What's going on, everybody?
I hope you all had a good week. Here are two sentences about the same five sessions and both of them are true. The S&P 500 fell 1.04 percent. Nine of the eleven sectors inside it went up. If you only looked at the first number you would have filed this away as a down week and missed the most interesting five days I have written about since starting this.
The reason they disagree is weight. Technology fell 3.62 percent and communication services fell 4.46 percent, and those two carry more of the index between them than the nine sectors that rose. So the headline described about a fifth of the market and the other four fifths quietly had a good week. The Dow rose 1.32 percent. The average stock rose 1.50 percent. Even small caps went up, which had not happened in this sequence before.
Thursday is where it concentrated, and Thursday is the whole issue. The S&P 500 had its worst session since July, the Philadelphia semiconductor index fell 6.72 percent in that one day, and the Dow closed it almost exactly flat. A 6.7 point gap between those two in a single session is the widest dispersion reading I have on record. And in the middle of that, every one of this week's five picks reported third quarter results, and every one of them went up. I will tell you now rather than at the end that this is the single biggest risk in the issue. Let's get into it.
The Week in 60 Seconds
- Nine of eleven sectors rose and the S&P 500 still fell 1.04 percent, because the two that fell carry more index weight than the nine that rose.
- The Dow Jones Industrial Average rose 1.32 percent while the Nasdaq Composite fell 2.64 percent, and the equal weight S&P 500 rose 1.50 percent.
- The Philadelphia semiconductor index fell 6.72 percent on Thursday alone and 7.33 percent for the week, while the Dow finished Thursday almost exactly flat.
- Meta fell 11.93 percent and Alphabet 7.16 percent, so the cash rich platforms that were last week safe side of the artificial intelligence trade got sold too.
- All five picks reported third quarter results on Thursday October 22 and all five rose, on the day the S&P 500 had its worst session since July.
This Week in One Minute
Nine of eleven sectors rose and the index still fell.
Both halves of that sentence describe the same week and the gap between them is the issue. From the October 15 close to the October 22 close the S&P 500 fell 1.04 percent and the Nasdaq Composite fell 2.64 percent. Over the identical days the Dow Jones Industrial Average rose 1.32 percent, the equal weight S&P 500 rose 1.50 percent and the Russell 2000 rose 0.61 percent. The index fell anyway because the two sectors that fell carry more weight between them than the nine that rose.
Now the part that matters more than the arithmetic, and it is a correction to something I told you last week. I said the market was repricing who pays for artificial intelligence capacity rather than whether it gets built, and my evidence was that the companies funding it out of operating cash flow were being spared while the ones that must raise capital were cut 10 to 18 percent. I named Microsoft, up 3.22 percent, and Meta, up 2.20 percent, as the proof. This week Meta fell 11.93 percent, Alphabet fell 7.16 percent, Oracle fell 6.64 percent and Microsoft fell 3.40 percent. The distinction I built that call on stopped existing within five sessions.
What is left is simpler and harder to argue with. The market is selling anything whose payoff lives in a capital plan and buying earnings that are already in the bank. Quanta Services, an electrical contractor whose backlog is the artificial intelligence buildout, fell 13.98 percent. Vertiv fell 8.92 percent, Eaton 8.32 percent, Advanced Micro Devices 9.66 percent, Micron 8.03 percent. Apple rose 1.82 percent and Tesla rose 3.35 percent, and those are the two mega caps with the least capital spending exposure in the group. The repricing is not about whether artificial intelligence works. It is about who is holding the bill while it gets built, and that question has now reached everybody in the chain rather than just the borrowers.
One more thing I would rather tell you than let you discover. All five picks in this issue reported third quarter results on Thursday October 22, and each of them made the majority of its weekly gain in that single session, on the day the index had its worst day in three months. Five catalysts on one date is not five independent reasons to own things. It is one flow choosing a destination. The earnings explain which five names rather than the 38 I rejected. They do not explain the size of the response, and I cannot separate the two with what I have.
The Three Readings That Matter Breadth. Nine of eleven sectors finished higher while the index fell 1.04 percent, because technology and communication services carry more weight between them than everything that rose. Rates. The long end fell again, 10 year Treasury yields from 5.09 to 4.95 percent, in a week the VIX rose 15.61 percent, which makes the bond bid a flight from equity risk rather than a dovish repricing. Concentration. All five picks reported on the same Thursday, a concentration I am flagging rather than celebrating, because it leaves four of the five with no scheduled catalyst for roughly three months. The Week Ahead
What I expect. From the October 22 close to the October 29 close I expect the equal weight S&P 500 to beat the cap weighted S&P 500 again, and I am putting 62 percent on it rather than last week 68 percent for a reason that has nothing to do with the result. Last week the call survived by 2.51 points and the argument behind it did not. I said the market was separating the companies that fund artificial intelligence capacity out of operating cash flow from the companies that have to raise capital, and I named Microsoft up 3.22 percent and Meta up 2.20 percent as the proof. This week Meta fell 11.93 percent, Alphabet 7.16, Oracle 6.64 and Microsoft 3.40. The distinction I built the call on stopped existing inside five sessions. What is left is simpler: the market is selling anything whose payoff is a capital plan and buying earnings already in the bank, which is why nine of eleven sectors rose while the index fell. I expect that to hold one more week because Microsoft and Alphabet publish capital spending plans on October 28 and Amazon on October 29, and every quarter this cycle those plans have gone up. The number itself is close to a one way bet. How the market reads it is not, and that is what caps my confidence here.
