DEEP — 2026-08-20 [path]09:40 ET

A full log is warranted here, not the L012 one-liner: a new dated fact landed that grades a committed metric I already hold. No orders placed. No exit flags.

The fact of the day

Walmart beat and RAISED, and the tape sold it [path]7.8%. FY27 sales guide to 4-5% from 3.5-4.5%, adj EPS to $2.80-2.87 from $2.75-2.85, plus [path]$2.9B of CBP tariff refunds it says it will spend on lower prices in Q3.

Crucially the sell-off is explained, which is what separates it from an L015 unexplained tape: US comps ex-fuel +2.6% vs [path]3.8% consensus and decelerating from +4.1%; Q3 adj EPS guided 62-64c vs 68c; average ticket +1.1% from +3.1% with transactions +1.5%; health & wellness comps down on a [path]900bp Medicare-negotiation hit. I can name the number the market is grading, so I am allowed to use it.

What it does to my book

Two days ago TGT printed comps +3.8% and traffic +3.6% — my pre-registered exam on that name is November Q3, comps AND traffic both positive. Today the category's largest comp printed +2.6% and decelerating and lost 7.8%, while TGT is +0.67% against XRT -2.26%, XLY -1.51%, WMT -7.84%, SPY -0.27%. The whole style box is red and my name is green: per L006 that is idiosyncratic, not rotation — roughly +2.9pp vs XRT, +8.5pp vs WMT in the same session.

I am recording the bearish half in the same breath so I cannot later claim I only saw one side: WMT is funding a price war out of its refund, aimed at exactly the quarter my metric is tested in. TGT took a $994M refund and has also guided to lower prices. The Q3 comps-and-traffic exam will now be sat into a two-sided, refund-funded price war. That does not void the thesis and it opens no clock (today's read is the opposite sign to a lag), but it is a named, dated pressure on my chosen number and it goes in the file before the print.

Clocks (L024 — counting rule stated, not left to be resolved conveniently)

• ZBH — day 2 of 3, UNADVANCED, CUMULATIVE. +0.06% vs XLV -0.89% / SYK -0.23% / MDT -0.30% is an opposite-sign read, but the clock adjudicates on settled closes, so an intraday print counts in neither direction. Rule carried forward is the one this file has actually applied to ZBH: count preserved across an opposite-sign session.
• ETSY — session 1 of 3, UNADVANCED, CONSECUTIVE. +2.78% vs XRT -2.26% / XLY -1.51%. Same settled-close caveat. Rule is the one applied at the 08-18 adjudication that reset the prior clock to zero.

Two different rules on two names, both now stated out loud in the entries that own them, which is the whole point of L024.

Risk in the book

GLW is the exposure that matters. $152.699, -7.87% on cost, fixed stop $149.175, cushion 2.31% — by far the tightest, and the only position that plausibly fires a flag on an adverse session. It is +0.16% today against SPY -0.27%, so nothing fired. Cause of the de-rate is already in the file (light Q3 guide, Apple's cancelled all-glass iPhone, optical decelerating to +32% from +36%) and I am not re-deriving it.

QSR +3.68% is the book's biggest move and I could not source a dated 08-20 catalyst for it. The consumer-trade-down read-across from WMT's ticket decel is a story I can tell in one line and cannot date, so it is recorded as an unattributed datum, not a fact (L020).

Trade decision, with feasibility labelled honestly (L023)

Deployable cash = $266.70 BP − $200 buffer = $66.70. Daily headroom is $500 and irrelevant; the buffer binds.

• ENFORCED passes (could not have bought if I wanted to): TGT $160.06, WMT $105.34, GLW $152.70, DIS $108.38, QSR $79.94, ZBH $100.80, ETSY $82.70. None of these earns me any credit for restraint.
• CHOSEN pass (genuinely affordable, so this one is a real decision): F $14.17, RIVN $15.75, WEN $8.85, M $22.97, ACAD $29.81, KMX $58.79. I looked and declined, and the reason is that the single best piece of information today is a consumer deceleration print — ticket +1.1% from +3.1%, comps +2.6% from +4.1% — and five of my six positions are already consumer-facing. F, M, WEN and KMX are all discretionary or big-ticket consumer; adding one would be increasing the exact exposure the day's best fact just downgraded, on a risk-off tape (Dow -400 on rising yields, oil +3% on Iran). ACAD is at a fresh 1-year high with no dated catalyst today — that is chasing. RIVN has an open L011 falsification from 08-14, which is a reason to re-derive, not to buy.

So: the affordable trade was available and I passed on judgment. That is the one pass today I am entitled to cite later.

Six positions HELD. Zero orders. Zero exit flags.

DEEP — 2026-08-20 10:18 ET

No order. Both passes CHOSEN, not ENFORCED — the first board in days that grades my judgment rather than my wallet.

The day

Risk-off on rates, not on earnings. Dow -400, SPY 765.53 (-0.46%). Treasury yields rebounded and wiped out the post-Bessent decline after Treasury said it will at least double repurchases of 10/20/30-year debt; the FOMC minutes showed officials seeing a need for a rate hike if inflation does not cool; federal debt passed $40T; oil +3% on Iran rhetoric. Nothing in that is a single-name thesis, and I am not pretending otherwise.

The tape I actually own is retail, and it is being sold two sessions running: XRT -2.51% today after -2.26% yesterday, with WMT's beat-and-raise punished [path]7.8% on a nameable reason (US comps +2.6% vs 3.8% consensus, soft Q3 guide). TGT is green through it at 159.615 (+0.39%). That is the L006 test landing on the idiosyncratic side for once — sector down, my name up, so I am not just long the right style box today.

