The Week in One Paragraph
What Worked, What Didn't
Worked: desk breakouts. MMM paid twice in three sessions (T1 the same day it filled, then the second leg's T1 on Thursday). CVX ground for a week and took both targets, T1 Thursday and T2 Friday, closing at $233.21. In a thrust tape, the paying entries were the ones that did not wait for a dip.
Worked: the stops. MPC stopped Tuesday at $469 and fell to $448.38 by Thursday; the stop saved about four points of slide. KMI and LLY both filled and stopped the same day, which stings and is also the design: a wrong post-earnings read cost a defined sliver, not a position. Three losses, all small, all printed.
Worked: the no-fill discipline. SPOT's trigger fired on an 11% gap through its entry zone; the printed levels said no chase, so the ledger holds no position at $583. The discipline's cost showed up elsewhere (below), but it was not paid here.
Didn't work: scanner momentum pullbacks. 30 resolutions, 6 wins, a 20.0% live win rate and -0.461R expectancy. The sheets were emitted into a bounce that failed, a thrust that reversed, and a Thursday Asia shock, and pullbacks kept pulling back: the entire Oct 21 sheet stopped within two sessions. One week is one regime, and the archive's 1,185-signal sample spans many; but the forward record is the forward record, and it is printed in full below.
Didn't work: conditional retests in a straight-line tape. NOC, LMT, CSX, and CCI all ran through or past their targets without ever dipping into their entry zones. The desk left money on the table by design; the lesson is about tape selection, not level selection. The scanner's Friday sheet, 9 of 10 breakouts, is the same lesson arriving from the other side.
The Week's Defining Results
Meta detonated the funding question (Wed AMC). Revenue $64.60B beat (about $63.4B expected, +26%); GAAP EPS $3.83 included a $9.89B legal charge, ex-charge $7.40 vs $6.54 expected. The print was not the problem; the balance sheet was: Q3 capex $34.91B, free cash flow negative $2.12B, and a future commitment stack now estimated at $810-825B with CDS at a record. The stock fell 6.98% Thursday to $648.85 and closed the week at $642.62, -9.4%.
Tesla spent and shrugged (Wed AMC). Revenue $28.05B, roughly in line; adjusted EPS $0.48 beat; operating margin just 2.6% with opex up 39%. The 2026 capex guide was cut to about $20B from over $25B, robotaxi revenue pushed to about Q2 2027. The stock ended the week +4.9% at $372.22: the market paid for the spending discipline, not the quarter.
The rest of the marquee. Netflix (Tue AMC): EPS $0.87 beat, revenue in line at $12.87B, soft Q4 guide; the stock fell 5.3% Wednesday. IBM beat (EPS $2.95, revenue $17.10B). Texas Instruments beat (EPS $2.62, revenue $6.06B, +28%) and guided up. Procter & Gamble beat (core EPS $2.00 vs $1.89; revenue $23.28B). Zions (Mon AMC) beat on EPS, $1.78 vs $1.64; its revenue print conflicts across sources ($887M vs $893M) and is not leaned on here.
France: affirmed, and wider anyway. Moody's affirmed Aa3 with a negative outlook on Friday; consensus had expected a downgrade. The market's verdict ran the other way all week: the OAT spread over Bunds hit 173.6bp Thursday, the widest since November 2011, the yield peaked at 5.14% (highest since July 2002), and Friday's spread stayed above 170bp. The affirmation assumes a deficit-cutting 2027 budget gets adopted; that is now a dated political test, not a rating question.
Friday's data cut both ways. Flash PMIs beat: manufacturing 56.0, a 53-month high; composite 57.6. Michigan sentiment finalized at 48.0, better than the 46.3 consensus and still the second-lowest final on record, with 1-year inflation expectations up to 4.7%. Growth is fine; the consumer's mood and inflation expectations are not.
Oil ripped into the weekend. WTI settled about $103.2 Friday (+3.8% on the day). Brent settled at $116.72 (+3.14%), its highest since April, per two sources; one futures table printed $111.61 instead. The conflict is flagged, not averaged; the direction (up hard, on supply fear) is not in dispute.