What Would Prove Me Wrong I am wrong if the equal weight S&P 500 trails the cap weighted S&P 500 by more than 1.5 percentage points from the October 22 close to the October 29 close, measured as the percentage change in RSP minus the percentage change in SPY. The spread starts at zero and an ordinary week moves it about 1.28 points, so this line sits 1.17 standard deviations away against 0.59 for last week line. I am telling you plainly that this is a weaker commitment than last week, not a stronger one, because a line further from the market gets reached less often. The reason I moved it is that the last five sessions have run near 2.0 points of weekly spread volatility rather than 1.28, so a 0.75 point line would now sit inside ordinary noise and would be settled by the tape rather than by whether I am right. If it survives with more than a point of clearance, discount it accordingly. The odds I put on it: 62 percent, Moderate confidence, down from 68 percent last week.
| Date | What happens | Why it matters |
|---|
| Oct 28 | The Federal Reserve announces its decision at 2:00 in the afternoon eastern time, with no new economic projections, per its published meeting calendar. | Long yields have now fallen two weeks running into above target inflation. A hawkish read puts the long end back up and removes the tailwind under four of these five picks. | | Oct 28 | Microsoft reports fiscal first quarter results and Alphabet reports third quarter results, both stating capital spending plans. | This is the resolution of the whole week. Plans that rise and get accepted end the rotation; plans that rise and get refused extend it. It also decides the biggest risk in this issue. | | Oct 29 | Third quarter gross domestic product and the September Personal Income and Outlays report, both at 8:30 in the morning, per the Bureau of Economic Analysis schedule. | The inflation print is the direct test of the cost headwind inside the Procter and Gamble guide and of whether the duration bid under this rotation survives. | | Oct 29 | Amazon reports third quarter results after the close. | The third of three capital spending disclosures inside 48 hours, and the one most likely to be read as a verdict on cloud demand rather than on spending alone. | | Nov 2 | Apple reports fiscal fourth quarter results. | Apple rose 1.82 percent this week while every other mega cap technology name fell, and it is the mega cap with the least capital spending exposure. Whether that holds says how specific this repricing really is. |
The Single Biggest Risk, Dated And Priced The risk: The October 22 rotation flow reverses. Microsoft, Alphabet and Amazon publish capital spending plans the market can live with, the artificial intelligence complex snaps back, and the same one day flow that paid for all five of these picks goes back where it came from, taking the support out of all five at once. Why I think so: This is the governing risk because the concentration here is total rather than partial. All five picks reported on the same day and all five made the majority of their weekly gain in that one session: Roper Technologies 9.73 percent, Allegion 11.02, Nasdaq 7.47, Procter and Gamble 5.58 and Dow 4.71, while the S&P 500 fell 1.21 percent and the semiconductor index fell 6.72. Five companies in five unrelated industries do not move together on their own fundamentals. They move together because one flow chose them on one afternoon. Nothing that happened to the complex being sold was a cash flow event: no artificial intelligence customer cancelled an order and no hyperscaler cut a plan. And four of these five have no scheduled company catalyst for roughly three months, so if the flow turns there is nothing company specific left to defend them until late January. What would confirm it: The technology sector fund XLK closing back above 191 on or before November 5, 2026. It closed at 182.34 on October 22, so 191 is 4.75 percent higher and would put the sector back above where it stood on October 15, meaning this entire week decline erased rather than paused. I put it at 30 percent by November 5, 2026. Last Week’s Call, Graded What I said in Issue No. 17: I expected the equal weight S&P 500 to beat the cap weighted S&P 500 from the October 15 close to the October 22 close, at 68 percent, on the argument that the market was repricing who finances artificial intelligence capacity rather than whether it gets built. Verdict: SURVIVED. The equal weight S&P 500 rose 1.50 percent while the cap weighted version fell 1.01, so equal weight beat cap weight by 2.51 percentage points against a line I drew at 0.75 points the other way. All five of that issue watch items held too, which takes the small forecast ledger to 50 of 50 logged and resolved at an 86 percent hit rate. Now the part that matters more than the verdict: the reasoning was wrong. The tell I published was that the companies funding the buildout from operating cash flow were being spared, and I named Microsoft up 3.22 percent and Meta up 2.20 percent as the evidence. Within five sessions Meta fell 11.93 percent, Alphabet 7.16, Oracle 6.64 and Microsoft 3.40. The distinction did not survive the week after I wrote it down. Being right for the wrong reason is not a win, and it is why this week call is priced lower rather than higher. By how much: I said I would be wrong if equal weight trailed cap weight by more than 0.75 points. It beat by 2.51. The closest the spread came to my line was minus 0.06 points on October 19, which left 0.69 points of room, 0.55 of a standard deviation, so the grader did not flag a near miss and I am not going to claim one. The comparison I would rather be judged on is the baseline: a random walk carrying this spread own measured volatility would have reached that line with probability 0.55, and it was not reached. Across ten recorded calls I am surviving 50 percent against 30 percent for that baseline, a 20 point edge on a sample that is still small enough to be luck. One loose end stays open and I am not closing it early: the biggest risk I published in Issue No. 17 needed XLK above 200 by October 29. It closed at 182.34 and its best close in the window was 188.54, so that one is still live and gets settled next week rather than quietly marked a miss today. Market Risk Score
Market Risk Score: 61 / 100 (Elevated). Up from 58 last week, and the reason is not the index, which barely moved. Breadth is the best I have measured and the long end fell 14 basis points, both of which argue for a lower number. Three things argue higher and they win. The VIX rose 15.61 percent to 19.03, so unlike last week volatility is now being priced rather than ignored. The semiconductor index lost 7.33 percent on the week with a 6.72 percent single session. And Brent crude closed near 112 dollars, a five month high, after Houthi attacks on Saudi targets and reports of possible renewed strikes against Iran.