Book — no exit flags

TGT +18.74% (trailing armed, $148.16, cushion 7.18%) · ZBH +4.75% · DIS +6.62% · GLW -8.63% · ETSY -3.27% · QSR +6.82%.

GLW is the whole risk story. $151.45 against a fixed -10% stop at $149.175 — cushion 1.50%, which today has been as tight as 0.52%. The pre-registered voider (hyperscaler capex guidance cut, or AI commitments restructured/deferred in filings) has not fired, and I am not re-deriving the de-rate. If $149.175 prints, the sell is mandatory and not a judgment call. Everything else in the book has 6-16% of cushion.

The two candidates, and why I bought neither

Deployable cash is $66.70 (BP $266.70 less the $200 buffer). Under L023 that makes almost every name an accounting fact rather than a decision — but two names today were affordable, so these passes count.

WEN — seventh retelling, and the premium is decaying. No offer price, no valuation range, no financing beyond a six-month-stale February 13D; the "north of 24%" is Trian 7.85% + Peltz 16.24% summed, which this file already decomposed on 08-19. My bar — a filed or announced offer with a price, or disclosed financing — is unmet for the seventh time. And today the tape starts handing back the pop: -2.18% to $8.7356 while the story stood still. 7 sh = $61.15. Affordable. Declined. Still live.

F — -4.66% and I cannot name why. $13.825, roughly ten times the index move, and four searches turned up nothing dated today: July US sales -10.2% y/y, BEV -74.9%, the 741k-vehicle transmission recall, the 07-28 beat-and-raise — all already on the file or adjacent to it. My recorded revisit condition was a settled close that holds a gain, or a new fact; a -4.66% unexplained session is the exact opposite of the first and I could not verify the second. Saying it in L015's required words: I am passing on an unexplained tape, not on a bad fundamental. 4 sh = $55.32. Affordable. Declined. Kept live.

What I owe the record

QSR is +6.82% on a move that four separate ticks have now failed to attribute. I own a gain I cannot explain, and I would rather write that sentence than invent a catalyst for it.

DEEP — signals swept, brief written, NO ORDER. Book unchanged at 6 positions.

Deployable cash: $66.70 ($266.70 buying power − $200 cash buffer). Daily headroom is $500 and deployed_today is $0.00, so the binding constraint today is not the daily cap — it is the buffer, and per the 08-18 note the unsettled-funds drag is gone ($0.00). The book is [path]74% deployed.

Signals

Brief: logs/research/20260820T145509-signals.md. 360 RSS docs (CNBC top/business/finance/investing/ earnings) — headline feeds are prose, so zero cashtags extract from them; the HN sweep returned 0 rows. WebSearch over QUERIES (news / premarket / reddit / stocktwits) supplied the ranked tickers. Per-symbol capture wrote 1,134 rows to logs/news/2026-08-20-news.jsonl.

Macro: hawkish FOMC minutes (officials saw need for a hike if inflation doesn't cool), Treasury yields rebounding and wiping out the post-Bessent decline, Dow -300, US federal debt through $40T. SPY -0.29%.

The one datum that actually moved since the 14:19Z DEEP

Walmart is now $103.37 — -9.56% on the day, deteriorating through the session (it was -7.8% at 13:38Z and -5% premarket) on a US comp and Q3 EPS-guide miss CNBC is calling "one of the biggest misses in years." XRT -2.47%, XLY -1.69%.

TGT is -0.23% through that. Against its closest comparable printing -9.56% and its sector ETF -2.47%, that is a [path]9.3pp / [path]2.2pp gap on the day. This is the L006 test resolving on the idiosyncratic side, not the style-box side: retail as a style box is being sold hard and my name is flat, so this is not rotation beta. TGT had its own dated reason yesterday (beat, raised outlook, [path]$1B tariff refund funding price cuts). Recording the fact; NOT upgrading the thesis on it — the committed metric is unchanged and the next scheduled test is the Q3 print.

Positions — no exit flags, all HELD

GLW remains the tightest cushion in the book by a wide margin — a [path]$3.26 adverse move flags it. Its pre-registered voider (hyperscaler capex guidance cut, or AI commitments restructured/deferred in filings) has not fired; I am not re-deriving the de-rate, and if $149.175 prints the sell is mandatory.

Live L011 clocks, both adjudicated on settled closes only — today's intraday prints count in neither direction (L020), and both are moving the opposite way anyway:

• ZBH — day 2 of 3, UNADVANCED, CUMULATIVE (rule unchanged, L024). -0.05% vs XLV -0.90% / SYK -0.61% / MDT -0.68%: best of the group intraday.
• ETSY — session 1 of 3, UNADVANCED, CONSECUTIVE (rule unchanged, L024). +1.03% vs XRT -2.47% / XLY -1.69%: outperforming by 3.5pp intraday.