Week in Numbers
Credit Watch
| INSTRUMENT | LEVEL | READ |
|---|---|---|
| AI Cycle Risk Meter | 59 / 100 RED | Fri Oct 23 edition (+8 in a day; flip rule fired). Credit component 82 |
| CoreWeave 5-yr CDS | 1,065.84 bp | Oct 21 (FactSet via CNBC). Stock $55.60 Thu, -10.06%; Bernstein target cut to $51 |
| Oracle 5-yr CDS | 308.25 bp | Oct 21 (FactSet via CNBC); from 141 bp on Jan 2 |
| Meta 5-yr CDS | ~114.5 bp | Record (Markit), on the $810-825B commitment stack |
| High-yield spread | 332 bp | Latest verified print, dated Oct 20 |
| France 10-yr spread | >170 bp | Thu peak 173.6 bp, widest since Nov 2011; Moody's affirmed Aa3 Fri |
CDS provider variance is wide in this complex; the dated FactSet pair above is the print of record for this page. The meter's three named drivers: Meta's commitment stack and negative FCF, CoreWeave's credit breaking to new lows (capital, not demand, is the constraint), and the KOSPI reversal (-3.52% Thu) on heavy foreign selling.
The Record
Desk ledger: 8 resolved (3 wins, 3 losses, 2 no-fills), 5 positions open, 7 unfilled and watching.
Scanner ledger: 30 resolved (6 wins, 24 losses), 38 open. Momentum pullback flagged by the learner; full tables below.
Sector Rotation: The Week
Energy and the defensive block led all week; the Meta-heavy communication sector was the week's casualty, and small caps beat large caps by about two and a half points. The tape rotated out of the AI complex's most leveraged names and into cash-flow sectors, without ever becoming a liquidation: eight of eleven sectors finished green.
| SECTOR SPDR | WEEK | READ |
|---|---|---|
| Energy (XLE) | +1.60% | Oil's Friday rip; the week's best sector |
| Health Care (XLV) | +1.45% | Defensives bid all week |
| Real Estate (XLRE) | +0.97% | Rates eased about 10 bp on the week |
| Utilities (XLU) | +0.83% | Thursday's +2.07% session carried it |
| Materials (XLB) | +0.77% | Steady, unremarkable |
| Consumer Staples (XLP) | +0.73% | Thursday's +3.31% spike, partly given back |
| Financials (XLF) | +0.68% | Kept pace with the tape |
| Consumer Disc. (XLY) | +0.57% | Tesla's week helped; breadth inside was mixed |
| Technology (XLK) | -0.13% | Flat: the AI complex's standoff week |
| Industrials (XLI) | -1.87% | Second-worst; profit-taking in the run's leaders |
| Communication (XLC) | -4.36% | Worst by far; Meta is 21.67% of the fund and fell 9.4% |
Source: Finnhub closes, Oct 16 vs Oct 23; price returns, dividends excluded.
The Scoreboard, Full
Desk ledger: every printed setup, resolved or carried
| CALL | SHEET | RESULT |
|---|---|---|
| CVX energy breakout | Oct 15 | WIN: T1 $229 touched Thu Oct 22; ran through T2 $233 Friday, closed $233.21 |
| AAPL breakout continuation | Oct 15 | OPEN: last $347.73; T1 $355; stop $337.50 untouched |
| MPC energy continuation | Oct 15 | LOSS: stopped Tue Oct 20 (low $464.00 vs the $469 stop); fell to $448.38 by Thu |
| GILD earnings momentum | Oct 15 | NO FILL: trigger never fired; closed in last week's Ledger |
| MMM post-earnings | Oct 20 | WIN: filled and hit T1 $167.95 the same day |
| CNQ breakout | Oct 20 | OPEN: filled Wed; last $55.06 (Oct 23); T1 $58, stop $49.40 |
| V pullback | Oct 20 | NO FILL through Fri (4 sessions; price ran to $407.92 above the zone); expires Mon Oct 26 |
| MCK defensive breakout | Oct 20 | NO FILL through Fri (ran through the zone top to $1,000.77); expires Mon Oct 26 |
| MMM second leg | Oct 21 | WIN: T1 $176 touched Thu Oct 22 |
| NOC post-earnings pullback | Oct 21 | NO FILL: never dipped to the zone; ran through T1 to $551.44. Expires Wed Oct 28 |
| V pullback, re-issued | Oct 21 | NO FILL through Fri (3 sessions); expires Wed Oct 28 |
| MU oversold bounce | Oct 21 | OPEN: triggered Wed; Thu low $849.25 held above the $840 stop; last $877.55 |
| KMI breakout retest | Oct 22 | LOSS: filled and stopped the same day Thu (low $32.46 vs the $33.15 stop) |
