Daily Polymarket Alpha

Daily US Global Rates Portfolio Archive Method

● Live —

Daily US Global Rates Portfolio Archive Method

● Live

Analysis Archive

Every published analysis — 69 Daily US and 68 Global Rates editions.

Daily US

Sep 9, 2026
Daily Macro US
THE GATE'S BILL IS FINAL: $42.91. The 10-year printed 4.80% on Sep 8 (Treasury par curve, primary source) and Polymarket's 4.8%-touch contract resolved YES at 22:21 UTC. We sold it Sep 1 at a volume-weighted 71.69c on a 66c-bid gate published since Aug 20 - 151.5152 shares, $108.61 taken, +$83.61 realised. At resolution those shares pay $151.52. Correct process, public trigger, clean fills, and a $42.91 invoice on a $25 ticket. WORSE: the barrier model we retired on Sep 1 priced this contract at 76.8 against a 67.5 market on Aug 31 and it resolved at 100 - the model we threw away beat the market on the last question it was asked. It stays retired on a pre-commitment written before this was knowable, but the retirement was not costless and we say so. BET CLOSED: pos-011 sold at a volume-weighted 40.11c, +$4.49 on $25, NOT on a gate. Re-deriving its fair value we found the published method is FV = market_4.0 x (our P(>=1 hike) / the market's hike mass) - the price of the thing being valued is on both sides, and it says nothing about whether the Fed hikes ONCE or TWICE, which is the whole contract. Priced properly off a per-meeting path (62.8 / 23.6 / 61.2, resolution rule verified this session - the December meeting counts), P(exactly one hike) is 39.4% at zero correlation and 35.0 / 30.5 / 25.4 / 18.8 as correlation rises, against a 40.2c bid. THE SIGN DOES NOT FLIP, so we acted. NO NEW TRADE: the same anchor makes the Fed's September hike worth 62.8% against a 53c ask, +9.8pp - under the 10-point bar, and on a dashboard we could not corroborate because CME FedWatch was unreachable and press citations of it span 52-58.7%. Pre-commitment written for tomorrow: >=60% from a citable source and we take it; 52-59% and it is refused permanently. NEW STANDING RULE: every published fair value must name which inputs are exogenous to the market being priced; if none are, there is no fair value, only a price. That test would have caught all three models we have now retired. pos-004 is the whole book - YES 9.75, flat a third session, sixth without an explanation, 3.8 points cheap on the bid, both gates untriggered and neither hostage to a print later today. One open, $100 staked, +$20.20 open, +$743.25 realised, +$763.45 total - identical to yesterday to the cent. 19/22.
Sep 8, 2026
Daily Macro US
The backtest we pre-committed to on Sunday killed the trade it was meant to justify, and reversed its sign. The August CPI MoM 0.3% bucket screened +5.6pp against the ask on an ASSUMED standard deviation of 0.05. The Cleveland Fed publishes an evaluation of its own nowcasting model - EC 2023-06, Knotek and Zaman, Table 1, real-time data Sep 2000 to Dec 2022 - and the RMSE for headline CPI at the last-day-of-target-month horizon is 0.164 percentage points. WE WERE 3.3x TOO CONFIDENT. On the measured number the bucket is 12.6pp RICH, not 5.6pp cheap. We are not banking the reversal either: 0.164 is unconditional across a sample including the 2021-22 surge, and at that dispersion our model flattens to 10.0/15.1/22.4/23.3/29.2 against a market peaked at 2.6/6.0/33.0/48.5/13.5 - and a flat model always calls the tails cheap, which is the exact shape the barrier model failed in. THE SIGN FLIPS BETWEEN sd 0.08 AND 0.10 (+5.7/+1.6/-2.5/-7.9/-12.6) and we cannot pin dispersion that finely. A position whose DIRECTION depends on an input we cannot measure is not a position, so the pre-commitment fires as written and the CPI-BRACKET CATEGORY IS RETIRED to diagnostic-only - the category that produced three of this book's biggest early winners, retired on evidence we went looking for ourselves. Reinstatement needs a CONDITIONAL dispersion estimate reproducing the market's peaked shape within 5 points across three consecutive releases. LIMITATION PUBLISHED: we promised a bias check and the paper reports RMSE but not mean error, so half that promise is