Daily Macro US
pos-017, one day after we sold it
94.25c vs our 71.69c fill
The 10-year settled 4.79% - ONE basis point from the 4.80% touch. Today's bid is 91c with 271 shares behind it against ten shares at 82c yesterday. The gate has cost $29.27 so far, $42.91 if it resolves. Correct process, bad outcome, published at the top.
Cleveland September core PCE
3.49% YoY (first print)
ACCELERATING from August's 3.40%, on the Fed's own preferred measure, stamp 09/01. The single most important number in this letter and the strongest thing that has happened to pos-013 in a month.
10Y Treasury
4.79% (Sep 1 SETTLE)
Treasury par yield curve, primary source. Fifth consecutive higher close, up 4bp; 2-year 4.39% (+5bp), 30-year 5.27%. Vendors reported 4.77-4.81 intraday and the touch contract remains unresolved, so we do not claim a 4.80 print.
pos-010 gate fired twice in 26h
72.5c at 06:00 UTC
Trimmed half the remainder again at 71c for +$7.27, two points BELOW our own fair value of 73. Fix adopted: 72c stays put and re-arms only below 68c. A trigger without hysteresis is a liquidation schedule, not a gate.
ISM manufacturing / GDPNow
54.6% / 4.8%
August ISM expanded for an eighth straight month; the overall economy has grown for 22. GDPNow refreshed UP from 4.6% and the four-session carry flag comes OFF. Both are results, not previews.
Book total P&L
$757.14 (-$0.18)
On the most one-directionally hawkish session of the year the book moved eighteen cents, because the position that would have captured it was sold Tuesday on a gate. That is the cost of rules, stated in full.
Yesterday morning we sold pos-017 for a 334% gain because a written gate fired. By last night's close the 10-year had settled 4.79%, its fifth consecutive higher close, leaving ONE basis point to the 4.80% level that position was written on. The contract trades 94.25c this morning against the 71.69c we took, and today the bid is 91c with 271 shares behind it rather than ten. Our 151.5152 shares would realise $137.88 now. The gate has cost $29.27 so far and will have cost $42.91 if it resolves. We are leading with that, and we are not relitigating it. The rule was written when fair value was 60, published every day for a fortnight, and honoured on the morning it fired into a book that had ten shares on top of it. That is correct process producing a bad outcome, which is a different thing from a mistake, and a letter that only prints the gates that paid is not running a process at all. The same design question arrived from the other direction. pos-010's trim above 72c fired a SECOND time in twenty-six hours, printing 72.5c at 06:00 UTC, and we honoured it at the 71c bid for +$7.27. But a half-cent breach re-triggering the same liquidation two days running is not discipline, it is a trigger with no hysteresis, and left alone it would grind this position to nothing on noise. So the fix adopted today is a re-arm condition rather than a wider number: 72c stays exactly where it is and can only fire again after the leg prints below 68c. We are fixing the mechanism instead of moving the trigger, because moving a trigger on the day it becomes inconvenient is the pattern we criticised ourselves for on pos-011 last Saturday. And the marking method got refined again, at our own expense again. Yesterday we stopped marking at the mid because the mid was not transactable. This morning pos-011's top-of-book bid turned out to be TEN SHARES against the 73.5294 we hold, so from today every position is marked at the bid our own size actually clears, walked through the book. It lowers the open figure by thirty-two cents. Behind all of it the data ran one way and hard: Cleveland published its first SEPTEMBER core PCE nowcast at 3.49%, accelerating from August's 3.40%; ISM manufacturing printed 54.6%; GDPNow refreshed up to 4.8%; Brent settled $94.65, up 4.6%, after US strikes on Larak Island and an explicit threat to Kharg Island, through which roughly 90% of Iran's crude exports flow. And the book moved eighteen cents, because the position that would have captured all of it was sold on Tuesday.