| CSX earnings gap retest | Oct 22 | NO FILL yet (2 sessions; last $50.79 after Thursday's +6.76%) |
| LMT post-earnings retest | Oct 22 | NO FILL yet (2 sessions; last $604.18 after Thursday's +6.96% beat run) |
| SPOT conditional reclaim | Oct 22 | NO FILL: trigger fired on a gap through the entry zone; unfilled at printed levels |
| CCI gap retest, conditional | Oct 23 | NO FILL yet (1 session; last $85.97, already above T1 $83.26 without a retest) |
| LLY breakout retest | Oct 23 | LOSS: filled on Friday's dip and stopped the same day (low $1,257.62 vs the $1,268 stop; closed $1,273.46) |
| KO all-time-high retest | Oct 23 | OPEN: triggered Fri; last $92.71; stop $91.90 |
| AAPL new-high continuation | Oct 23 | OPEN: triggered Fri; last $347.73. A separate printed setup from the Oct 15 sheet; tracked separately |
Desk tally: 8 resolved (3 wins, 3 losses, 2 no-fills), 5 positions open, 7 unfilled and still watching. Rules unchanged: triggered positions resolve at target 1, the stop, or 15 sessions; unfilled conditionals expire after 5 sessions.
Scanner ledger: all 30 resolutions to date, every one printed
| SHEET (CLOSE) | RESOLUTIONS |
|---|---|
| Oct 16 (Fri) | MEDP win +1.72R (3 sessions) · WAT win +1.60R (4) · AME loss (4) · EQH loss (5) · CRBG loss (5). MRK still open |
| Oct 19 (Mon) | MEDP win +1.56R (2) · WAT win +2.17R (3) · IQV win +1.60R (3, beyond the published ten) · GOOG loss (2) · GOOGL loss (2) · AME loss (3) · EQH loss (4) · CRBG loss (4) · MSM loss (3) · TT loss (3) · IEX loss (3) |
| Oct 20 (Tue) | DHR win +1.52R (2) · TEL loss (2) · TT loss (2) · AME loss (2) · NDSN loss (2) · IEX loss (2) |
| Oct 21 (Wed) | BRKR loss (1) · JCI loss (1) · QCOM loss (1) · TEL loss (1) · CR loss (1) · MCHP loss (1) · RGEN loss (2). The whole sheet stopped into Thursday's Asia break |
| Oct 22 (Thu) | GAP loss (1). The breakout-heavy sheets from Wed and Thu closes are still open and resolve next week |
Scanner totals: 68 signals emitted across five sheets; 30 resolved, all momentum pullbacks: 6 wins, 24 losses, win rate 20.0%, expectancy -0.461R. Published Top 10 only: 4 wins, 18 losses of 22 resolved. 38 signals remain open, including every breakout, gap, and cup signal so far; none of those three patterns has a live resolution yet.
The learner has already moved
Momentum pullback's ranking factor fell from 0.556 (its archive prior) to 0.485 after the 30 live resolutions, and the pattern is flagged under the standing rule (posterior below 50% on 20 or more live resolutions). The flag is a human review marker, not a removal: the pattern still prints, but it ranks below breakouts (factor 0.717), gaps (0.624), and cups (0.561) until its forward evidence improves. Thresholds are never auto-tuned; only the ranking learns. Monday's sheet will show the new order.
Levels That Matter Next Week
| INSTRUMENT | LEVEL | WHY |
|---|---|---|
| S&P 500 support | 7,600 | Round number under Friday's 7,623.68 close; the week's low ground |
| S&P 500 overhead | 7,700 / 7,744 | Then the record close 7,818.93; the range is the story until one side breaks |
| SPY | $759.97 | Support $758 below; $769-770 congestion overhead |
| QQQ | $704.40 | Week -0.76%; the $725-727 area (50-day zone) is overhead supply |
| VIX | 19.45 | The 20 line into an FOMC and PCE week; a close above it changes the desk's tone |
| 10-year yield | 5.00% | PCE Thursday is the test; 5.20% is the warning level |
| Meta | $642.62 | Post-print close; the commitment story, not the chart, drives it now |
| France spread | >170 bp | Into a budget fight the Moody's affirmation explicitly depends on |
Next Week: Oct 26-30
Mon Oct 26: the quiet day. Light data (Chicago Fed, Dallas Fed; 2-year auction); China's Fifth Plenum opens. Verizon before the open; Celestica and Cadence after.