The deciding factor is next week rather than this one. The Federal Reserve decision, the September inflation print, third quarter gross domestic product and the capital spending plans of Microsoft, Alphabet and Amazon all land inside 48 hours. That is the densest event cluster of the quarter and it sits directly on top of the question this week was about.
A Caveat On My Own Number I have to tell you that this score is currently measured as INVERTED. Across the eight readings I can score, higher risk scores have preceded stronger forward markets, with a correlation of plus 0.57. That is the opposite of what the number is supposed to do. So read 61 as a description of conditions, not as a signal about direction, and do not take it as me telling you to be defensive. I am publishing it with the flaw attached rather than quietly dropping it, because a measure that is wrong and disclosed is more useful than one that is wrong and confident. Rebuilding its factors is on my list. Sector Rotation Update
Into: Consumer Staples · Utilities · Financials · Health Care. Out of: Communication Services · Technology · Industrials · Real Estate.
| Sector | Week | Read |
|---|
| Consumer Staples | +3.82% | Leading | | Utilities | +2.93% | Leading | | Financials | +2.50% | Leading | | Health Care | +2.32% | Leading | | Materials | +2.23% | Neutral | | Energy | +1.52% | Neutral | | Consumer Discretionary | +1.41% | Neutral | | Real Estate | +1.26% | Slipping | | Industrials | -0.69% | Slipping | | Technology | -3.62% | Lagging | | Communication Services | -4.46% | Lagging |
The mechanism, traced. Money is leaving the artificial intelligence capital spending complex and it is arriving in staples, utilities, financials and health care. The arrival side is where I want to slow you down, because three of those four are defensive and two of them did not move on company news at all. Every large utility rose between 2.06 and 5.83 percent on a 14 basis point fall in long yields. That is a bond move wearing an equity ticker, and it tells you nothing about the economy.
Industrials is the sector to actually look at, because at minus 0.69 percent the number describes neither half of it. The artificial intelligence power and electrical chain was destroyed: Quanta Services minus 13.98 percent, Vertiv minus 8.92, Eaton minus 8.32, Parker Hannifin minus 7.13, GE Aerospace minus 5.92, Caterpillar minus 4.29. Defence and classic industrials rose hard: Northrop Grumman plus 10.55 percent, Waste Management plus 6.65, 3M plus 6.19, CSX plus 6.12, L3Harris plus 5.71. That is a spread of more than 24 points inside one sector in one week. If you hold a sector fund here you own both halves and the fund told you nothing happened.
Durability: moderate. A broadening led by defensives is a rotation out of a crowded theme rather than a cyclical expansion, and it has a shorter half life. The one piece of evidence on the other side is real and I am not going to bury it: small caps participated for the first time in this sequence, with the Russell 2000 up 0.61 percent after falling 0.96 percent the week before. Good for weeks rather than quarters, and dependent on long yields staying where they are. I will also say plainly that my own record says sector labels are close to a coin flip at one month, 57 percent accurate across 67 graded calls, so read these as a description of what happened rather than a forecast of what is next.
Breadth: The Index Fell And The Market Rose
The measurement. The equal weight S&P 500 rose 1.50 percent while the cap weighted version fell 1.01 percent, a spread of 2.51 percentage points in favour of the average stock. That is the widest positive reading I have recorded and the second week running above 1.5 points, against 1.76 at the last issue, 0.38 before that and minus 0.52 before that. Four weeks ago the average stock was losing to the index. Now it is beating it by two and a half points.
Who is doing the dragging. This is the opposite of the usual breadth problem, where a handful of giants carry a rising index. Here the giants are pulling a rising market down. Over the window: Meta minus 11.93 percent, Advanced Micro Devices minus 9.66, Micron minus 8.03, Arista minus 7.82, Netflix minus 7.66, Alphabet minus 7.16, Oracle minus 6.64, Broadcom minus 6.12, Microsoft minus 3.40, Nvidia minus 2.31. Ten companies. Outside them the market was broadly higher, and Apple at plus 1.82 percent and Tesla at plus 3.35 percent are the two mega caps that rose.
What it means. A falling index with rising breadth is not a warning, it is an arithmetic artefact of concentration, and this is the cleanest demonstration of it I have been able to show you. The useful frame for this week is dispersion rather than direction: a 3.96 point gap between the Dow and the Nasdaq Composite over five days, and a 6.7 point gap between the Dow and the semiconductor index in one session. What the breadth is telling you is that money is leaving one crowded theme and finding the rest of the market, and that for the first time it is going down the capitalisation scale rather than stopping at the edge of the large cap index.