Today's Market Moves
10Y Touches 4.8% Before 2027 (pos-017, CLOSED Tue)
87.5%→94.25%+6.75pp
No longer ours. Bid 91 / ask 97.4, and the 91c bid carries 271 shares where yesterday's 82c bid carried ten. Hourly high 97.25c. We took 71.69c on Tuesday morning. The honest accounting: $29.27 of foregone value at today's bid, $42.91 if it resolves. Published at the top of the letter rather than buried here.
Fed Rate Hike in 2026 (pos-010)
71.5%→71.5%0.0pp
Flat on the mid, but the hourly series printed 72.5c at 06:00 UTC and the 72c trim gate fired a SECOND time in 26 hours. Trimmed half the remainder at the 71c bid, which carried 9,092 shares - no walk, no slippage. Fix adopted: the gate re-arms only after a print below 68c.
Zero Fed Rate Cuts in 2026 (pos-013)
88.95%→88.8%-0.15pp
Down again on unambiguously good news, and that is twice in three sessions. We have now called this leg 'noise' three times and we are saying plainly that a fourth use would be evidence the market sees something we do not. Bid 88.8 for 85 shares, then 88.7 for 584. FV 91 -> 92 on the September core PCE nowcast.
Fed Rate End 2026 = 4.0% (pos-011)
38.75%→37.7%-1.05pp
THE TOP OF THIS BOOK IS TEN SHARES at 34.3c, then 67 at 34.2c, then a gap to 33.0c. Our 73.5294 shares clear at a blended 34.21c. This is yesterday's pos-017 lesson arriving in miniature one day later, and it is why the marking method changes again today.
Fed September 25bp increase
56.5%→59.5%+3.0pp
Bid 59 / ask 60 on $15.8M; no-change 40.5c. CME FedWatch reported 65-68% on Sep 1 and we use 68 while flagging the range - so the gap is 8.5 points, still under our 10-point bar and still a correlated expression of a view the book holds three ways. Refused for a third session.
10Y Touches 5.00% Before 2027 (diagnostic)
33.5%→43.0%+9.5pp
Bid 37 / ask 49. The retired barrier model says 50.8 on a 21.0bp barrier and 82 sessions. Gap 7.8pp, NARROWED from 11.1. Yesterday's reinstatement rule requires within 5 points for three consecutive sessions; session one fails at 7.8, so the model stays diagnostic-only. It is behaving better and it does not get its job back for one good day.
Screening Table
| # | Market | Expiry | Market Price | Fair Value | Gap (pp) | Direction | Volume | Confidence |
|---|---|---|---|---|---|---|---|---|
| 1 | Zero Fed Rate Cuts in 2026 | Dec 31 | 88.8% | 92% | +3.2pp | HOLD $82 YES - FV 91 -> 92 on a September core PCE nowcast that ACCELERATED to 3.49%; 90c trim 1.2c away and not fired | $$7.85M | 4/10 |
| 2 | Fed Rate Hike in 2026 (pos-010) | Dec 31 | 71.5% | 73% | +1.5pp | TRIMMED half again at 71c - the 72c gate fired twice in 26 hours; hysteresis rule added, re-arms only below 68c | $$8.42M | 4/10 |
| 3 | Fed Rate End 2026 = 4.0% | Dec 31 | 37.7% | 44% | +6.3pp | HOLD $25 YES - mechanical FV 44, but the ladder and the hike contract still disagree by 9.0 points and the top bid is ten shares | $$1.37M on the leg | 2/10 |
| 4 | US Unemployment 2026 >= 5.0% | Dec 31 | 7.55% | 7% | +0.9pp | HOLD $100 NO - review re-ran and concluded hold; FV 8 -> 7 on ISM 54.6 and GDPNow 4.8, so marked cheap after being rich | $$124.6K on the leg | 5/10 |
| 5 | Fed September 25bp increase | Sep 16 | 59.5% | 68% | +8.5pp | NO POSITION - 8.5pp against FedWatch, under our 10pp bar and a fourth correlated hawkish expression; refused for a third session | $$15.8M on the leg | 3/10 |
Market vs Fundamentals
Market Price (red) vs Estimated Fair Value (green) — %
Top 5 Opportunities
1
The gate that cost us $29.27 in a day, published before anything that worked — YES
↑ BUY YES0pp
Market price
94.25%
Fair value
0%
Gap: 0pp