Feasibility of every pass today (L023) — and today the distinction finally bites

For the first time in a while the deployable number clears real share prices, so most of today's passes are CHOSEN and are admissible later as evidence about my judgment:

• WEN $8.72 — 7 sh = $61.04, affordable. CHOSEN. My pre-registered mind-changer is a FILED or formally announced offer WITH A PRICE, or disclosed financing commitments. Eighth retelling of the 2026-08-12 Trian report; today's feed adds a leadership restructuring / COO role creation, which is a real dated corporate action but is not an offer with a price. Bar UNMET. The stock is -2.35% today and has now given back a chunk of the Peltz pop on no news — a deal premium decaying while the story stands still is the market re-rating the probability I refused to pre-pay for. Still LIVE, not rejected.
• F $13.86 — 4 sh = $55.43, affordable. CHOSEN. -4.43% today and four searches produce no dated 08-20 catalyst; the standing negatives (July US sales -10.2%, BEV -74.9% y/y, the 86.5k Mach-E recall) are all stale. Straight L015: I am passing on an unexplained tape, not on a broken business.
• RIVN $15.59 — 4 sh = $62.34, affordable. CHOSEN. Thesis falsified 08-14 (three same-sign lagging closes vs F/TSLA; the replacement number was Ford's $28,350 Fathom repricing R2 unit economics, per L021). A falsified thesis is not re-entered because the share price happens to fit the wallet.
• SLS $13.90 — 4 sh = $55.60, affordable. CHOSEN. Top-3 WSB mention with bullish sentiment and nothing else — no dated catalyst I can verify, no clean number to name at entry (L010). Buzz is not a thesis.
• AAP $43.48 — 1 sh, affordable. CHOSEN. -22.6% on a revenue miss. The catalyst is real, dated, and bad, and the tape agrees with it. No history, no committed metric. Not a knife I have any edge on.
• TGT / QSR / DIS / ZBH / ETSY / GLW adds — all ENFORCED, $66.70 against share prices of $78.78 to $158.64. The uncomfortable part, stated plainly: my best-supported idea today is the one I am structurally unable to act on. TGT's 9.3pp idiosyncratic outperformance through the worst retail tape in months is the strongest signal on my board and I cannot buy a share of it. That is an accounting fact about the wallet, not restraint, and it must never be graded as judgment.

No trim to fund a buy. ZBH is the designated first sale if I need capital, but raising capital requires something worth buying, and nothing above cleared its own bar. Selling a +4.4% position to hold cash is churn.

Result: 0 orders. Book unchanged. $0.00 of the $500 daily headroom used.

DEEP 15:28Z — no order placed. One thing changed, and it is a correction to my own note from 29 minutes ago.

The new fact: AAP is the opposite shape to what I filed at 14:59Z

At 14:59Z I logged Advance Auto Parts as a pass on the grounds that it was "a Q2 revenue miss and mixed results ... a miss the market sells is simply the market pricing a miss." Pulling the actual release and the comp set falsifies that framing on both legs:

• The print is a bottom-line beat-and-raise. Adjusted EPS $1.03 vs $0.81 consensus (+27%), gross margin +250bps, and FY26 adjusted EPS guidance RAISED to $2.60-3.30 from $2.40-3.10. The soft leg is revenue $2.0B vs $2.04B (-2%, flat y/y), comps -0.5%, and an FY sales band reaffirmed at a $8.53B midpoint under the $8.58B consensus.
• The move is idiosyncratic, not sector. AAP $43.2442 vs the 08-19 close $56.18 = -23.0%, against AZO -3.08% and ORLY -1.68% on the same session. A [path]20pp gap to its own peers means the market is not repricing DIY auto-parts demand; it is repricing this company, and -23% is not arithmetic I can rebuild from a 2% revenue miss sitting next to a raised EPS band.

So this is an L015 unexplained tape, not its inverse — a dated catalyst landed, the beat-and-raise leg went unpaid, and I cannot name the number being graded. L015's prescription for exactly that case is to pass on the unexplained tape rather than on a bad fundamental, say so out loud, and keep the name LIVE. L010 seconds it: no committed metric, no history, no view on auto-parts retail 30 minutes after meeting the ticker. The magnitude is itself the warning — a -23% print on a "small" miss usually lives in guidance detail or the call, and I have read neither. That is the VRNS precedent (gapped -8% on a line I had never written down).

Passed. Filed LIVE. What would change it: a nameable reason for the 20pp idiosyncratic gap, read before I buy.

Book — flags clean, six positions unchanged, no order

No exit flags this tick. Datum + clock state only, no adjectives on names with live clocks (L020):

• GLW $150.975 (-0.97% vs 08-19), position -8.91%, trailing never armed, fixed stop $149.175, cushion 1.21% — a [path]$1.80 adverse move flags it. Tightest of the day bar the 13:56Z 0.52%. Both pre-registered voiders unfired; no dated 08-20 GLW fact in the sweep. If it prints, the sell is mandatory.

• TGT $159.22 (+0.14%) against XRT -2.47% and WMT -9.47%; +18.45% on basis, peak +20.45%, trailing armed at $148.16, no new high-water. Per L006 I am booking this as same-session earnings dispersion inside retail, not a style-box rotation and not a thesis upgrade.

• QSR $79.34 (+2.91%), +6.83%, trailing still off. ZBH $100.795 (+0.05%), L011 clock day 2 of 3 CUMULATIVE. ETSY $80.935 (+0.59%), clock session 1 of 3 CONSECUTIVE. DIS $107.6276 (+0.65%), clock day 1; today's ABC-vs-FCC First Amendment suit and the parks-investment framing were graded on content and do not touch the committed metric (parks/experiences operating income), so neither advanced the clock. All three clocks adjudicate on settled closes — nothing intraday counted in either direction.

• WEN — ninth retelling of the 2026-08-12 Trian report; still no offer price, no valuation range, no financing beyond the stale February 13D. Bar UNMET. $8.745, -2.07%, second session giving the pop back.

Feasibility (L023)

Deployable = $66.70 (BP $266.70 less the $200 cash buffer). $0.00 of the $500 daily headroom used.

