 pts	+2.86%
Gap-Open Regime	2	56.5%	0.0%	-56.5 pts	-6.58%

What changed

Since 2026-10-26: new trades VZ long · no longer trades none · new on watch GDX/GDXJ short, JNJ long, SMH/IWD long, SPY/RSP long, SPYV/SPY short, XLK/XLP long, ^GSPC/RSP long, ^OVX/^VIX short.

Research only - nothing here places a trade. Option strikes are indicative; price the structure on the live chain at entry. Engine: candidates.py · flash.py.