Two Saturdays ago I sat the September letter down and marked it: five confirmed, two pending, none wrong — and the least useful kind of right, because every reason had arrived and the thing they were supposed to cause hadn’t. A week later it still hasn’t. The S&P 500 closed Friday at 7,743.41, 0.71% from its Aug. 13 record, and it got there through what was supposed to be the weakest fortnight of the year, up 1.2% since expiry. Meanwhile the reasons kept arriving. On Wednesday the 10-year Treasury had its biggest one-day jump since the April 2025 tariff shock; it closed the week at 5.17%, a level last seen in 2007. Jobless claims fell to 197,000. The 30-year mortgage hit 7.03%. Heating oil in New York is $6.14, up 66% in a year.
So the question a client put to me this week is the right one, and it has three parts. Are we on path? How are we positioned, exactly — if the drop comes now, or on Oct. 2, or Oct. 14, or the day after the Fed? And what are all the things that can go wrong: not the ones that fit the thesis, all of them? This letter is the answer, in that order. The meter above is the summary. The rest is the working.
1. On path: the ledger, three weeks on
A thesis without dates is an opinion. The September letter had nine dates and six reasons, and each can be checked against a newspaper. Here is the ledger, with the two weeks since the last grading added.
| What the letter said | What happened | Grade |
|---|---|---|
| Sept. 11 inflation: hot confirms the income sleeve; cool, and “my 60 drops toward 45.” Named as the specific way the thesis would be wrong. | Core +0.3% against the 0.1% test. The exit closed. | Confirmed |
| Sept. 16 Fed: a hike would be the first this year — the trigger the bear had waited on since March. | Raised to 3.75–4%. Sixteen of 18 officials see at least one more this year, per Thursday’s paper; Governor Michael Barr says more increases are likely needed. | Confirmed |
| The Fed is not coming to help — the one reason the letter would keep if it had to throw away the other five. | Our Friday pull of Polymarket: about 63% on a quarter-point increase on Oct. 28, 36.5% on no change, under 1% on a cut. | Confirmed |
| The bond market is voting: a 10-year toward 5%, the term premium rebuilt, the foreign buyer base thinning. | Wednesday brought the biggest one-day rise in the 10-year since the April 2025 tariff shock, to 5.113%, on a 58.4 flash PMI, Barr, Iran at the U.N. and a weak five-year auction. Friday: 10-year 5.17%, 2-year 4.81%, 30-year 5.49%. It just “doesn’t make sense to a lot of people to own bonds here,” a credit-union investment chief told the paper. | Confirmed — the number the letter feared is the number on the screen |
| The fuel and the consumer: diesel, wages behind prices, the average stock underneath the index. | Heating oil $6.14 in New York, up 66% in a year; diesel above $6.50; the 30-year mortgage hit 7.03%; the winter forecast raised from +31% to +43%. And jobless claims fell to 197,000: the consumer is strained and employed at the same time. | Confirmed — with a twist: the economy is too strong for its own bonds |
| Sept. 18 expiry: the two weeks after are on average the weakest fortnight of the year. | Week one: the S&P 500 rose 1.2%. The fortnight ends Friday, Oct. 2. | Pending — running against us |
| The drop: 6–10% from the Aug. 13 high, the low between late September and mid-October, bought into the vote. | 0.71% from the record. VIX 14.87. Nothing has moved. | Pending |
| The bear’s veto: no hedges while the curve stays in contango and vol-of-vol under 90. | Measured this week, not assumed: 12.76 / 14.87 / 17.93, vol-of-vol 87.84 (it was 90.57 on Thursday). No hedge bought. | Honored |
| Nothing new into a binary; Micron at no more than 1% waits for the print. | Nothing bought in six sessions; the conditions pass one of three. Micron reports Wednesday. | Honored |