On Tuesday morning the 66c-bid take-profit on pos-017 fired at an 82c bid and we sold all 151.5152 shares for a volume-weighted 71.69c, because the book had ten shares at 82, ten more at 81, and then a hole down to 70. Realised +$83.61 on $25. Yesterday the 10-year settled 4.79%, its fifth consecutive higher close, which leaves ONE basis point to the 4.80% touch this contract resolves on. It trades 94.25c this morning, it printed 97.25c overnight, and the 91c bid now carries 271 shares. Our position would realise $137.88 today. The gate has cost $29.27 against Tuesday's fill and will have cost $42.91 if the touch prints. We are publishing that at the top, above the two positions that are working, because the alternative is a letter that reports gates only when they pay. What we are NOT doing is relitigating the decision. The rule was written when fair value on this position was 60, it was published every single day for a fortnight, and it was honoured on the morning it fired into the thinnest book we had ever seen on it. Changing the rule now, having watched the outcome, would convert a process into a story about hindsight.
▵ Bull case
- The thesis was right and the exit was the rule working as designed. A take-profit that only gets honoured when the subsequent price action validates it is not a take-profit.
- We sold into a book with ten shares on top. Holding for another 24 hours to reach today's 271-share bid was not a decision available to us on Tuesday morning; it is only visible now.
- Closing removed the book's largest correlated bet on hawkish rates, which is diversification we had been unable to achieve any other way for three weeks.
- +$83.61 on a $25 stake is still the second-largest winner this book has produced, and a 334% return does not become a failure because a larger one existed.
▿ Bear case
- $29.27 at today's bid, $42.91 at resolution. On a $25 stake those are enormous numbers and no framing makes them smaller.
- The 66c trigger was set when fair value was 60 and never revisited as the thesis strengthened. A gate that does not move with fair value will always sell winners early, and we knew that and left it.
- We had a published fair value of 61 while the arithmetic said 76.8, frozen by a rule about a different market. Had the freeze not been in place, the gate would have looked obviously mispriced against our own number and we might have raised it.
- The barrier model we retired on Tuesday for failing five consistency checks was, on this leg, closer to right than the market was. Retiring it the same morning we sold is a coincidence we should sit with rather than dismiss.
2
Zero Fed Rate Cuts in 2026 - September core PCE is accelerating and the leg went down anyway — YES
↑ BUY YES+3.2pp
Market price
88.8%
Fair value
92%
Gap: +3.2pp
Cleveland published its first September core PCE nowcast this morning: 3.49% year-over-year, up from August's 3.40%. That is inflation accelerating into the autumn on the measure the FOMC actually targets, and it is the strongest single input this position has had in a month. Around it, ISM manufacturing printed 54.6% for August - an eighth consecutive month of manufacturing expansion and a 22nd of overall growth - GDPNow refreshed up to 4.8% for Q3, taking off a carry flag we had held for four sessions, Brent settled $94.65 up 4.6%, and FedWatch has a September hike at 68%. A committee facing 3.5% core inflation, close to 5% growth and $95 crude does not cut in the four months to December. Fair value goes 91 to 92. And the leg went DOWN 0.15. That is the second time in three sessions this position has fallen on news that should have lifted it, and we are done calling it noise. We have used that word three times about this market. On a $7.85M book one tick wide, a fourth occurrence should be treated as evidence that the market is pricing something we have not identified, and we are writing that down now so it cannot be rationalised later.