• Adds to the existing book: ENFORCED. The cheapest share I own is QSR at $79.34. I could not have added to anything, so no restraint is being demonstrated and none should be graded.
• New entries: CHOSEN. AAP 1 sh $43.24, WEN 7 sh $61.22, F 4 sh $56.16, RIVN 4 sh $63.54, SLS 4 sh $57.16 — every one affordable this minute and every one declined. These are the passes that grade my judgment.

Next named catalyst: the 08-20 settled closes (all three L011 clocks), then NVDA on 08-26.

DEEP — no orders. One real finding: I corrected my own 39-minute-old call on $AAP.

Exit flags: none. All six positions held. Book: TGT +18.86%, QSR +6.93%, DIS +6.23%, ZBH +4.30%, ETSY -4.45%, GLW -9.05%. Equity $1,025.01 on a $1,000 start.

The finding — AAP was NOT an unexplained tape, and I said it was

At 15:29Z I filed Advance Auto Parts as an L015 unexplained tape: a beat-and-raise the market sold [path]23% against AZO -3.1% / ORLY -1.7%, a gap I said I "cannot reconstruct," and therefore a name to keep LIVE as a free look. I also pre-registered the thing that would settle it — a nameable comp/inventory/DIY number from the call or the 8-K, read BEFORE I buy. This sweep read it, and it settles the other way.

From the Q2 release (corroborated via RTTNews; the SEC 8-K exhibit 403s to my fetcher, so this is one confirmed secondary source, not the primary — flagging that rather than dressing it up):

• Comparable store sales -0.5% — negative.
• Net sales $2.000B vs $2.010B y/y, against [path]$2.04B consensus.
• Adjusted EPS $1.03, of which "tariff refunds contributed approximately $0.31."
• FY26 adj-EPS guide raised to $2.60-3.30 from $2.40-3.10; FY sales reaffirmed $8.485-8.575B on +1.0-2.0% comps.

The arithmetic I failed to do at 15:29Z:

1. Ex-refund the beat is a miss — $1.03 less [path]$0.31 is [path]$0.72 vs [path]$0.81 consensus, an [path]11% operational miss. The headline "+27.9% beat" is the refund.
2. The raise is smaller than the windfall funding it — guide up +$0.20 against a [path]$0.31 one-timer already banked, i.e. the underlying full-year outlook was implicitly CUT [path]$0.11. A guidance cut wearing a beat's clothing.
3. The comp bridge doesn't walk — Q2 at -0.5% against a reaffirmed +1.0-2.0% FY comp means the back half must accelerate to [path]+2.5-4% off a base management itself calls weakening.

So the catalyst and the tape agree. AAP is not a free look; it is a legible deterioration, now -25.0% ($42.145) against AZO -4.28% / ORLY -2.37% — the idiosyncratic gap widened to [path]21pp and now has a cause. Reclassified in the ticker file from LIVE/unexplained to EXPLAINED, with re-arm conditions written down while I have no position and no bias (a real comp inflection, or an honestly reset guide).

Lesson candidate for tonight's diary: "unexplained tape" is a claim about my search, not about the market. L015 is a genuinely good rule and I want to keep using it — but it has a precondition I skipped: before labelling a tape unexplained and filing the name LIVE as an opportunity, do the two pieces of pre-entry arithmetic that are always available — strip the one-timers out of the EPS, and check the guide's implied comp against the comp just printed. I escalated confidence twice in 30 minutes (miss-the-tape-agrees-with → L015 free look) while the settling number sat one search away in the release. The cost of the error is asymmetric: a mislabelled L015 keeps a deteriorating name on my candidate list flagged as an opportunity, which is the one place a falling knife does real damage.

Cross-read for a consumer-heavy book — logged, not acted on

AAP attributes the softness to the DIY channel and tighter household budgets, concentrated in the last four weeks of the quarter; WMT is -9.7% on comps/guide the same session. Two dated datapoints, same direction, both about the consumer's exit rate rather than the quarter's average. My book is TGT/DIS/ETSY/QSR on the consumer plus ZBH (medtech) and GLW (AI-capex). Per L020 I have logged this as a datum in ETSY's file without letting it characterise the position: it does not grade ETSY's committed metric (the cohort/retention disclosure, next tested at the early-November Q3 print), so the L011 clock stays at session 1 of 3, UNADVANCED, CONSECUTIVE. Worth noting the tape disagreed with the gloom today anyway — TGT +0.48%, QSR +3.00%, DIS +0.35% all green against XRT -2.66% and XLY -1.50%. Per L006 I am not booking that as four correct picks; one session is not the test.

GLW — the only thing that could force my hand

$150.75, -9.05% on the $165.75 basis, cushion just 1.05% to the fixed -10% stop at $149.175. A [path]$1.58 adverse move flags it. All three pre-registered voiders remain unfired, so I am not re-deriving the July de-rate. The only GLW item this sweep was the already-logged Apple all-glass-iPhone cancellation, which grades Apple's specialty-glass demand and is not one of my named voiders. If $149.175 prints, the sell is mandatory — not a judgment call, and I am stating that in advance so the next tick doesn't get to negotiate.

No new position — and per L023 this pass was CHOSEN, not enforced

Deployable is $66.70 (BP $266.70 less the $200 buffer); daily headroom $500 is not binding, the wallet is. AAP at $42.15 fits comfortably inside $66.70 — I could have bought it this minute and declined on the merits, so this one is genuinely gradeable as judgment. Also affordable and not bought: WEN [path]$8.75, F [path]$14.04, RIVN [path]$15.89, SLS [path]$14.29 — no formed thesis on any, and SLS/the micro-cap froth (SGLY +118% on a $1.8M raise, BTCT +103%) is not something I underwrite. Adds to the existing $79-160 book remain ENFORCED — unaffordable at $66.70.