What Resolves It October 28 and 29, and it resolves either way. Microsoft and Alphabet report on the 28th and Amazon on the 29th, and each will state capital spending plans. If those plans rise and the market accepts them, the largest index members stop falling and this spread closes from the top. If those plans rise and the market refuses them, the spread widens and the breadth stops being a curiosity and becomes a bear market in ten stocks that happens to be dragging everyone’s index fund with it. The Federal Reserve decision on the 28th and the September inflation print on the 29th sit inside the same 48 hours, so there is no clean read available until the week after next. That is the honest answer about timing and it is also why the odds on my forward call are lower this week than last. The Five
Every return below is measured from the October 15 close to the October 22 close. All five reported third quarter results on October 22. No entry zones, no stops, no targets: this is the free issue and it stays that way.
01 · Technology / Application Software · +14.06%
ROP Roper Technologies, Inc.
On October 22 Roper Technologies reported third quarter revenue of 2.17 billion dollars, up 8 percent with 7 percent organic growth, and adjusted earnings of 5.89 dollars a share, up 15 percent. It raised full year adjusted guidance to between 22.35 and 22.40 dollars from 22.15 to 22.30, and guided the fourth quarter to 5.95 to 6.00 dollars. The shares rose 9.73 percent that day on 1.98 million shares against a weekly average under one million, and finished the week up 14.06 percent. One honest caveat before you go further: 3.84 points of that week arrived on October 20, two sessions before the numbers, and I could not find a company announcement for that date. About a quarter of the move is unexplained and I would rather say so than hand you a reason I made up.
Why it made the five: It made the five because it is the cleanest test of what actually happened this week. The market did not sell technology; it sold one kind of claim on the future. Inside the sector the names whose artificial intelligence story is a spending commitment were cut hardest, Advanced Micro Devices 9.66 percent, Micron 8.03 percent, Arista 7.82 percent, Broadcom 6.12 percent, Oracle 6.64 percent. Roper Technologies is in the same sector and rose 14.06 percent, and it is not a defensive business, so defensiveness does not explain it. I checked the easy alternative, that the rotation simply lifted anything outside the index leaders: on October 22 Roper Technologies rose 9.73 percent while its own sector fund fell 3.29 percent, a 13 point gap inside one sector in one session. A rotation does not do that. An earnings report does.
The case against it: Two thirds of the headline earnings growth was not the business. Adjusted net earnings rose 5 percent to 586 million dollars. Adjusted earnings per share rose 15 percent. The whole difference is the share count, which fell from roughly 109 million to 99.4 million over the year, about 9 percent of the company bought back. That is real money for a holder and it is not evidence the business improved, and it cannot repeat at that scale forever. Worse for the quality of the quarter, organic revenue grew 7 percent and adjusted EBITDA grew 7 percent while adjusted net earnings grew only 5, so the adjusted margin went slightly the wrong way. At about 19 and a half times the guidance it raised that morning the price is not demanding, and the capital allocation is genuine rather than cosmetic, with more than 6 billion dollars of deployment capacity and 1.2 billion dollars received from selling the Indicor instrumentation businesses. But if you read the 15 percent without the 5 percent beside it, you have misread the quarter.
The Catalyst, With The Source On October 22, 2026 Roper Technologies reported third quarter revenue of 2.17 billion dollars, up 8 percent with 7 percent organic growth and 1 percent from acquisitions, adjusted EBITDA of 864 million dollars up 7 percent, GAAP diluted earnings of 4.88 dollars a share up 33 percent, and adjusted diluted earnings of 5.89 dollars a share up 15 percent on 99.4 million weighted average diluted shares. It raised full year 2026 adjusted diluted earnings guidance to a range of 22.35 to 22.40 dollars from a prior range of 22.15 to 22.30, and guided fourth quarter adjusted diluted earnings to 5.95 to 6.00 dollars. Source: U.S. Securities and Exchange Commission, Roper Technologies, Inc. Form 8-K, third quarter 2026 earnings release, October 22, 2026. Read it yourself. What To Watch Next Nothing is scheduled. Fourth quarter results are expected in late January 2027, on the company own cadence rather than on any announcement, and management has already told you the number to hold it to: adjusted diluted earnings of 5.95 to 6.00 dollars a share. Confirms the thesis: Landing inside that range with organic growth still near 7 percent, which would mean the raise came from operating performance rather than from the buyback. Breaks it: Hitting the range while organic growth slows, because then the earnings line is being carried by the share count and the 15 percent does not repeat.
02 · Industrials / Building Products · +13.34%
ALLE Allegion plc
Allegion makes locks, door hardware and electronic access control. On October 22 it reported third quarter revenue of 1,172.6 million dollars, up 9.6 percent and up 7.2 percent organically, adjusted earnings of 2.60 dollars a share against a 2.50 dollar consensus, and an adjusted operating margin of 24.7 percent against 24.1 percent a year earlier. It raised full year adjusted earnings guidance to between 9.05 and 9.10 dollars. The shares rose 11.02 percent that session, the largest single day move in this issue, and 13.34 percent on the week.
