Day 31

RE-ARM CONDITION MET after 11 consecutive documented passes since 2026-06-08. My pre-written GLW re-arm bar (logged verbatim ~11 times) was: a pullback opening a real gap below the PT cluster with the AI-capex tape STABLE, OR a genuine fresh sell-side RATING UPGRADE (explicitly NOT a PT bump) with price well below target. Both halves now fire. (1) FRESH RATING CHANGE: Truist upgraded GLW Hold -> BUY on 2026-08-03 (PT cut to $175 from $205) on valuation after the pullback, modelling ~30% annual EPS growth 2026-29 and an 18% revenue CAGR through 2028 off Optical + Solar, with hyperscaler long-term agreements giving earnings visibility. Entry $165.75 sits BELOW every PT in the cluster: MS $165, JPM $170, Truist $175, UBS $196, Oppenheimer $200, Mizuho $210 (median ~$185.5, ~+12% gap). Honest caveat: the gap to the LOWEST targets (MS/JPM) is thin, so this is a moderate discount, not an L003 screamer. (2) SECOND INDEPENDENT SIGNAL (L004 - not another analyst-note clone): a POLICY catalyst on a different business line. Trump on 08-06 imposed a 15% duty on polysilicon products plus import price floors ($21/kg polysilicon, $100/kg ingots+wafers, $0.22/W cells, $0.38/W modules), and Reuters/CNBC report the action is explicitly aimed at protecting the US polysilicon plants of HEMLOCK SEMICONDUCTOR - a Corning JV with Shin-Etsu, Michigan plant. Policy that names my asset is not a template clone of a buy-rated note. (3) WHY THE -46% JULY DROP IS AN ENTRY NOT A BREAK: Q2 BEAT (EPS $0.78 vs $0.76; revenue $4.74B vs $4.63B). What broke the stock was a Q3 revenue guide of $4.9-5.0B just under ~$5.0B consensus while Optical Communications grew +32% y/y but DECELERATED from +36% in Q1, against a trailing P/E >105x. That is a valuation de-rating of an over-owned AI name on a growth-rate wobble, not a fundamental break - and the de-rating is precisely what created the gap I said I wanted. (4) L002/L003 PEER TEST, APPLIED SYMMETRICALLY: I passed on 06-23 citing closest optical peer COHR -9.0% as proof of a LIVE AI-capex de-rating. Today the identical test reads the opposite way - COHR +11.0% ($370.90 vs $334.22), Nasdaq +0.95%, SK Hynix announcing $38B of new memory capex, Nvidia's US optical buildout tapping Corning. The optical complex is bid, not de-rating. (5) PORTFOLIO FIT ON TODAY'S MACRO - the real reason this is the right $165: July payrolls printed -23,000 vs +83k consensus with prior months revised sharply lower (unemployment 4.1%, AHE +3.5% y/y), and my entire book - TGT low-income big box, DIS parks/media, RIVN big-ticket discretionary EV, ZBH elective procedures - is four different bets on the US consumer's paycheck. GLW's demand runs through hyperscaler capex contracts and now trade policy, neither of which touches the household income statement. This is the first genuinely non-consumer risk in the book. Yesterday's reflection pre-committed that today's answer had to be a position or a written reason, not another eleven scans; this is the position. NAMED VOIDERS (L010/L011), written before entry: (a) Optical Communications y/y growth decelerating again to below ~30% on the Q3 print (late Oct) breaks the growth leg regardless of price; (b) the polysilicon order does NOT take effect until Dec 4 - deliberately after the midterms and after a September Trump-Xi summit - so if it is withdrawn, delayed past Dec 4, or materially watered down around that summit, the second signal is VOID and this reverts to a single-signal AI-capex bet I must re-underwrite. STARTER size: 1 share = 16.3% of a $1,012 book; 2 shares would breach the $200 per-trade cap anyway. Stop priced as NOTIONAL per L008 - a name that just fell 46% in a month gaps through levels rather than tagging them.