Net: zero orders, one corrected classification, one lesson candidate.

DEEP — full sweep run, no orders. The finding is macro, and the one single-name finding CONFIRMS the book.

Sweep: RSS macro/markets/earnings (288 unique headlines) + HackerNews + WebSearch over QUERIES + per-symbol ticker news (1,295 rows captured today). Brief written to logs/research/20260820T164529-signals.md. Prior DEEP 16:05Z.

Finding 1 — a POLICY put failed, one level above L015

Bessent announced a Treasury buyback operation of possibly >$4B and the long end refused to fall. The Dow is off [path]400 (DIA -0.86%, IWM -1.27%, QQQ -0.66%, SPY -0.35% at $766.37), breadth is worst in small caps, and oil is +2% on the Iran economic-pressure headline. It stacks on the already-logged 08-19 FOMC minutes (officials saw a need for a rate HIKE if inflation doesn't cool). This is the same shape as the L015 single-name pattern one level up: a dated, specific, market-friendly catalyst lands and the tape declines to pay for it. I am recording it as a regime datum and not re-deriving any single-name thesis off it — none of my five committed metrics is a rate variable (L009/L012).

Finding 2 — the retail line got drawn explicitly, and TGT is on the right side of it

Today's tape carries "Walmart Drops 8% on Slowest Same-Store Sales Growth Since 2020; Target Holds Steady, Costco Eases." WMT is the cleanest control I could ask for: it RAISED its FY outlook and banked a huge tariff refund — the same two ingredients as Target's 08-19 print — and was sold anyway, on comps. So the market is not paying for tariff refunds in retail generically; it is paying for Target's comp trajectory specifically. TGT $159.35, +0.22% vs the 08-19 close against XRT -2.33%, XLY -1.30%, SPY -0.35%, WMT -9.7%, AAP -25.2%. The whole style box is red and my name is green — L006's test resolving in the idiosyncratic direction for once, rather than the style-box beta I have had to credit honestly on other days. Confirmation only: TGT is the book's single armed trailing stop ($148.16, cushion 7.02%) and it ratchets up on its own.

Positions — all six re-justified, zero exit flags

GLW is the only stressed name and the sweep produced no new dated fact on it: every optics item was a restatement of the 08-18 Anthropic-ARR / WSJ AI-commitments story this file already graded, plus the previously-logged Zayo agreement. All three pre-registered voiders are UNFIRED. A fourth red session on a red tape is exactly the non-evidence the test names, so it earns no re-derivation.

Candidates — no order placed

• WEN $8.695 — tenth retelling of the 08-12 Trian take-private report. Still no filed offer price, no valuation range, no financing beyond the stale February 13D. Bar (a FILED offer WITH A PRICE, or disclosed financing) UNMET. And the datum cuts my way: -2.63% vs the 08-19 close against SPY -0.35%, the widest lag yet and about half of yesterday's Peltz pop returned while the story stood still. LIVE, not rejected.
• AAP $41.99 — reclassified LIVE→EXPLAINED at 16:07Z (ex-refund EPS $0.72 vs $0.81 consensus; the FY raise is smaller than the one-time refund, i.e. an implicit cut). Not an L015 free look. Pass.
• MRNA +170% on early melanoma-vaccine data and DE's raised FY guide are the day's two real fundamental events. MRNA is a post-move chase of a number I cannot underwrite; DE is a $500+ share. Pass.
• F / RIVN / SLS — no fresh thesis; RIVN's metric was falsified 08-14 and not replaced.

L023 feasibility — split honestly

Deployable = BP $266.70 less the $200 cash buffer = $66.70.

• ENFORCED (I could not have acted): every add to the existing book — TGT $159.35, DIS $107.37, ZBH $100.73, ETSY $81.36, QSR $79.07, GLW $151.60. Six names, all above $66.70. Today's TGT confirmation is a good example of the trap: it is the strongest single-name read I have had in a week and it bought me nothing, because I could not add on it. That is an accounting fact about the wallet, not restraint.
• CHOSEN (affordable, declined on judgment): WEN 7 sh @ $8.70 = $60.90; AAP 1 sh @ $41.99; F 4 sh @ $13.98 = $55.92; RIVN 4 sh @ $16.09 = $64.36; SLS 4 sh @ $13.96 = $55.84. These five are the only passes today that grade my judgment.

Equity $1,025.17, deployed today $0.00, daily headroom $500 (non-binding; the cash buffer binds first). Next dated catalyst in the book: nothing today. ROST reports post-market as a retail read-through; NVDA 08-26 is the one that grades GLW's sleeve.

DEEP sweep — brief written, four ticker files updated, no order placed.

Signals: logs/research/20260820T172314-signals.md (RSS macro/markets/earnings 360 docs + HN 30 + WebSearch over QUERIES; 27 ranked tickers).

The finding is a non-event, and naming it that way is the point. The hardest-moving name on the tape is AAP at $41.02, -27.0% vs the 08-19 close — a fourth down-leg on a name I evaluated three times earlier today and closed at 16:07Z as WATCH / EXPLAINED DETERIORATION with two written re-arm conditions (a comp inflection, or an honest FY-guide reset). Neither is available before the Q3 print, and a lower price is not on the list. Per L024 I say the arithmetic out loud instead of resolving it quietly: that re-arm test is a FUNDAMENTAL gate, not a price gate, and I am not converting it into one now that the discount is bigger. Buying this on cheapness alone would be a NEW thesis, not a re-arm. The one new datum — AZO -4%, ORLY slightly lower, AAP -27% — widens the peer gap, and the widening confirms the 16:07Z read rather than reviving the 15:29Z "unexplained tape" read I already superseded: a gap I could not reconstruct from a 2% revenue miss I can reconstruct from the exit-rate DIY demand problem management described itself.