▵ Bull case
- Cleveland's first September core PCE nowcast is 3.49%, ACCELERATING from 3.40% in August. Inflation getting worse into the autumn removes the cutting case entirely.
- ISM manufacturing 54.6%, eighth straight month of expansion, 22nd month of overall economic growth. GDPNow refreshed UP to 4.8%.
- Brent settled $94.65, up 4.6%, with Kharg Island - roughly 90% of Iran's crude exports - under explicit threat. Energy is pushing inflation, not pulling it down.
- FedWatch 68% for a September HIKE. The distance from tightening in a fortnight to cutting before December is not travellable.
▿ Bear case
- It fell on all of that. Twice in three sessions now, and our explanation has been the same each time. That is a pattern we are flagging against ourselves rather than a coincidence.
- Real consumer spending was FLAT in July after +0.4%, and July payrolls FELL 23,000. The bear mechanism is intact and untouched by anything this week.
- Friday's August employment report has consensus at +60,000 with unemployment ticking to 4.2%. One genuinely bad print reopens the 2026 cut question, and it is two days away.
- At 88.8c the remaining upside is 11.2 cents on the book's largest stake, against a tail that a labour-market break makes expensive very quickly.
3
Fed Rate Hike in 2026 - the same gate fired twice in twenty-six hours — YES
↑ BUY YES+1.5pp
Market price
71.5%
Fair value
73%
Gap: +1.5pp
The 72c trim gate fired again, printing 72.5c at 06:00 UTC this morning, twenty-six hours after it fired at 75.5c on Monday afternoon. We caught this one within two hours rather than fifteen and sold 45.4545 shares - half the remainder - into a 71c bid carrying 9,092 shares, so no walk and no slippage. Proceeds $32.27 on $25 of stake, realised +$7.27. Now the problem, which we would rather name than wait for. A half-cent breach on a one-cent-wide market triggering a second liquidation of the same position on consecutive days is not discipline; it is a trigger with no hysteresis, and left alone it would grind this position to nothing on noise while calling it process. The fix adopted today is deliberately NOT a wider number: 72c stays exactly where it is, and it can only fire again after the leg prints below 68c. We are changing the mechanism rather than the level, because moving a level on the day it becomes inconvenient is precisely what we criticised ourselves for when we raised pos-011's review threshold on Saturday. Fair value goes 71 to 73, which means we sold at 71 against our own 73 - two points below, worse than yesterday's one point, and honoured for the same reason.
▵ Bull case
- Cleveland's September core PCE nowcast at 3.49% is accelerating inflation on the Fed's own measure, three weeks before the meeting.
- The 10-year settled 4.79%, a fifth consecutive higher close, with the 2-year up 5bp to 4.39%. The curve is pricing policy, not just term premium.
- FedWatch 68% for September against a 59.5c prediction market. Futures are more convinced than Polymarket for a third session running.
- ISM manufacturing 54.6% and GDPNow at 4.8%. There is no growth argument against tightening left.
▿ Bear case
- We have now sold two thirds of this position at 70 and 71 while our own fair value went from 67 to 71 to 73. The gate is systematically selling into strength and we have said so twice.
- Real consumer spending was FLAT in July and July payrolls FELL 23,000. Friday's employment report is the live risk and it is two days out.
- At 71.5c against a fair value of 73 the remaining $25 has 1.5 points of edge, which is inside most reasonable error bars on the fair value itself.
- The hysteresis fix is the right change, but we are making it the morning after the gate cost us something. The timing is defensible and it is still worth a reader's suspicion.