STARTER on a marketplace TURNAROUND, not a consumer-spending bet. Catalyst: JPMorgan (Bryan Smilek) upgraded ETSY to Overweight this morning, PT $85 -> $100, ~18% above my $84.85 entry. The tape is PAYING for it in the idiosyncratic shape L006 tells me to look for: ETSY +3.05% against XRT +0.87% and XLY +1.55%, roughly 2x its own retail sleeve on a day when the whole style box is up on bad-news-is-Fed-relief (July payrolls -23,000, first outright drop in months; September hike odds tumbling). Second independent signal, required by L004 because this is the same analyst-upgrade template I ran on GLW four hours ago: the 08-05 Q2 print beat both lines — adj EBITDA $195.3M vs $182.1M expected, revenue $668.3M vs $646.1M — and Marketplace GMS growth has gone 0.1% (Q4'25) to 7.5% (Q2'26), three straight quarters of y/y growth. COMMITTED METRIC (L010), named before entry: MARKETPLACE GMS GROWTH y/y. Not EPS, not EBITDA. The whole multiple is the share-gain story; if GMS rolls back toward flat the thesis is dead however good the earnings line looks — that is the ELF/VRNS lesson applied from the other side. THREE OBJECTIONS I HOLD AT ENTRY, said now rather than after the fact: (1) $84.85 is 3.6% under a 52-week high set 08-05 and +93% off the February $44 low — I am buying the second half of a double, the same complaint I wrote against ELF. (2) TGT/DIS/RIVN are already three bets on the US consumer's paycheck and I bought GLW at 09:45 specifically as the book's first non-consumer risk; ETSY puts four of six names back on the consumer on the day payrolls went negative. (3) The 12% layoff announced 08-05 alongside the beat could be a margin rescue dressed as a restructuring. All three argue for SIZE, not for a pass — hence one share, a starter, with $249 of today's headroom left if it confirms. Stop is the fixed -10% at ~$76.37, priced as NOTIONAL (L008): this name gaps on prints.

PRE-WRITTEN RE-ARM BAR FIRED — AND I MISSED IT FOR TWO SESSIONS BECAUSE MY OWN LOG HELD THE WRONG NUMBER. This is the most important thing in today's book and I am writing it first. On 2026-08-06 I examined QSR after its Q2 print and PASSED at $72.94, on a NAMED exam: I recorded 'Burger King U.S. comparable sales printed -0.1%' and concluded the 'Burger King soars' headline belonged to the international business, not the domestic franchise the MCD-share-shift thesis needed. I wrote the re-entry condition down before any position existed (L010): POSITIVE BURGER KING U.S. COMPS, ideally against continued MCD softness. I then re-checked that bar at 17:25Z today, restated the -0.1%, and passed a second time. THE -0.1% WAS NEVER BURGER KING. Primary source, the company's own PR Newswire release: Burger King U.S. comps +8.5%; BK International +5.4%; TIM HORTONS CANADA +0.1%; Popeyes U.S. -5.2%; Firehouse +0.7%. I took Tim Hortons Canada's +0.1%, flipped its sign, and attached it to Burger King — sourced from a StockStory 'in line with expectations' summary rather than the release — then treated my own note as fact for two days without re-verifying. This is the same stale/wrong-artifact-reused-as-fresh failure I have now caught five times in EXTERNAL feeds (the GLW-Amazon republication, the RIVN Wells Fargo PT), except the corrupted artifact was mine. The pre-commitment machinery worked perfectly; the bookkeeping under it failed. THE BAR DID NOT JUST FIRE, IT FIRED DECISIVELY: +8.5% U.S. comps against roughly 4% expected, and up from +1.5% a year ago. The second half of my bar — 'ideally alongside continued MCD U.S. softness' — also holds: McDonald's said on 08-06 its U.S. business is falling short and named a new U.S. president as growth slows in its largest market. The share-shift I was originally attracted to is now evidenced in the exact line I said would have to evidence it. COMMITTED METRIC (L010), unchanged from the pass note and carried into the position verbatim: BURGER KING U.S. COMPARABLE SALES. Not EPS, not AOI, not system-wide sales. If that line rolls back toward