Not more prose. The decision layer: how the view is expressed, how big, what it assumes, and the number that closes it. Below is the paid half of a weekday note, section by section, with real excerpts from a published edition.
Every weekday note opens on the view, one claim per bullet, and under every claim the thing that kills it. This is free, on principle, because a reader should be able to check the view against the tape without paying. From the September 9, 2026 edition:
THE BARREL. Brent 100.58 is being paid in crude and refused everywhere else. Long the premium at a quarter unit, no more. I'M WRONG IF energy equities lead the tape.
THE FRONT END. Third consecutive front-end-led close, this one with no policy input at all. Short duration at 1.5 units into it. I'M WRONG IF the 30Y rallies 15bp on today's operations.
FINANCIALS. Tuesday's 256bp of subsector dispersion sorts by liability duration, not by margin. Own the cohort ratio, not the sector. I'M WRONG IF high yield clears 290bp with the 10Y still rising.
THE ADMISSION. Six factors on the sheet and 2.00 of 6.50 units sit in one combined inflation bet. Concentration, stated. I'M WRONG IF the 2Y breaks 4.25 and Brent settles below 95.
One table, one row per live line. The factor column is what makes the netting checkable by the reader rather than assertable by the writer; the assumption column names what the kill is actually testing, which is the part that usually breaks. Entry, mark, target and stop live on the consolidated sheet and the note says so. The first row from the same edition:
| # | EXPRESSION | FACTOR | SIZE | STATUS | KEY ASSUMPTION | I'M WRONG IF |
|---|---|---|---|---|---|---|
| 1 | Duration short, 30Y and TLT puts | Long end | 1.50 | Add executed at half on the 8 Sep close | Dominant supply eventually makes the long end lead, and three failed tests are a sequencing problem rather than a refutation | 30Y rallies 15bp on today's operation tape |
| Thirteen more rows in the note, then the netting paragraph, in units, ending on the concentration admission. | ||||||
Then the netting, written out. Which legs monetize the same factor, how many units that really is across how many tickets, what is uncorrelated, what hedges the rest. It ends on an admission, in the note's own words, of how many bets the sheet really is. On September 9 the answer was six factors, and 2.00 of 6.50 units in one combined inflation bet. A sheet that nets to one bet and presents as four is a failure, and the note says so when it happens.
The home sector. Banks, regionals, investment banks and brokerage, exchanges and data, insurers, life, BDCs, consumer finance, payments, fintech and mortgage REITs do not trade together, and the edge is in saying which is which. At least one level or spread every day. It ends on a positioning read, not a recap, including on the days nobody else is writing about financials. Especially those days.
Positioning, vol, skew, dealer gamma, fund flows. What the crowd owns and what it would have to do if it is wrong.
Four branches, A through D, with weights that sum to a hundred, labeled as subjective rather than market-implied, and each one carrying its change against the prior run in tens. A static tree is a forecast. A moving one is a decision framework, and the movement is the information. Each branch says what happens to the sheet, position by position, not just what happens to the market.
The forward section. Every bullet carries a level or a date, because a risk invoked with no number is a mood rather than a watch item. The note ends on the last one. No hook, no sign-off, no curtain call.
Weekly Musings walks every sector and marks every live line against its stop as distance to trigger, one table and one chart, with last week's call graded against Friday's close. Paid readers get it every Saturday. Everyone gets the first Saturday of the month, which also carries the monthly scorecard, losers first, and a section titled THINGS I GOT WRONG, because a visible feedback loop is the difference between a decision process and a prediction-selling machine.
The ten-section Mega Morning Brief, PDF and HTML, every weekday before the open. There is a dated sample on this site.
$45 a month, $450 a year. The price of one bad lunch a month.