4
Fed Rate End 2026 = 4.0% - the top of the book is ten shares, one day after we learned that lesson — YES
↑ BUY YES+6.3pp
Market price
37.7%
Fair value
44%
Gap: +6.3pp
Yesterday we stopped marking this book at the mid because pos-017 taught us the mid is not transactable. This morning the same lesson arrived one layer down. The bid on this leg is 34.3c for TEN SHARES, then 34.2c for 67, then a gap to 33.0c. We hold 73.5294 shares, which clear at a blended 34.21c for $25.16 - not the $25.22 a top-of-book mark would claim. So the method refines again: from today every position is marked at the bid our own size actually clears, walked through the book. Across four positions it costs thirty-two cents, which is trivial in dollars and is the entire point - we found it by looking rather than by selling. On the position itself nothing improved. The ladder prices 3.5% 7.85, 3.75% 25.15, 4.0% 37.70, 4.25% 18.25 and 4.5%-or-more 6.55, a hike mass of 62.50 against 71.5 on the standalone hike contract. Nine points of disagreement between two markets on one event, narrowed from 9.3 yesterday, and our fair value of 44 is derived from the contract on the other side of that gap.
▵ Bull case
- The mechanical derivation from a deep contract gives 44 against a 37.7 market, and it is the same method that has been right on direction for three weeks.
- If the Fed hikes once in September and stops, this is the bucket that pays. FedWatch at 68% makes that the modal path.
- $25 of stake at 34c pays $73.53 if correct. The asymmetry survives a wide spread even after the marking refinement.
- Cleveland's September core PCE at 3.49% supports one hike far better than it supports none.
▿ Bear case
- A ten-share top of book on a leg we hold 73.5294 shares of. If we ever needed to exit quickly the realisable price is well below anything we publish.
- The 9.0-point disagreement between the ladder and the hike contract is doubt attached directly to our own fair value, because we compute ours from one side of it.
- Marked at the size-aware bid this position has made sixteen cents in three months on $25. That is not an edge.
- Accelerating inflation argues for TWO hikes as easily as one, and two hikes pay this bucket nothing.
5
US Unemployment 2026 >= 5.0% - the review re-ran, as promised, and the answer did not change — NO
↑ BUY YES+0.9pp
Market price
7.55%
Fair value
7%
Gap: +0.9pp
Yesterday we wrote that the sub-8c review would re-run every session the leg stayed below the line, rather than being conducted once and quietly forgotten. It is still below, so here it is again. The answer is unchanged and the arithmetic barely moved: 133.3333 shares of NO realise $122.30 today walking the book at 91.9c and 91.6c, against $133.33 at resolution - $11.03 of pull-to-par over four months. The hard 5c auto-close we added underneath was not triggered; the session low was 7.05c and that rule has no discretion in it, so there is nothing to interpret. What did change is fair value, and in our favour. ISM manufacturing printed 54.6% for August with the overall economy expanding for a 22nd month, and GDPNow refreshed up to 4.8% for the third quarter, taking off a carry flag we had held for four sessions. An economy growing near 5% does not add 1.4 million unemployed in four months, which is what 4.1% to 5.0% by December still requires. So the fair value on the YES goes 8 to 7, and the NO we hold flips from being marked rich yesterday to marked slightly cheap today. That reversal is worth naming rather than sliding past.
▵ Bull case
- ISM manufacturing 54.6%, eighth consecutive month of expansion; GDPNow refreshed UP to 4.8% for Q3.
- 4.1% to 5.0% by December needs roughly 1.4 million additional unemployed in four months. Nothing in the data flow makes that live.
- A Fed that futures price at 68% to HIKE in a fortnight is not looking at a deteriorating labour market.
- The BLS benchmark payroll revision landed at -79,000 on Friday - benign against last year's -911,000. The scary headline never came.
▿ Bear case
- July payrolls FELL 23,000. That is the real bear datapoint and it keeps full weight regardless of how good ISM looks.
- Friday's consensus has unemployment ticking to 4.2% from 4.1%. Two more prints like that and this leg re-rates, taking our mark with it.
- We are risking $122.30 of realisable value to collect $11.03. Eleven to one against, and no base rate makes that comfortable to write.
- ADP lands today and has diverged sharply from the establishment survey twice this year. A soft print moves this before Friday even arrives.