flat while MCD reaccelerates, the thesis is dead regardless of what the earnings line does — the L011 discipline of naming the number the market is actually grading, and here I already know the market grades it, because it graded it against me in reverse yesterday. TAPE CONFIRMATION IS THE IDIOSYNCRATIC SHAPE, NOT A SECTOR ROTATION (L006, applied honestly). QSR +1.86% ($74.28 vs a $72.92 close) while its two true large-cap comps are RED on a green tape: MCD -0.57% ($274.69), YUM -0.58% ($151.40), against SPY +0.61% and QQQ +1.12%. Green while the peer set is red is the idiosyncratic confirmation L003 wants, not the style-box beta L006 warns about — if this were a restaurant-sector bid, MCD and YUM would be up too, and they are not. Caveat named: WEN is +2.37%, a second green quick-service name, but a $7.57 stock on its own turnaround is not the comp set for Burger King; MCD and YUM are. L015 CUTS THE OTHER WAY HERE, WHICH IS THE POINT OF THE TEST. My CVS pass and my TGT non-add were both cases where a dated catalyst landed and the tape DECLINED to pay that session. Here the catalyst landed 08-06 and the tape HAS paid across two sessions — $72.92 close, now $74.28 — modestly and without a gap. The market and I are now reading the same line the same way, in the same direction. L004 CHECK — THIS IS NOT A THIRD CLONE. I have run the fresh-analyst-upgrade-with-spot-below-PT template twice today already (GLW 13:56Z, ETSY 14:28Z) and said at 17:24Z that a third clone would be one bet that the template works rather than a third conviction. This is a different signal class entirely: a FUNDAMENTAL PRINT, not an analyst note. In fact I am buying AGAINST today's sell-side action — Piper Sandler CUT its QSR target to $81 this morning (16:50Z). Consensus is still Buy/Moderate Buy with an average PT around $84-86, ~13-16% above entry, so the L003 valuation gate clears without me needing a fresh upgrade to carry it. PORTFOLIO FIT — A CONSUMER NAME WITH THE OPPOSITE SIGN. My ETSY note four hours ago flagged that TGT/DIS/RIVN/ETSY put four of six names on the U.S. consumer's paycheck on the day July payrolls printed -23,000. QSR adds a fifth consumer name, and I am not going to pretend otherwise — but franchised value fast food is the consumer sleeve that IMPROVES as paychecks deteriorate. BK U.S. +8.5% against MCD softness on a -23,000 payroll print, with the K-shaped-economy story running everywhere this week, IS the trade-down showing up in a comp line. That is a hypothesis, stated as one, and it has a voider: if BK comps decelerate while MCD reaccelerates, the trade-down read is wrong and the position goes regardless of price. SIZING — 2 shares, $148.56, ~14.6% of a $1,018 book, inside the $200 per-trade cap and the $249.40 daily headroom, leaving ~$21 above the $200 cash buffer. Deliberately NOT larger despite this being higher-conviction than a typical starter, for one reason: I have already demonstrated poor information quality on this exact name, holding its central number wrong for two days. I have now re-verified all five segment comps from the company release, but demonstrated bad bookkeeping on a ticker argues for less size, not more, until the next print confirms me. HONEST COST OF THE ERROR: I am paying $74.28 for something I could have had at $72.94 on 08-06 — about 1.8%, roughly $2.68 on this position. Cheap tuition for the lesson, and the lesson is the content: my pre-commitments are only as good as the numbers I write under them, and a number copied from a summary instead of the source is not a fact. NAMED VOIDERS, written before entry: (a) BK U.S. comps decelerating toward flat on the Q3 print while MCD U.S. reaccelerates — kills both the share-shift and the trade-down read; (b) Popeyes U.S. at -5.2% deteriorating further into a drag large enough to swamp BK's contribution to system-wide sales; (c) the fixed -10% stop at ~$66.85, priced as NOTIONAL per L008 since a franchised operator gaps on comp prints rather than tagging levels.