| Valero: the trim signal is a White House move toward restricting diesel exports — the policy, not the price. | The President backed an export ban on Tuesday; the Energy Secretary says any limits would be voluntary, details soon; nothing published or ordered. Held at weight; if a plan is published, the trim decision gets made in writing. | In play — honored as written |
| What would change my mind: a record close above 7,798.99 before Oct. 2; a 0.1% core print on Oct. 14; a Fed on Oct. 28 that says one was enough. | One week left on the first; the second and third are on the calendar. | Open |
Eight confirmed or honored, three open, none wrong — the same uncomfortable shape as two weeks ago, only sharper. A thesis that is right about every cause and hasn’t produced the effect is either early or wrong, and the market still hasn’t said which. What it has done is take away the cushion. On Sept. 18 the Journal’s weekly data put the index at 20.1 times forward earnings — a 4.98% earnings yield against a 4.995% ten-year, a premium of zero. Hold the estimates where they were, move the index up 1.2% and the 10-year to 5.17%, and by my arithmetic the earnings yield is about 4.9% against 5.17%. Negative. A lender to the Treasury is now paid more than an owner of the S&P 500, with a contract instead of a hope. Markets have run this way before — third-quarter profit growth near 29% is the argument that they can — but every dollar from here has to be earned, on time, because there is nothing left to absorb a miss.
What I take from three weeks of being right about the causes: the plan was never a bet that the index would fall on a date. It was a bet that the dates would arrive and that we would still be solvent, paid and unforced when they did. On that reading it is on path. On the reading that matters to a scoreboard — has it paid yet — it hasn’t, and the fortnight it was written for is half over, going the wrong way.
2. How we’re positioned, sleeve by sleeve
This is the part a client actually wants on a Saturday, so here it is from Friday’s bake, which prices every book at the close. Four books tell the whole story: the flagship dividend book most clients hold, the North Star, and the two aggressive books at the far end.
| Book | YTD | Bills & cash | Energy | The build-out | Gold | Volatility | Worst drawdown | Since |
|---|---|---|---|---|---|---|---|---|
| Midterm Dividend $100K — the flagship | +9.0% | 13.6% | 5.5% | 2.0% (Microsoft) | 1.7% | 10.2% | −10.0% | Sept. 2021 |
| North Star 250 | +20.7% | 14.9% | 13.1% | 17.3% | 5.2% | 12.1% | −12.5% | Mar. 2024 |
| Aggressive 250 | +30.4% | 1.5% | 10.4% | 9.5% | 2.6% | 17.2% | −18.4% | Mar. 2024 |
| Aggressive 250 · Tactical | +28.3% | 1.5% | 12.0% | 30.2% | 2.8% | 20.0% | −21.0% | Mar. 2024 |
| S&P 500 (SPY) | +14.0% | — | — | the index is the bet | — | — | — | — |
Volatility is annualized from daily closes since inception; the drawdown is the worst peak-to-trough close in that span. “The build-out” counts Microsoft, Meta, Alphabet, Nvidia, Micron, Taiwan Semiconductor, Broadcom, Apple, Amazon, Oracle and Vertiv where held. Weights sum to 99.5%; the half-point is the operating reserve.
The reserve is the plan’s whole argument, and it is where the letter said it would be: Treasury bills and floating-rate paper — the iShares 0–3 Month Treasury Bond ETF (SGOV) and WisdomTree’s floating-rate Treasury fund (USFR) — about 15% of the North Star and Midterm Dividend books. A three-month bill yielded 4.24% at Friday’s close; the average bank money-market account, per Bankrate’s table in Friday’s paper, pays 0.44%. If October brings the hike the crowd expects, the floaters reset up. The reserve isn’t waiting for free. It is paid four percent to wait, which is the difference between a plan and a hunch.