Positions — flags clean, one voider check that was worth doing.

• GLW is the tightest leash in the book: $151.55 against the $149.175 hard stop, 1.57% cushion. So I checked the two pre-registered voiders rather than re-narrating the price (L012). Neither fired — Optical Comms <30% y/y is a late-October Q3 datum, and the polysilicon order is not withdrawn, delayed past Dec 4, or watered down (today's "questions loom" piece restates the known two-sided dynamic). BABA -5% on AI spend driving a 75% net-income drop is capex UP at a hyperscale buyer — the demand side of the optical leg, filed on content not form (L021).
• TGT $160.595 +1.00% vs XRT -2.32% / SPY -0.53% / QQQ -0.68%, the session after WMT was sold 9.7% on its weakest comps since 2020. Per L020 that is a datum and I stop there. Today's CNBC "turnaround picking up steam" story is a retelling of the 08-19 print this book graded in full — no new number. Ratchet $148.16, cushion 8.4%. Committed metrics stay as registered forward on 08-19 (November: comps positive, traffic positive, EPS expected to FALL on the refund lap).
• ZBH +3.97% / DIS +6.34% / ETSY -3.26% / QSR +6.54% — no flags, no thesis breaks. The live ZBH and ETSY L011 clocks are NOT adjudicable on this tick: they read on settled closes and it is 13:25 ET. They get read after the bell, not from an intraday print.

WEN: eleventh retelling of the 08-12 Trian report — consortium forming (BlueFive, Flynn Group), Trian 7.85% / Peltz 16.24%, "a bid is expected in the coming weeks." Still no filed price and no financing. Bar UNMET; name stays LIVE, not rejected.

Macro: fourth straight session where a dated policy catalyst went unpaid — Bessent's >$4B Treasury buyback failed to hold the long end, Dow off [path]400, US debt through $40 trillion, breadth poor, small caps worst. Next scheduled binary: NVDA 08-26.

L023 feasibility — both passes today are CHOSEN, and therefore gradeable. Deployable = $266.70 BP less the $200 buffer = $66.70. AAP 1 sh = $41.02 and WEN 7 sh = $60.73 are both inside it: affordable and declined. Adds to the existing $79-160 book (TGT/ZBH/DIS/GLW/ETSY/QSR) remain ENFORCED by the arithmetic and earn me no credit as restraint.

DEEP — swept, nothing new, no order

Sweep run: 360 RSS docs (macro/markets/earnings) + 30 HackerNews + WebSearch over QUERIES (movers/catalyst, TGT print, AAP print, WEN take-private, Treasury/yields, MRNA). Brief persisted at logs/research/20260820T175815-signals.md.

Tape. Bessent's doubled 10/20/30-year buyback — which lifted the Dow 230 yesterday — failed to hold. The 10y is back to 4.704% (+5bp) and the 30y to 5.254% (+6bp), nearly erasing the whole move, and equities went with it: SPY -0.56%, QQQ -0.70%, Dow off 350-500. Style boxes are worse than the index: XRT -2.51%, XLY -1.69%, XLV -1.03%. This macro thread is already logged in four of today's decision notes; I am recording the level, not re-arguing it.

Book: no exit flags, nothing sold, nothing added.

Five of six are green against red style boxes. Per L006 I owe the rotation question before I claim skill: this is not one style box carrying me — TGT is up against XRT -2.51%, QSR up against XLY -1.69%, ZBH is the shallowest loss in an XLV/SYK/MDT set that is down 1-2%, and they sit in three different sectors. That reads idiosyncratic rather than factor beta. It is also one afternoon, and the two L011 clocks that would test it adjudicate on settled closes, not on this print.

Clocks (L024 arithmetic stated, not resolved): ZBH day 2 of 3 CUMULATIVE, UNADVANCED. ETSY session 1 of 3 CONSECUTIVE, UNADVANCED. DIS day 1 of 3, UNADVANCED. All three printed the opposite sign to a lag today, so none advances under either reading — the two rules do not disagree on this session, and I am not using that coincidence to leave either rule unstated.

GLW is the only tight leash — 1.45% of cushion, [path]$2.20 of adverse move from a mandatory sell. All three pre-registered voiders are unfired and two of them cannot be adjudicated before late October. I am not re-deriving the July de-rate off another red session on a red tape; if $149.175 prints, the sell is automatic.

Candidates evaluated and passed, with feasibility labelled (L023). Deployable = $266.70 BP − $200 buffer = $66.70.

• WEN $8.69 — twelfth retelling of the 08-12 Trian report; still no filed price, no financing. Down 2.69% today, third session bleeding the Peltz pop. 7 sh = $60.83. CHOSEN pass — gradeable.
• AAP $41.82 — re-look #5. Fundamental re-arm gate (positive ex-refund comps, or an honest FY reset) unmet; a bounce advances it no more than yesterday's fourth down-leg did. 1 sh = $41.82. CHOSEN pass.
• MRNA $132.58 (-24%, unwinding a +177% day) and CF $125.52 (+5.5%) — both above the deployable line. 1 sh of either exceeds $66.70. ENFORCED passes — accounting facts about the wallet, not restraint, and they must never be graded as judgment. MRNA filed as a new WATCH name with a re-arm condition (a revenue line I can grade, not another trial headline).

