## Session overview Overnight, gainers were led by CVS up 4.80%, BKNG up 4.15%, QCOM up 3.97%, MSFT up 3.66% and LRCX up 2.62%. Decliners were led by VNET down 4.05%, INTC down 3.45%, META down 3.33%, CRM down 1.76% and ORCL down 1.75%. ## Compared with p

Verdict: Real edge, but the trigger has not fired — wait for the EMA reclaim

The idea buys oversold pullbacks in ten quality names, and the evidence base is real: the strongest tested window returned 97.5% on 12 daily trades with a 75% win rate and a 16.7% maximum drawdown, with positive results too on the 24-month and 12-month windows. The strongest point for the trade is that the buy list is fundamentally rich where it counts — CVS just posted Q2 free cash flow of $5.65B (up 66.3%) and Microsoft grew fiscal Q4 revenue 8.6% to $90.0B — so dips arrive in names with earnings support. The strongest point against is posture and precision: every one of the ten names shows net open-market insider selling in the filings for the period ending June 30, 2026 — CVS alone at about $171.6M — while the 2.4% stop sits inside a tape where INTC fell 3.45% and VNET fell 4.05% in a single session. No robust parameter setup was established by the sensitivity evaluation, so the published rules stand unrefined. The verdict: wait — RSI and trend-strength conditions are met, but every name still trades below its 50-period EMA, so the entry is not armed. A daily close above the EMA on CVS (about $95.2) with RSI still at or below 35 would flip us to actionable; a break under the $86.0 first support before that reclaim would kill the setup.

Trade now: no entry armed yet — the trend filter is the missing piece

**Bottom line: do nothing today — the setup is not armed.** The strategy wants three things on the daily chart for each name: RSI (14) at or below 35, price above the 50-day EMA, and ADX (14) above 20, plus a close above the first support level. Right now the oversold part is largely in place — CVS's daily RSI sits around 20.7, ORCL near 17, BKNG near 13.7 — and the trend-strength test is met nearly everywhere. What is missing is the momentum flip: every name still trades below its 50-period EMA, with CVS roughly $7.67 below its EMA near $95.2 and BKNG about $19.69 below its EMA near $181.5. Until price reclaims that line, the entry stays off. **What wait means concretely.** For CVS, a daily close above roughly $95.2 with RSI still at or below 35 (and above the $86.0 first support) would arm the trade. Because the strategy requires the close back above the EMA *while* still oversold, the trigger window is tight — typically a violent one- to two-day reversal. The closest analogous earnings and cash-flow profile supports the thesis: CVS just posted Q2 revenue of $106.1B, up 5.7% quarter over quarter, with free cash flow of $5.65B, up 66.3%. **Risk framing from the completed backtest.** The five-year daily backtest on this exact setup produced 12 trades at a 75% win rate, a 97.5% total return, and a maximum drawdown of 16.7%. Once live, each position carries a fixed 2.4% stop and a 4.8% take-profit, with exits also possible at the first resistance level — for CVS that is about $95.63, which would cap the reward near +4.8% against roughly 2.4% of risk, an effective reward-to-risk of about 2:1. Note the disclosed limitation once: exits were filled on daily bars rather than intraday ticks, so reported win rates are coarse. **Action plan.** Keep orders out of the market. Set alerts at the EMA reclaim levels for each ticker, size any eventual entry at no more than 10% of capital per position, and stand down if price breaks the first support before the EMA is reclaimed — that invalidates the reversal premise rather than creating a discount.