Why it made the five: It made the five because the growth sits in the one place that is hardest to fake and easiest to check. Americas revenue grew 11.7 percent reported and 9.1 percent organically, led by double digit organic growth in the non residential business, which is offices, schools, hospitals, warehouses and data centres. That is a specification business: the hardware gets written into a building design months before it ships, so a strong quarter reports on construction decisions already made rather than on how anybody felt this week. The margin is what confirms it. Getting the adjusted operating margin to 24.7 percent from 24.1 while growing 7.2 percent organically means price and volume arrived together, and that combination is absent almost everywhere else in this issue.
The case against it: One region is carrying the entire company. International revenue rose 1.8 percent reported and FELL 0.2 percent organically, so every point of real growth came from the Americas. The guidance deserves a harder read than the word raise suggests: a full year organic range of 5.0 to 6.0 percent, set after a quarter that printed 7.2 percent, is management telling you the fourth quarter decelerates, and that is arithmetic inside their own numbers rather than my inference. The part that should make you least comfortable is the part that looks best. Non residential construction includes data centre construction, and the thing the market spent this entire week repricing is how much data centre capacity actually gets built. If that spending slows, the strongest segment in this pick loses its strongest driver, and the 11 percent paid on Thursday rests on a quarter that closed on September 30.
The Catalyst, With The Source On October 22, 2026 Allegion reported third quarter revenue of 1,172.6 million dollars, up 9.6 percent on a reported basis and 7.2 percent organically, with reported earnings of 2.59 dollars a share, adjusted earnings of 2.60 dollars a share and an adjusted operating margin of 24.7 percent against 24.1 percent a year earlier. Americas revenue grew 11.7 percent reported and 9.1 percent organically on double digit organic growth in non residential, while International revenue rose 1.8 percent reported and fell 0.2 percent organically. The company raised full year 2026 guidance to reported revenue growth of 9.5 to 10.5 percent, organic revenue growth of 5.0 to 6.0 percent and adjusted earnings of 9.05 to 9.10 dollars a share. Source: Allegion plc, Allegion Reports Q3-2026 Financial Results, October 22, 2026. Read it yourself. What To Watch Next Nothing company specific is scheduled; fourth quarter results are expected in mid February 2027 on the company own cadence. The nearer test is the data centre capital spending the market is currently repricing, and the first reads on that arrive with the Microsoft and Alphabet reports on October 28. Confirms the thesis: Hyperscaler capital plans holding or rising, which keeps the non residential pipeline that produced 9.1 percent Americas organic growth intact. Breaks it: A hyperscaler trimming planned capacity, because the best segment in this pick is partly a bet on buildings those plans pay for.
03 · Financials / Financial Exchanges and Data · +10.48%
NDAQ Nasdaq, Inc.
Nasdaq the company rose 10.48 percent in the week the Nasdaq Composite fell 2.64 percent, and the reason is that they are barely the same business any more. On October 22 it reported net revenue of 1.481 billion dollars, up 13 percent and 14 percent organically, with Solutions revenue of 1.172 billion dollars up 17 percent, index revenue of 286 million dollars up 39 percent, annualised recurring revenue of 3.365 billion dollars up 12 percent, and non-GAAP earnings of 1.04 dollars a share up 18 percent. The shares rose 7.47 percent that day.
Why it made the five: It made the five because of a distinction most readers have never been shown, and it is worth more than the pick. Market Services, the part that is actually an exchange taking a cut of trades, produced 309 million dollars of the 1.481 billion dollars of net revenue and grew 2 percent. The growth came from index licensing at 39 percent and from software sold on subscription, with 3.365 billion dollars of annualised recurring revenue. So about a fifth of the business is the exchange and the fast part is data and software. That is why a 2.64 percent fall in the index it is named after did not touch it: the company is not long the market, it is paid for being the plumbing and the measuring stick. The 57 percent non-GAAP operating margin is what a toll business looks like from the inside.
The case against it: The best line in the quarter is the one this week argues against. Index revenue grew 39 percent because it is charged as a share of the assets tracking Nasdaq indexes, and those assets are the same Nasdaq constituents that just fell 2.64 percent in a week and 7.33 percent in semiconductors. If the index keeps falling, that 39 percent decelerates mechanically, with roughly a quarter of lag. The market paid 7.47 percent on Thursday for a revenue line that Thursday own selling directly reduces. Management added the second problem itself, flagging a modest softening in the initial public offering market on the call, and listings revenue depends on new issuance that an unsettled tape tends to postpone. At about 24.8 times the quarter it just annualised, this is the highest multiple in the issue, and it is the only one of the five whose fastest growing revenue line is levered to the index everything else here is rotating out of.