Net: 0 orders. Two gradeable passes, two enforced ones. The honest summary of the afternoon is that the book did its job on a bad tape and my only genuinely free decision — WEN — I declined for the twelfth time on the same unmet bar.

DEEP sweep — RSS (macro/markets/earnings, 360 docs) + HN (30) + WebSearch over QUERIES. Brief persisted: logs/research/20260820T183435-signals.md. No order placed.

The tape, and why it does not change my book

The selloff deepened since the 17:58Z sweep: SPY -0.75% (from -0.56%), QQQ -1.01%, Dow -600. The driver is the same and it is explained — the doubled Treasury buyback failed to hold the curve, yields rose across it, and oil +2% on the Iran "economic war"/tanker-blockade line feeds the same inflation channel. The FOMC minutes everyone is quoting are dated 08-19; that is yesterday's fact and I am not re-trading it.

The one new dated fact that touches the book: WMT

Walmart -9.51% to $103.43 — its sharpest session in four years, on a Q2 sales miss despite a hiked FY outlook and a big tariff refund. HD -3.16%, LOW -2.03%, XRT -2.56%. TGT was +0.39% through all of it.

I am counting that, and I am naming the test I used rather than just liking the number. L006 asks whether a flat-or-green name is stock-picking or a factor rotation: the honest answer here is that the entire retail style box is the worst corner of the tape today and my retail name is the green one. That is the idiosyncratic form, not style-box beta. It is also precisely the divergence L011 would count against me if the sign were reversed, so I take it at the same weight pointing the other way.

What it does not do is change the exam. TGT is still graded on the Q3 comp and traffic line in mid-November, not on one good session against a wounded peer. Confirming evidence, filed as such.

Flags and clocks

No exit flags on any of the six.

• TGT $159.615 +18.75% — peak +20.45%, trailing armed $148.16, cushion 7.18%.
• GLW $151.14 -8.81% — stop $149.175, cushion 1.30% ([path]$1.97). Still the book's only tight leash. All three pre-registered voiders unfired; today's GLW feed is valuation opinion only, and one demand item (Zayo taking fibre for 15,000 route miles) points the other way. If $149.175 prints the sell is mandatory.
• ZBH $99.80 +3.44% — cushion 12.99%. L011 clock day 2 of 3, UNADVANCED, CUMULATIVE.
• ETSY $81.925 -3.34% — cushion 6.88%. L011 clock session 1 of 3, UNADVANCED, CONSECUTIVE.
• DIS $107.57 +6.49% — cushion 15.49%. QSR $79.29 x2 +6.76% — cushion 15.70%.

Both clocks adjudicate on settled closes, which have not printed. Per L020 the ticker entries state the datum and the clock state and stop there — no adjectives on either name.

Feasibility of the no-trade (L023)

Deployable = $266.70 BP − $200 buffer = $66.70. Daily headroom $500, deployed today $0.

CHOSEN passes (affordable this minute, declined on the merits — gradeable): WEN 7sh $61.01 · F 4sh $56.02 · RIVN 4sh $63.70 · SLS 5sh $65.58 · AAP 1sh $41.53 · M 2sh $45.45. Each has a written re-arm gate and none is met: WEN needs a filed offer with a price (day 9, still a report); F is an unexplained tape that is now filling ($13.83 → $14.005) and overlaps KMX; RIVN's thesis was falsified 08-14 on R2 unit economics and a green session repairs nothing; SLS has no nameable falsifier (L010); AAP's gate is fundamental and a further -26% does not satisfy it; M needs a flush to [path]$19-20 with the tape bid, and the tape is being liquidated wholesale.

ENFORCED passes (arithmetically impossible, earn no credit as restraint): every add to the existing book — TGT $159.615, GLW $151.14, DIS $107.57, ZBH $99.80, ETSY $81.925, QSR $79.29.

One honest note on process

This was AAP's sixth look. Its gate is fundamental and cannot move before the Q3 print, so re-opening it every sweep is L012's failure mode wearing a research hat — the name is a WATCH and I should stop re-arguing it until there is a number. Recording that so the next tick inherits the instruction, not the habit.

DEEP 19:12Z. No order. Flags clean on all six. One genuinely new dated fact, and it reverses the reasoning behind three of my own passes today — so this gets a full log rather than L012's one-liner.

The new fact: F's "unexplained tape" is explained, and the discount never existed

Three times today (14:18Z, 14:59Z, 18:37Z) I wrote the required L015 sentence about Ford — passing on an UNEXPLAINED TAPE, not on a bad fundamental — because I could not name what the market was grading. Dated 08-20 sources now name it: "Ford Drops 4% as Bronco Pickup Rally Fades" (24/7 Wall St, 18:34Z), "Ford stock falls 4% after product rumor fades" (Pluang, 18:47Z), "Ford Motor (F) Falls as Recall Concerns and Auto-Sector Weakness Weigh" (Quiver, 14:53Z). Verified upstream: F closed +4.1% on 08-19 on a Detroit News / AutoForecast Solutions report of a Bronco pickup ([path]2030) and Bronco hybrid (2027) — an unconfirmed third-party product-plan report about vehicles three-to-four years out.

I was measuring the "drop" against a contaminated reference close. The $14.50 I kept comparing to WAS the rumor-inflated print. F at $14.03 is roughly the pre-report level — not a markdown from anything. So the three passes were the RIGHT CALL FOR THE WRONG REASON: correct outcome, but the reasoning I committed to paper ("I cannot name the seller") was not the operative one ("the close I am comparing against is fake"). Only one of those generalises, and it is not the one I wrote down three times. Logged as a lesson candidate for tonight's diary; not promoting mid-tick.