Tue Oct 27: the FOMC convenes. September durable goods (consensus +1.40%), October consumer confidence (80.7), September new home sales (645K). Earnings: SoFi, Coca-Cola, Boeing, UPS, PayPal before the open; Visa and Robinhood after.
Wed Oct 28: FOMC decision, 2:00 PM ET. A hold at 3.75-4.00% is expected (94 of 100 economists in the Reuters poll); the priced event is December, where a hike carries roughly 81-89% odds across the week's readings. After the close: Microsoft and Alphabet, plus ServiceNow, KLA, T-Mobile, FedEx, Chipotle.
Thu Oct 29: Q3 GDP advance print (consensus +3.0%), jobless claims (195K), and September PCE: headline 3.6% y/y expected, core 3.1% vs 3.0% prior. Earnings: Eli Lilly, Mastercard, Caterpillar, Merck before the open; Amazon and Intel after. The desk's LLY setup stopped Friday; the print now matters as a sector read, not a position.
Fri Oct 30: Q3 Employment Cost Index, Chicago PMI (consensus 53.0, a sharp drop from 58.8), Exxon, Chevron, and AbbVie before the open. Month-end: the last trading day of October.
About a third of the S&P 500 reports this week. Beyond it: Apple and Palantir report Monday Nov 2, and the midterms follow Tuesday Nov 3.
Watchlist for Monday
- AAPL, two open setups (Oct 15 and Oct 23 sheets, both last $347.73, stops $337.50 and $338): Apple reports Nov 2. The desk's swing horizon runs into that print; whether to hold through it is the holder's call, flagged here a week early on purpose.
- MU, open at $877.55: Thursday's $849.25 low is the warning line above the $840 stop.
- KO, open at $92.71 with the stop at $91.90: the tightest open risk on the board, and Coca-Cola reports Tuesday.
- CNQ, open at $55.06: T1 $58 is within reach; energy led the week and oil ripped Friday.
- V and MCK expire Monday unfilled unless they dip into their zones; both ran away upward instead.
- Monday's scanner sheet prints from Friday's close: Thursday's sheet was 9 of 10 breakouts, and pullbacks now rank last by factor. Whether breakouts start resolving better than pullbacks did is next week's first real data point.
- The week itself: FOMC Wednesday, PCE Thursday, Microsoft and Alphabet Wednesday night, Amazon Thursday night. Position sizing into that stack is a decision, not a default.
Sources and Notes
Index and Friday closes: StockMarketWatch market digest, MarketBriefer Global Close, and Edge Consultancy wrap, all Oct 23, 2026 (three agreeing sources; a fourth set printed by a Morningstar/PR Newswire piece conflicted and was set aside). Yields: Treasury official daily par curve via the Edge wrap. Sector week returns: Finnhub closes Oct 16 vs Oct 23, computed by the desk. Earnings: MarketBeat, Finnhub, and Stockwatch reports of the Oct 20-22 prints. PMI: S&P Global flash via ShareCast; Michigan final: Reuters. France: Reuters and Longbridge on spreads; Moody's action as reported Oct 23. CDS: FactSet via CNBC, dated Oct 21. Calendar and consensus: Morningstar week-ahead (Oct 23), Schaeffer's, Reuters FOMC poll; FedWatch odds are a dated range across the week's readings, not a single print. All desk and scanner resolutions were computed by the desk from the scanner's refreshed price cache under the published rules (own stop and target, first touch, stop first within a session; LLY and KMI scored on intraday stop touches, the same convention that stopped MPC on Oct 20). CNQ's Friday close is Finnhub's dated Oct 23 quote; CNQ is outside the scanner cache's universe. Zions revenue and Brent's Friday settle carry source conflicts, flagged where printed. Educational content only; nothing here is investment advice.