When: RBC pre-templated gross-margin-with-tariff-refund-offset narrative plays out

Then: Maintain 90% core; watch sell-the-news risk

When: Gross margin compression >50bps, soft FY27 guide, renewal <90%

Then: Trim additional 10-15% to defensive tail; reassess moat integrity

TRIM 10-15% AT 895, STOP AT 870

• Trim 10-15% of COST exposure at current levels near 895, not waiting for 50 SMA at 936 which has rolled over from ~974 to 936
• Set hard stop on remaining position at 870, below lower Bollinger band at 876
• Size remaining position so a 15% post-earnings drawdown is survivable
• Do not redeploy until macro regime shifts: ten-year backs off 5% or Fed signals dovish pivot
• This is risk management and capital preservation, not an alpha-generating short-side bet
• Reject waiting for 50 SMA at 936 because it is falling, not rising, and price may never reach that level
• MACD histogram is widening negative, moving averages stacked bearishly, volume at 4.16 million elevated, ATR compression suggests imminent volatility expansion
• Reject stop at 958 (aggressive analyst) because ATR is 14.26 and 2-day adverse move is ~28 points; 958 gives back nearly 7% before exit
• Quality businesses can compress 30-40% on multiple re-rating (2007 precedent with Walmart and P&G)
• Costco could give back 25% in 12 months on multiple compression alone even with 12% earnings growth, producing -13% total return
• 45x trailing multiple in 5% yield environment with oil above $100 is mathematically challenged regardless of business quality
• Beta of 0.855 protects from market crashes but not from valuation compression in rising rate regime
• If price breaks below 870, that is a confirmed technical breakdown and further compression is likely
• If stock rips up on special dividend, smaller trimmed position still participates; if it gaps down, capital is preserved with dry powder to redeploy at better levels

→ vs aggressive: Framing trimming as an alpha-generating short-side bet is wrong; trimming is risk management and capital preservation, not return-seeking

→ vs aggressive: Entry at 50 SMA 936 is flawed because that level is falling (rolled from ~974 to 936), and price may never tag it before earnings, making the plan a mental stop that never executes

→ vs aggressive: Stop at 958 is too wide; with ATR of 14.26, a two-day adverse move is ~28 points, making the 958 stop give back nearly 7% of position before exit

→ vs neutral: Framework stacks too many contingencies: trim conditioned on clean beat, re-entry conditioned on ten-year backing off 5%, making probability of sequential execution vanishingly small

→ vs neutral: If print disappoints modestly, plan has no defined response; holding through a 10-15% draw waiting for confirmation that never comes is capital destruction through inaction

→ vs neutral: High bar for clean beat: must beat and raise at 45x trailing multiple in 5% yield regime with oil above $100 and no DoorDash contribution

Trim 10-15% of COST exposure now into 9.06-9.10 zone (10 EMA resistance) or at market; post-print triggers at 8.76 closing break or 8%+ gap down; redeploy on 8.44 capitulation volume or 10-year below 5%.

• Trim 10-15% exposure now into 9.06-9.10 zone (10 EMA resistance) or at market if price bleeds lower without testing this level
• Post-print: hold reduced position if COST gaps down 8%+ on September 24 print; cut remaining position if 8.76 lower Bollinger band breaks on closing basis within two weeks of print
• Redeploy triggers: 8.44 52-week low with capitulation volume OR 10-year Treasury yields break below 5%—either trigger, not both; do not wait for Fed dovish pivot
• Maintain reduced position sizing below benchmark weight regardless of print outcome
• Technical: downtrend with stacked bearish moving averages, MACD and volume confirm downside, ATR compression directionally ambiguous; elevated probability of breaking lower Bollinger band with non-trivial upside reversal chance on catalyst
• Fundamentals: net cash 11.86 billion, ROE 29%, FCF nearly 9 billion; business can compound through hostile macro if multiple re-rates
• Bear case analytically superior on 3-6 month horizon; trim captures both probabilities without overcommitting to either
• Special dividend not fully priced in at 100% probability—consensus awareness differs from complete price-in; magnitude and accompanying buyback authorization leave upside surprise room

→ vs aggressive: Trim trigger at 9.36 is 'a hope dressed up in technical jargon'—the 50 SMA has rolled over from 9.74 in late July to 9.36 today, actively walking away from the level; price may never tag it

→ vs aggressive: Hard stop at 9.58 is 63 points above current price, allowing 7% adverse move before exit—far too generous for a thesis with asymmetric downside into a binary event

→ vs aggressive: Special dividend dismissal as 'fully priced in' is 'lazy'—consensus expectation of above-average probability differs from 100% price-in; treating a binary catalyst as continuous variable is a category error

→ vs conservative: Stop at 8.70 'bets on direction' and is 'a short position disguised as risk management'—only triggers on continued downside, caps upside participation through implicit position sizing

→ vs conservative: 'Do not redeploy until macro regime shifts' is dogmatic and will cost real returns—2007 precedent shows Walmart and Procter bottomed 18 months before macro rates actually rolled over; waiting for academic macro signal will cause 6-12 month miss of the bottom

→ vs conservative: Stacking too many contingencies in sequence makes probability of all gates opening low—original framework's execution ambiguity is valid criticism