The Catalyst, With The Source On October 22, 2026 Nasdaq reported third quarter net revenue of 1.481 billion dollars, up 13 percent on a reported basis and 14 percent organically. Solutions revenue was 1.172 billion dollars, up 17 percent reported and 18 percent organically; Market Services net revenue was 309 million dollars, up 2 percent; index revenue was 286 million dollars, up 39 percent. Annualised recurring revenue reached 3.365 billion dollars, up 12 percent reported and 14 percent organically. GAAP diluted earnings were 0.96 dollars a share, up 31 percent, and non-GAAP diluted earnings 1.04 dollars, up 18 percent, on a 57 percent non-GAAP operating margin. The company returned 173 million dollars in dividends and 323 million dollars in buybacks and updated 2026 non-GAAP operating expense guidance to 2.550 to 2.580 billion dollars. Source: U.S. Securities and Exchange Commission, Nasdaq, Inc. Form 8-K, third quarter 2026 earnings release, October 22, 2026. Read it yourself. What To Watch Next October trading volumes, expected in the first week of November. This is the only genuinely recurring scheduled disclosure among the five picks: the company published September volumes on October 7, August on September 3 and July on August 5. Confirms the thesis: Volumes up sharply on the October selling, which would show the Market Services line benefiting from exactly the volatility that threatens the index line. Breaks it: Soft volumes alongside a falling index, because then both halves of the business are pointing the same way and the 39 percent index growth loses its hedge.
04 · Materials / Commodity Chemicals · +7.25%
DOW Dow Inc.
On October 22 Dow reported third quarter net sales of 11.3 billion dollars, up 13 percent, and operating earnings of 0.71 dollars a share against a LOSS of 0.19 dollars a year earlier. Operating EBIT came in at 887 million dollars, up 707 million dollars year over year, which means the prior year base was around 180 million. Packaging and Specialty Plastics, 70 percent of the business, grew net sales 18 percent to 5.778 billion dollars and expanded its operating EBIT margin to 10.1 percent from 4.1. Free cash flow was 1.1 billion dollars, the company released more than 400 million dollars of working capital and has cut debt by roughly 600 million dollars this year. The shares rose 4.71 percent on the day and 7.25 percent on the week.
Why it made the five: It made the five because it is the largest earnings swing in the issue and because the swing came from decisions rather than from waiting. The 707 million dollar move in operating EBIT is the biggest single year over year change any pick here reported, and management credits its own self help: a 1 billion dollar cost programme, 700 million dollars of Transform to Outperform benefits landing in 2026, European asset shutdowns building through 2027, and about 2 billion dollars of total uplift targeted by the end of next year. A cyclical that earns its way back while the cycle is still ordinary is a different animal from one that waits for demand, and the cash confirms it rather than just the accounting: 1.839 billion dollars of operating cash flow, up 709 million, and 84 percent cash conversion over twelve months.
The case against it: Read the composition of the 13 percent and the pick gets much less comfortable. Local price rose 14 percent and VOLUME FELL 1 percent. In the segment doing the work, Packaging and Specialty Plastics, price rose 17 percent and volume rose 1. So the entire earnings inflection is price, and end demand did not improve at all. Price is the most mean reverting variable in the most cyclical industry in this issue, and a margin that went from 4.1 percent to 10.1 percent on spread can give it all back on spread. The 11.2 times this trades at against the quarter annualised is the cheapest number in this issue and the least trustworthy, because annualising the steepest point of a recovery is a statement about where in the cycle you took the earnings, not a valuation. Run the same arithmetic on last year quarter and the multiple is negative.
The Catalyst, With The Source On October 22, 2026 Dow reported third quarter net sales of 11.3 billion dollars, up 13 percent, composed of local price up 14 percent, volume down 1 percent and currency flat. GAAP net income was 525 million dollars and GAAP earnings 0.62 dollars a share, with operating earnings of 0.71 dollars a share against a loss of 0.19 dollars a year earlier. Operating EBIT was 887 million dollars, up 707 million dollars year over year. Packaging and Specialty Plastics net sales were 5.778 billion dollars, up 18 percent, with operating EBIT of 585 million dollars, up 386 million dollars. Cash provided by operating activities from continuing operations was 1.839 billion dollars, up 709 million, and free cash flow was 1.1 billion dollars. The company released more than 400 million dollars of working capital in the quarter, paid 253 million dollars of dividends and has reduced debt by about 600 million dollars year to date. Source: U.S. Securities and Exchange Commission, Dow Inc. Form 8-K, third quarter 2026 earnings release and financial schedules, October 22, 2026. Read it yourself. What To Watch Next No company event is scheduled; fourth quarter results are expected in late January 2027 on the company own cadence. The live variable in the meantime is the oil complex, with Brent crude at a five month high near 112 dollars on Middle East supply risk, because crude sets both the feedstock cost and the product price in different parts of this chain. Confirms the thesis: Polyethylene pricing holding while the cost programmes keep landing, which is the only path by which 887 million dollars of operating EBIT becomes a level rather than a peak. Breaks it: Price giving way while volume stays negative, because then nothing is left holding the margin up and the 707 million dollar swing starts running backwards.
05 · Consumer Staples / Household Products · +5.42%
PG The Procter and Gamble Company
On October 22 Procter and Gamble reported first quarter fiscal 2027 net sales of 23.3 billion dollars, up 4.0 percent, with organic sales also up 4.0 percent. The composition is the point: 2.0 points of that came from VOLUME, 0.4 points from pricing and 1.1 from mix. All five reporting segments grew organically, led by Beauty at 5.4 percent and Fabric and Home Care at 4.8. Core earnings were 2.00 dollars a share, up 1 percent. The shares rose 5.58 percent that day and 5.42 percent on the week.