Buying quality dips when the trend holds: the case for the oversold-reversal setup

This idea is a mean-reversion strategy that buys oversold pullbacks (RSI below 35) in stocks still trading above their 50-day trend with rising trend strength, targeting ten large-cap tech and consumer names. It is a completed backtest, not a paper concept: on daily bars over the last 60 months, the CVS pair alone — the strongest of the three tested windows — returned 97.5% on 12 trades with a 75% win rate, and the deepest equity drawdown was held to 16.7%. The same ruleset also tested positive over the 24-month window (42.9% return, 7 trades) and the 12-month window (35.0% return, 3 trades), so the edge is not an artifact of one lucky stretch. The fundamentals of the buy list support buying weakness rather than chasing strength. Booking Holdings is the standout: Q2 2026 revenue of $7.35B was up 32.9% quarter over quarter, free cash flow of $3.64B rose 17.2%, and operating margin expanded more than 11 points to 34.0%. Microsoft's fiscal Q4 delivered $90.0B in revenue (up 8.6%) and $19.6B in free cash flow, with a 45.1% operating margin that ranks at the 99th percentile of its sector. CVS paired its recent 4.80% session gain with genuine cash generation — Q2 free cash flow of $5.65B, up 66.3% — and its full-year free cash flow sits at the 99.6th percentile of health-care peers. The pullback-buying logic also fits where each name is in its cycle. Salesforce just printed $3.53B in net income, up 67.4% quarter over quarter, with net margin jumping to 31.1% from 18.9% — a company whose 1.76% session dip looks more like noise than a thesis change. Lam Research grew revenue 15.1% to $6.72B with a 33.9% net margin, and its dividend growth rate of 16.8% a year signals management confidence. Even the laggards on the tape — Intel down 3.45%, Meta down 3.33% — are names where cash generation was recently strong (Meta's fiscal 2025 free cash flow was $46.1B), so the strategy's premise that sharp down days in quality names create buying opportunities has earnings support behind it. The rule structure itself is conservative: a 2.4% stop against a 4.8% take-profit, position risk capped at roughly 2.4% of equity, and a maximum 10% position size — the kind of asymmetric setup that produced the 75% historical win rate rather than a few outsized gambles.

What could break the thesis

The backtest that anchors this idea rests on just 12 completed trades across five years, and it did not win everywhere it could have. The 24-month window returned 42.9% on 7 trades with only a 57.1% win rate, and the 12-month window returned 35.0% on just 3 trades. Small trade counts mean the headline 97.5% return and 75% win rate are dominated by a handful of episodes, and exit fills were approximated on daily bars rather than intraday data — treat the reported drawdown and win rate as coarse, as the analysis itself flags. Meanwhile, the strategy's hard stop sits at 2.4% and take profit at 4.8% in a group of stocks where daily moves of 3% to 5% are ordinary: on the most recent session alone, INTC fell 3.45%, META fell 3.33%, and VNET fell 4.05%, meaning a single normal trading day can stop out a position before any thesis has time to work. Insider posture is uniformly negative. Every one of the ten names shows net open-market selling in the disclosures filed for the period ending June 30, 2026 — including $171.6M at CVS, $58.3M at Lam Research, $54.7M at Meta, $47.7M at Microsoft, and $45.0M at Oracle. Management figures offer little offset: Microsoft's Satya Nadella took $12.3M in fiscal 2025 pay and Meta's Mark Zuckerberg $25.1M, compensation sized to scale rather than signs of insider conviction at these levels. The fundamental backdrop is split, and several of the strategy's buy-list names are deteriorating. Meta's free cash flow collapsed 86.8% quarter over quarter to $1.7B and net margin fell 21.5 points to 26.1%, as $21.4B of quarterly capex weighs. Qualcomm's Q3 free cash flow fell 74.2% to $495M and net margin compressed 49.4 points to 20.1%. Intel swung to an $11.0B quarterly loss with net margin at negative 68.4% and debt-to-equity rising to 0.58 from 0.40. Oracle burned $5.4B of free cash flow in its latest quarter, and the Reuters piece on an Oracle–Blue Owl project delay sending ripples through AI financing undercuts the capex narrative supporting the whole cluster. Salesforce's operating margin slipped to 20.5% from 21.1% despite the earnings beat. Entry frequency is thin: even the best-performing pair produced only 12 trades in 60 months, so the setup waits for rare oversold-plus-trend conditions rather than harvesting the frequent mean reversion these volatile names offer. And because no robust parameter setup was established — the sensitivity evaluation exceeded its time budget and returned no recommendation — the published rules stand as-is, with no evidence they sit on a stable plateau rather than a favorable historical pocket. The unresolved question is the divergence inside the group: does a rule that buys weakness across all ten names equally capture quality dips like Booking and Microsoft, or does it catch falling knives in the cash-burning names like Intel, Oracle, and Meta?