The energy sleeve — Exxon Mobil (XOM), Chevron (CVX), ConocoPhillips (COP), Cheniere Energy (LNG) and, in the two aggressive books that hold it, Valero Energy (VLO) at 1.5% — is held, not chased, at 5.5% to 13% by book. Crude fell 2.3% Friday to $92.44 and most of the sleeve fell with it; that is fine. It isn’t there for the week. It is there because a war with a patient supplier behind one side — Friday’s front page counted about 1,300 Chinese shipments to Iran’s defense ministry through June — keeps the fuel line in every household budget high, and the dividends arrive in the same season as the heating-oil invoice. The crowd puts 0.75% on Hormuz traffic normalizing by Sept. 30. Valero’s written tripwire is in play and it stays at weight until a plan exists, which is the rule working, not the rule being ignored.
The build-out is owned at weight through the spenders and the makers, and the weight is the decision: 2% in the dividend book, 17% in the North Star, 30% in the tactical book — which is the one that would hurt most on a bad week, and is labeled that way on purpose. Nothing was added into a $10.3 trillion bet that Thursday’s paper says is increasingly financed with debt; that same debt is one of the reasons the 10-year sits where it does. Micron Technology (MU) reports Wednesday. It was the print the letter said the memory add waits for, and it still waits: a good print puts the name on the reserve’s shopping list, it doesn’t hand the reserve a shopping bag.
The core positions — Apple (AAPL), Amgen (AMGN), the gold fund (IAU), Berkshire Hathaway (BRK.B) through its chairman’s transition, CrowdStrike (CRWD) at its reduced weight — are held. The equal-weight index (RSP) is first on the list at −8%, because the concentration we can’t avoid inside the index we can at least offset beside it. Watched, not traded: Oracle, Vertiv, Constellation, NRG. No long nominal duration in any book, anywhere. No hedges while the veto stands. And the only thing in the books that resembles a target is a written maximum for the theme, not a price for a stock.
For a household the four checks are unchanged, and each is worth more than it was: how much duration sits in the bond fund you didn’t choose, inside the target-date account, with the long-Treasury index bleeding at a 5%-plus yield; twelve months of expenses in cash at four percent rather than 0.44%; the growth fund that turned out to be a momentum fund; and the beneficiary pages. None of them needs the meter to be right.
3. Five dates, and what we do on each
Here is the calendar from here to the vote, the way the letter promised: each date, each branch, and the move already written for it. The only forecast in this section is that these days will arrive.
| Date | If it goes this way… | …then this, on the rule | And if it goes the other way… | …then this |
|---|---|---|---|---|
| Wed · Sept. 30 Micron reports | Beats and raises: the memory shortage goes on the record with a number attached. | Micron goes on the reserve’s shopping list at no more than 1%, to be bought at the −8% trigger or the week of the vote — not on the print. A good print is the test the thesis wanted, not a permission slip for new money. | Misses, or cuts its outlook. | The build-out’s first crack. The guidance rule applies to the name that day in the books that hold it, and the sleeve’s written maximum is re-checked, not raised. |
| Fri · Oct. 2 September jobs; the falsifier expires | Hot — claims at 197,000 say strong: more hikes priced, the 10-year higher, growth multiples squeezed. | Nothing to do. The reserve is short paper that resets up, and there is no duration to lose. If the index also closes above 7,798.99, the fall caution was wrong on timing; I say so in print, and the week-of-the-vote purchase stands. | Cold, or negative — the first growth scare of the cycle. | The drawdown can finally start. A third of the reserve goes to work at 7,175 (−8%), on the rule. A cold jobs print beside a hot inflation print is the stagflation book’s week, and the dividend book’s. |
| Week of Oct. 12 September inflation (Oct. 14); the banks open earnings | Core 0.3% or hotter: strike three. Two hikes get priced; the 10-year heads toward 5.5%. | No duration in any account the owner didn’t choose on purpose — already true. The reserve stays; nothing new. | Core 0.1% or less. | The exit the Sept. 11 print closed reopens: the tactical adds come back on, Micron first; the no-drop case goes back above 40 and the meter under 4. The banks either way: deposit costs (the cash-is-king fight), card losses, private-credit marks and AI-bond exposure are the lines to read. Any name in the books that cuts its outlook gets sized down that day, on the rule, not in a meeting. |