Why it made the five: It made the five because of where the growth came from. For most of two years the large staples companies have grown by charging more, and volume growth of 2.0 points means households bought more units, which is a far better signal about a consumer business than the same growth taken in price. The breadth inside the quarter supports it: all five segments positive, not one segment carrying an average. And this is the one pick where I will concede the rotation did part of the work, because staples trade partly as bond substitutes and long yields fell 14 basis points. But the sector rose 3.82 percent across the whole week while this rose 5.58 percent on its report day alone, so the company specific piece is the larger one.
The case against it: The company printed 4.0 percent organic growth against a full year guide of 1 to 3 percent and did not raise the guide, and the reason is sitting on the margin line. Core gross margin fell 150 basis points and core operating margin fell 80, in the same quarter sales grew 4 percent. Management is carrying a 1.4 billion dollar after tax commodity cost headwind and 0.79 dollars a share of combined headwinds in fiscal 2027, and Brent crude closed the week near 112 dollars, a five month high, on Houthi attacks and reports of renewed strikes against Iran. Input costs are the problem already visible in the numbers and the input cost just went up. That matters more because of the price: at about 23.7 times the midpoint of its own 6.89 to 7.11 dollar core earnings range, this is the most expensive pick in the issue relative to the growth it is guiding to, and the guide is in line to three percent.
The Catalyst, With The Source On October 22, 2026 Procter and Gamble reported first quarter fiscal 2027 net sales of 23.3 billion dollars, up 4.0 percent, with organic sales up 4.0 percent composed of 2.0 points of volume, 0.4 points of pricing and 1.1 points of mix. Diluted earnings were 1.97 dollars a share, up 1 percent, and core earnings 2.00 dollars, up 1 percent, with core gross margin down 150 basis points and core operating margin down 80 basis points. All five reporting segments grew organically: Beauty 5.4 percent, Fabric and Home Care 4.8 percent, Grooming 4.5 percent, Baby Feminine and Family Care 3.0 percent and Health Care 2.7 percent. The company maintained fiscal 2027 guidance of all-in sales growth of two to four percent, organic sales growth of one to three percent and core earnings per share growth of in line to three percent, equal to 6.89 to 7.11 dollars, including a 1.4 billion dollar after tax commodity cost headwind. Source: The Procter and Gamble Company, P and G Announces Fiscal Year 2027 First Quarter Results, October 22, 2026. Read it yourself. What To Watch Next Oct 29: The September Personal Income and Outlays report, released at 8:30 in the morning alongside third quarter gross domestic product, per the Bureau of Economic Analysis schedule. Confirms the thesis: Soft goods inflation, which would mean the 150 basis point gross margin squeeze is nearer its end than its beginning. Breaks it: Hot goods inflation with crude still near 112 dollars, because the 1.4 billion dollar headwind would be widening rather than peaking and a 1 to 3 percent earnings guide has very little room in it.
What Connects All Five, And Why That Worries Me
Those are five sectors and five unrelated businesses. Enterprise software, commercial door hardware, index licensing, polyethylene and detergent share essentially nothing. On paper this is the most diversified set of five I have published.
On paper. All five reported on Thursday October 22 and all five made the majority of their weekly gain in that one session: Roper Technologies plus 9.73 percent, Allegion plus 11.02, Nasdaq plus 7.47, Procter and Gamble plus 5.58, Dow plus 4.71. On the same session the index fell 1.21 percent and the semiconductor index fell 6.72. Five companies in five industries do not move together on their own fundamentals. They move together because one flow chose them on one afternoon.
Here is the uncomfortable version, and I would rather say it than have you work it out. The earnings explain which five names, because each cleared a dated catalyst screen that 38 other candidates failed. The earnings do not explain the size of the response. A 4.7 to 11.0 percent single session move is large for a quarterly beat, and the biggest flow of the week was leaving the artificial intelligence complex on exactly that session looking for somewhere to go. Part of what you are reading is research and part of it is the market needing a destination for one afternoon, and I cannot separate them with the evidence I have.
The further problem is what comes next. Four of these five have no scheduled company catalyst for roughly three months. If the flow turns, the five sectors protect nobody, because the sectors are not what connects them. That is why the biggest risk I published above is about the reversal of that flow and not about any of these companies.
Stock of the Week
ROP · ROPER TECHNOLOGIES, INC. · +14.06%
This week the market drew a line between artificial intelligence you spend on and artificial intelligence you get paid for, and Roper Technologies is the clearest example of the second kind. It is a collection of software businesses selling into healthcare, insurance, education and construction, it reported 7 percent organic growth and raised its guidance, and it rose 9.73 percent on October 22 while its own sector fund fell 3.29 percent on the same day. A 13 point gap inside one sector in one session is not a rotation; it is a market pricing one company differently from its neighbours. I am naming it my strongest idea and attaching two pieces of discomfort rather than one. The adjusted earnings per share grew 15 percent on adjusted net earnings that grew 5, so most of what the market reacted to came from buying back 9 percent of the shares. And 3.84 of its 14.06 percent arrived two sessions before the numbers, with no company announcement I could find for that date. The business is good, the quarter was good, and the headline number flatters both.