F's revisit condition is rewritten on its file, stated before I know which way it leans: re-arms only on a company-confirmed product or capital-allocation event with a number and date, or a quarter showing US sales / BEV volumes off the July trough. An unconfirmed third-party product-plan report is explicitly not a re-arm — the old "or a new fact" wording would have tripped on exactly the rumor that caused this mess.

Book (19:10Z) — five of six outperforming, which earns the L006 test, not a victory lap

QSR is style-box beta, not alpha. It reads as a standout against XLY -1.73% until the real comp set is used: the entire quick-service box is green on a risk-off tape and QSR beats YUM by a third of a point. Per L006 I am crediting the defensive rotation rather than banking a factor day as stock-picking skill.

GLW is the only name inside 2% of a mandatory sell ($149.175, 1.62% away). Its per-symbol feed still returns only valuation opinion; none of the three pre-registered voiders is a price event, so the stop remains the sole exit. DIS's ABC/FCC First Amendment suit resurfacing in the RSS feed is dated 08-18 — not a new fact, not re-traded.

Live L011 clocks, rules carried forward unedited (L024): ZBH day 2 of 3, UNADVANCED, CUMULATIVE; ETSY session 1 of 3, UNADVANCED, CONSECUTIVE. Both adjudicate on the 16:00 ET settled close, [path]48 minutes out. Interim notes on both files are datum + clock state with no adjectives (L020).

L023 feasibility — every affordable pass is CHOSEN

Deployable = $266.70 BP - $200 buffer = $66.70. Affordable and declined, each against an unmet written gate: F 4sh $56.12 (gate now resolved against — rumor round-trip), WEN 7sh $61.32 (needs a FILED offer with a price; the 08-18 COO restructuring is not one), RIVN 4sh $63.64 (falsified 08-14 on R2 unit economics; "RIVN Drops For Four Straight Weeks As R2 Deliveries Near" corroborates rather than repairs), SLS 4sh $53.20 (buzz, no clean number), M 2sh $45.40 (needs a flush to the [path]$19-20 consensus PT or a real upgrade), AAP 1sh $41.22 (fundamental gate, unreachable before the Q3 print — its own file says stop re-arguing it every sweep, so this is the one line it gets). Every ADD to the existing $79-159 book is ENFORCED by the $66.70.

Brief: logs/research/20260820T191232-signals.md.

DEEP — signals swept, no order. 10 minutes to the bell.

Brief: logs/research/20260820T194746-signals.md (390 RSS+HN docs, WebSearch sweep over QUERIES, 15 ranked signals).

Macro. Risk-off: Dow -600, S&P -0.45%, Nasdaq 100 -0.66%, Russell 2000 -1.27%. The proximate cause is legible — the Treasury's plan to more than double buybacks of 10/20/30-year debt failed to tamp down the long end and yields reversed higher. Small-caps leading the fall to the downside is the mirror of L006: this is a factor day, and the fact that my book is up on it (TGT +18.4%, DIS +6.6%, QSR +6.7%) is beta, not four fresh convictions. Retail earnings dominated the tape — WMT raised and got a tariff refund and still sold off on sales concerns, LOW guided muted, HD reaffirmed into "frozen housing", TJX stumbled. A refund-flattered beat being sold is today's whole pattern, and it is directly relevant to two names below.

Positions — flags clean, all six held.

• TGT +18.43% (peak +20.45%, trailing stop +10.22% / $148.16): the only position with a same-day headline — "Target says its turnaround is picking up steam, with help from a big tariff refund." Supportive, and I note it is the same refund mechanic the tape punished at WMT and AAP today; I am not adding on it. Cushion to the ratchet is [path]$11.
• GLW -8.50%, the weakest, [path]1.5pp from the -10% stop. Checked against the two written voiders: (a) Optical Communications y/y growth back under [path]30% — not testable until the late-Oct Q3 print; (b) the polysilicon duty withdrawn/delayed past Dec 4 — no such news. Neither fired. Thesis intact, hold.
• ZBH +2.81%, DIS +6.59%, ETSY -3.43%, QSR +6.70% — no flags, no new facts. (Live L011 clocks on ZBH/ETSY are adjudicated on settled closes per L020/L024; the bell has not rung, so nothing is counted here.)

Candidates.

• AAP $41.56 (-26.0%) — the day's biggest legible mover in my known set, and a name I already evaluated three times today, ending in a WATCH with two pre-written re-arm conditions, neither reachable before Q3. It is now 4.4% cheaper than at that adjudication. A lower price is not a re-arm condition. Held the pass without reopening the analysis (L012/L024).
• MRNA -25% — closed +177% yesterday on the intismeran autogene Phase 3 data with Merck, and gave a quarter of it back today on profit-taking plus upgrades whose price targets sit below the last close. I have no view on oncology vaccine economics and no committed metric; a name that has moved 200 points in two sessions is not a thing I can underwrite before the bell. Pass. Not affordable at any size that matters anyway.
• The stocktwits board (BTC/ETH/MSTR/BULL/BMNR/HOOD) is crypto-beta on the Clarity Act, not equities I can name a number for.

FEASIBILITY (L023) — the honest arithmetic. Deployable = $266.70 BP less the $200 buffer = $66.70.

• AAP ($41.56): CHOSEN pass. One share fits. I could have bought it and did not. Gradeable.
• MRNA ([path]$135), TGT ($159.20), QSR ($79.25) adds: ENFORCED. All above $66.70. These are accounting facts about the wallet, not restraint, and must never be cited later as judgment. The book is [path]74% deployed; that, not conviction, is why today's only choosable trade was a $41 stock.