| Tue–Wed · Oct. 27–28 The Fed; the largest tech companies report | A hike, as about 63% expect: priced. The dividend book earns its keep the same afternoon. | Nothing new. If two more are priced by then, a 5%-plus ten-year is the base case — which it already is. | No hike, and a chairman who says one was enough. | The relief rally: the no-drop world. The plan doesn’t need the call to be right; the last third of the reserve starts spending the week of the vote regardless. Guidance from Microsoft, Meta, Alphabet and Apple that week runs through the same rule. |
| Tue · Nov. 3 The vote | Either way. | The reserve starts spending the week of the vote, drop or no drop: the twelve months after every midterm vote since 1950 have been positive, averaging about +18.8%. Anthropic’s $100 billion offering, moved to November, is the liquidity drain at the far edge of the window and the reason the last third isn’t spent early. | — | The household dates ride alongside: Medicare open enrollment from Oct. 15, the prescription-import rule on Oct. 22, Opportunity Zone deferrals ending Dec. 31. |
Notice what the table doesn’t contain: a move that depends on knowing the number in advance. Each branch has a rule, each rule was written in September, and the only judgment left for the day itself is reading which branch we are on. That is the point of a dated plan. It converts the hardest question in investing — what happens next — into the easiest one: what did we say we would do if it did?
4. Everything that can go wrong
A risk you can name has a defense; a risk you can only feel has a therapist. Here is the full list — the Risk Atlas’s eight, re-graded for the third time since July, plus the six specific to the next six weeks — with what covers each one and, where the honest answer is “not much,” that.
| # | What can go wrong | What it would look like | Assigned to | Coverage |
|---|---|---|---|---|
| 1 | The falsifier prints | A record close above 7,798.99 by Oct. 2: the cohort carries the index through the weakest fortnight; the fall caution was wrong on timing. | The reserve earned 4.2% waiting; the Nov. 3 purchase stands; I say so in print. | Covered — by design; the cost is an index I didn’t chase |
| 2 | The drop comes and doesn’t stop | The bear case, 20%: −12% comes and goes, and the largest stock sale in history lands in the week of the vote. | Two written triggers at 7,175 and 6,863; the dividend book’s 0.37 beta; the last third held for the vote. | Partial — below −12% nothing is written except Nov. 3; the reserve is 15%, not 50% |
| 3 | Rates | A failed auction, a 5.5% ten-year, a negative equity premium that widens: multiples compress with no cushion, the long-Treasury index bleeds, target-date bond sleeves lose. | No duration anywhere; bills and floaters reset with the Fed. | Thin by design — the price risk is covered because we own no duration; the multiple risk lands on the equity sleeves |
| 4 | Inflation re-accelerates | Energy into core, wages behind for a sixth month, two hikes: stagflation. The consumer names in the dividend book (12% consumer discretionary) feel it first. | The energy sleeve and dividends that grow; the stagflation theme book. | Covered |
| 5 | The strait stays shut — or opens | Shut: heating oil past $6.14 and a $2,500 winter; open: crude down 20% in a headline and the energy sleeve gives back a year. | Sized as a hedge (5–13% by book), not a bet; the December 2027 curve already expects a reopening. | Covered both ways — the hedge’s cost is the premium for the household bill |
| 6 | A diesel-export order | The tripwire fires: Valero trimmed on the rule in the two books that hold it; refiners repriced; European diesel spikes. | The written signal; 1.5% positions. | Covered — the rule is the whole point |
| 7 | China and Taiwan | An export regime, a blockade, a summit that curdles; the front page that hosted Xi also documented China arming Iran. | Taiwan Semiconductor sized at 1.5–3%; no position that can’t take a bad day. | Covered by size, not by hedge |