Company context

### BKNG **Ownership and insiders.** For the period ending 2026-06-30, 13 reporting holders were counted holding 38619212 reported shares. Insider filings indicate net open-market selling, netting US$-26005137.22. Coverage: Deadline Passed. **Management.** Mr. Glenn D. Fogel is identified as chief executive. Mr. Glenn D. Fogel is listed as President, CEO & Director, with reported total pay of US$9.2M for fiscal 2025. Source: [Yahoo Finance quoteSummary](https://finance.yahoo.com/quote/BKNG/profile), retrieved 2026-09-26T04:23:26.99+00:00. **Dividends.** Trailing twelve-month distributions total USD 1.644 per share; the latest observed ex-date was 2026-09-11 at USD 0.42 per share. The latest annual comparison is +9.4%. Source: [Yahoo Finance corporate actions](https://finance.yahoo.com/quote/BKNG/history?filter=div), retrieved 2026-09-26T04:23:26.99+00:00. **Fundamental changes.** Latest reported comparisons: - Free cash flow: USD 3.6B for Q2 ending 2026-06-30 versus USD 3.1B for Q1 ending 2026-03-31 (+17.2%) · quarter over quarter - Operating Cash Flow: USD 3.7B for Q2 ending 2026-06-30 versus USD 3.2B for Q1 ending 2026-03-31 (+15.7%) · quarter over quarter - Operating margin: 34.0% for Q2 ending 2026-06-30 versus 22.9% for Q1 ending 2026-03-31 (+11.0 percentage points) · quarter over quarter - Revenue: USD 7.4B for Q2 ending 2026-06-30 versus USD 5.5B for Q1 ending 2026-03-31 (+32.9%) · quarter over quarter - Shares Outstanding: 751.4M shares for Q2 ending 2026-06-30 versus 774.9M shares for Q1 ending 2026-03-31 (-3.0%) · quarter over quarter Source: Reported company financial data. Comparisons use consecutive reported periods at the same cadence and compatible units. Relative percentage changes require a positive baseline; rate changes are expressed in percentage points. **Unavailable comparisons.** - Debt to equity: The latest period's value is unavailable. - Gross margin: The latest period's value is unavailable. - Net Income: The latest period's value is unavailable. - Net margin: The latest period's value is unavailable. - Return on equity: The latest period's value is unavailable. ### CRM **Ownership and insiders.** For the period ending 2026-06-30, 17 reporting holders were counted holding 51225795 reported shares. Insider filings indicate net open-market selling, netting US$-173295.25. Coverage: Deadline Passed. **Management.** Mr. Marc R. Benioff is identified as chief executive. Mr. Marc R. Benioff is listed as Co-Founder, Chairman & CEO, with reported total pay of US$11.4M for fiscal 2026. Source: [Yahoo Finance quoteSummary](https://finance.yahoo.com/quote/CRM/profile), retrieved 2026-09-26T04:23:26.99+00:00. **Dividends.**…

Scores

• Conviction score breakdown: 59
• Thesis support: 68
• Trade readiness: 45
• Risk quality: 55
• Backtest evidence: 60
• Fundamentals trend: 65

Watch items

• CVS — Daily close vs 50-period EMA
• CVS — RSI (14)
• CVS — Daily low vs first support
• ORCL — Daily close vs 50-period EMA
• ORCL — RSI (14)
• BKNG — Daily close vs 50-period EMA
• VNET — Daily close vs 50-period EMA
• CVS — Insider net open-market selling (2026-09-30 window, filings mid-October)
• CVS — Q3 2026 earnings report (period ending 2026-09-30, expected early November)