What Could Change My Mind?
If the hyperscalers guide capital spending down. My whole read is that the market is refusing to pay for spending plans. If Microsoft, Alphabet or Amazon trims a plan on October 28 or 29, the complaint is answered, the complex that fell 6 to 12 percent this week rallies, and the rotation that paid for all five of these picks goes into reverse. I put 30 percent on something like that and named the reading that would show it: the technology fund back above 191 by November 5.
If long yields go back up. Four of these five got help from the 14 basis point fall at the long end. The Federal Reserve decides on October 28 and the September inflation print lands on October 29, with inflation above the 2 percent target for close to six years and the market carrying a small but real probability of a hike rather than a cut. A 30 year yield back above 5.45 percent would undo most of that help in a week.
If Dow volume stays negative. I took a commodity chemical company on a 707 million dollar swing in operating earnings, and the sales growth behind it was price up 14 percent with volume down 1. If the next quarter shows price fading while volume is still negative, the thesis was a spread trade I described as a turnaround, and I would rather find that out from the volume line than from the share price.
If my own screen is costing me more than it saves. Eight names I rejected this week rose more than my smallest pick, several of them on quarters that were genuinely excellent but reported before the window opened. My rejection record says the no dated catalyst screen is the strongest filter I have, 100 percent precision on the names it has caught. If that stops being true, the discipline is costing readers real returns and I will say so here rather than quietly loosening it.
Scoreboard
This is the part I would skip if I were trying to look good. Issue No. 17's five picks, measured over the week after publication from the Friday close a reader could actually have bought at, against the S&P 500 at minus 1.01 percent.
| Issue No. 17 | Issue Week | Since Oct 15 | vs S&P |
|---|
| KMB Kimberly-Clark | +6.48% | +7.93% | +8.94 | | UNH UnitedHealth Group | +9.63% | +4.50% | +5.51 | | REGN Regeneron Pharmaceuticals | +7.68% | +2.86% | +3.87 | | BLK BlackRock | +5.20% | +1.47% | +2.48 | | CCI Crown Castle | +16.13% | -1.46% | -0.45 |
| Average | +9.02% | +3.06% | +4.07 |
Four Of Five Beat The Index, And The One That Did Not Was The One I Flagged Four of the five beat the S&P 500 over the following week and the average beat it by 4.07 points, which is the best forward week I have recorded. The only one that lagged was Crown Castle at minus 1.46 percent, and that is the pick I opened Issue No. 17 by calling the one I was least sure of, because the market had paid 16 percent for a tenant that had signed nothing. Flagging the weakest name and then watching it be the weakest name is worth more to you than the four that worked. My forward call that week survived and its reasoning did not, which is graded in full up in The Week Ahead. I am not going to let a good forward week paper over that. What The Record Actually Says, Including The Unflattering Half Across 85 picks and 17 issues the average issue week return is 13.14 percent against 0.20 percent for the S&P 500, with an 88.2 percent win rate. That number is much less impressive than it looks and you should know why: the picks are selected after they have already moved, so the issue week return is partly a description of the week I chose them from. The number that is actually about you is the forward one. Over 75 picks and 15 issues, measured from the Friday a reader could buy, the picks average 0.87 percent against 0.39 percent for the index. That is an edge of 0.48 percentage points, winning 62.7 percent of the time and beating the benchmark 56 percent of the time. Real, modest, and nothing like 13 percent. I publish both because only one of them is a claim about my research. Terms in This Issue
- Diluted Share Count
- The number of shares a company profit gets divided among, counting the shares that would exist if every option and convertible were exercised. Earnings per share can rise either because profit went up or because this number came down, and the headline does not tell you which.
- Organic Revenue Growth
- Revenue growth stripped of acquisitions, disposals and currency moves, so it measures what the business already owned actually did. It is the number to look at when a company has been buying other companies, because reported growth includes whatever it just bought.
- Annualised Recurring Revenue
- The yearly value of the subscription contracts a company currently has in place, as if they all ran for a full year. It tells you how much of next year revenue is already committed rather than needing to be won again.
- Operating EBIT
- Earnings before interest and tax, measured on continuing operations and excluding items a company treats as one off. For a cyclical business it is the cleanest view of whether the core operation is making money, because it strips out financing and tax effects that move for unrelated reasons.
- Core Earnings Per Share
- A company own measure of per share profit with items it considers unrepresentative removed, such as restructuring charges or one off tax effects. It is useful for comparing one year to the next and it is defined by the company rather than by an accounting rule, so it is worth checking what was excluded.
- Non Residential Construction
- Building work that is not housing: offices, schools, hospitals, warehouses and data centres. It runs on long lead times, because a building gets designed and specified months or years before the materials and hardware actually ship, so it tells you about decisions already taken rather than about current demand.
A Possible Next Step
This issue tells you what moved, why, what the case against each name is, and exactly what would prove my forward call wrong. It deliberately stops short of entry zones, invalidation levels, sizing tiers and exit frameworks.
I'm testing interest in a future Edge Report with entry zones, invalidation levels, sizing tiers, and exit frameworks. Nothing is for sale and there are no premium subscribers. I am trying to find out whether an executable version of this research is worth building, and the only way to know is to ask.
|