| 8 | Credit inside the AI trade, and beneath it | Debt-financed data centers ($10.3 trillion, much of it off balance sheet); private credit reaching for the card applicants a bank declined; $757 billion of apartment debt due through 2028. | The avoid list: nothing priced by its own manager, nothing that borrows to buy what it holds. JPMorgan is held, so its own October numbers are the test. | Covered, with one exposure named |
| 9 | Concentration | The index is the bet. Meta moved 11% up in a day and 3% down in another this week; the tactical book carries 30% in the build-out. | Sleeves; the equal-weight fund first on the reserve’s list; a 3.5% cap on any single name. | Partial — the offset is real and small next to the index’s own bet |
| 10 | Event risk | A Micron miss; a hack of an AI agent holding someone’s credentials (“negative for the entire space,” Truist’s analyst said); an agent that goes looking for answers in a government portal; Berkshire’s transition; a bank outlook cut; Anthropic’s November offering. | Position caps; the guidance rule; no single outcome large enough to change the year. | Covered |
| 11 | Plumbing and platforms | New York sued Polymarket; regulators are examining a million look-alike Kalshi trades; a stablecoin “reward” isn’t a deposit; a broker outage on a red day. | We quote the odds and don’t bank the venue; speculation small, separate, withdrawable; FDIC deposit or fund, known for every dollar of cash. | Covered |
| 12 | The volatility regime breaks | Vol-of-vol through 90, the curve into backwardation. A market with the tail bid and the body asleep moves as a gap, not a grind. | The veto lifts and hedges are considered; until then the reserve is the hedge. | Thin by design — we said we wouldn’t fight a 110-day regime with a put, and we haven’t |
| 13 | Policy | Election-year levers that are good for stocks in the short run and bad for bonds; tariffs; the October fiscal fights; the ACA income cliff; the pension veto; the Oct. 22 prescription rule. | Dated and nonpartisan: the household lines sit in the budget with their dates; the market lines run through the rates square. | Covered where it’s a date, thin where it’s a headline |
| 14 | Us | A rule broken on a red Tuesday; buying the 355-foot drive; a plan that’s right about the causes and wrong about the effect — a Fed that stops, a strait that opens, a drop that already happened underneath the index and never shows in the average. | Dates, in writing, graded in public; the North Star and Midterm Dividend books graded on the average stock, not the index. | The hard one — passing, six sessions in |
Six covered by things that pay cash on a schedule, three by rules written before the week they are needed, two thin by design, three that are me being wrong. The tally I’d flag is the last three, because they are the ones the market has been arguing for since March: the cohort keeps carrying the index, earnings grow 29%, and the drop that six new lows for every new high described two weeks ago never reaches the average. That is why the plan buys on dates and drawdowns rather than on my view of them, and why the books that matter most to clients are graded on the average stock, not the index.
Where the meter sits, and what it says to do
6.3 of 10. The reason at ten, the calendar at eight, the breadth carried at eight, the price at one, the hedging at two. Fuel loaded; still no spark; and the fortnight it was loaded for is half over, going the other way.
So: the reserve stays a reserve — a third at 7,175, a third at 6,863, a third the week of the vote regardless; the shopping list is the equal-weight index, then the power builders, then the memory names after Wednesday’s print. No hedges while the veto stands. No duration while the 10-year is above the earnings yield, which it now is by about a quarter point. The energy sleeve held, not chased; Valero held at weight until a plan exists. A name that cuts its guidance in October gets sized down that day. And on Oct. 2, if the index prints a record, I will write that the fall caution was wrong on the date — and the plan will still buy in November, because that was never the part that depended on me being right.
You don’t wait for the first drop to find the umbrella. You also don’t stand in the doorway for six weeks refusing to go out. The plan is what lets us keep walking. Bring a statement; the four checks take fifteen minutes, and none of them needs the weather to cooperate.
December can grade the rest.