unfulfilled. CORRECTION AT THE TOP: MONDAY SEP 7 WAS LABOR DAY and yesterday's letter treated a shut US session as a normal trading day, running a daily-settle gate check against a day with no settle - one letter after adopting that rule. Treasury curve has no Sep 7 row. The trades STAND (Polymarket runs continuously, published bids, verified depth) but the description was wrong; and the Cleveland 09/04 stamp we had flagged for three sessions was correct all along - Monday was not a business day. pos-017 keeps billing: the 4.8% touch trades 98.40 against our 71.69c fill, $38.81 foregone at the bid and $42.91 at resolution, two basis points from printing. FV: pos-011 41 unchanged (and its cross-check is GONE, deleted by our own Sep 7 close), pos-004 6 unchanged. No trades, no gate fired. Two open, $125 staked, +$24.69 open, +$738.76 realised, 18/21. FIFTH SCREEN IN EIGHT SESSIONS, NO TRADE, PUBLISHED AS THE FINDING: today we hunted CONSTRAINT VIOLATIONS rather than forecasts - 56 exclusive ladders where a bid sum above 100 locks a profit with no model needed. Two hits, both net negative. Fed October sums 101.4 (a real 1.4pp edge on $864K depth, but a $2.45 prize across five simultaneous legs); China inflation 2026 sums 104.2 against a ~3.4pp nine-leg half-spread cost on $9.2K of total liquidity - the arb exists BECAUSE nobody tightens those quotes and collecting it means crossing every spread that creates it. THE BINDING CONSTRAINT IS OUR OWN SIZE, not the opportunity set: at $125-350 a position, trades must clear execution costs we have now measured twice (pos-017 surrendered 16 points to a ten-share book; pos-011's top bid holds 41 shares against 73.5 owned). We ACCEPT LOW DEPLOYMENT AS THE ANSWER - a $1,763 book with a 10pp bar in efficiently priced macro will sit in cash most of the time, and 91.5% cash is what that looks like, not a malfunction. The tiered entry bar proposed Sep 7 is SHELVED UNBUILT for the same reason.
Sep 7, 2026
Daily Macro US
The letter Sunday owed, and it closes two positions. pos-013 FULLY CLOSED at the 92.8c bid for +$8.03: Sunday trimmed half on the 90c gate, said in the same note that the arithmetic demanded ALL of it, and admitted to picking the trim convention over the pos-002/pos-003 precedent - so the pre-commitment written then put a full exit on the table today if the leg was still at or above fair value. It is. The meeting ladder prices cuts at 0.50% Sep / 3.70% Oct / 8.15% Dec, so the PERFECT-CORRELATION ceiling - the most generous construction available - is 91.85 against a 92.85 market, with $3.80 of upside left over 115 days against $49.03 at risk. 12.9 to 1 against. pos-010 ALSO CLOSED at 70c for +$3.41 AND NO GATE REQUIRED IT: fair value cut 70 -> 65 on two independent routes that agree (meeting blend 64, ladder hike mass 65.1) against a 70.5 market. Having closed pos-013 at ONE point rich, holding this at 5.5 would mean a stricter test on the position we wanted to sell than the one we wanted to keep. Counterweight published: the stub had BETTER convexity than pos-013, 2.3:1 vs 12.9:1, and selling on a consistency argument costs something. THE STRUCTURAL FAILURE IS FIXED RATHER THAN CONFESSED A FIFTH TIME - every gate since Aug 31 was honoured late because it fired intraday with no run scheduled, so effective today EVERY PRICE-TRIGGERED GATE EVALUATES ON THE DAILY SETTLE. A once-a-day desk cannot honour an intraday trigger; matching trigger frequency to observation frequency removes the failure by construction, at the cost of the best intraday exits. AND THE TRADE WE REFUSED CLOSED ITS OWN GAP: Fed September was 59.5c vs a 68% FedWatch on Sep 2, refused at 8.5pp under our 10pp bar; today 48.5c vs 49.4%, a 0.9pp gap closed by the MARKET falling eleven points to meet the futures. The bar saved an eleven-point drawdown into a nine-day expiry and we did not earn it - Waller did, and we did not forecast Waller. pos-004 rose a FOURTH straight session against the best news it will ever get (payrolls +162k, July revised positive, unemployment HELD at 4.1%) and we still have no explanation. FV: pos-010 70->65, pos-011 44->41, pos-004 unchanged at 6. Brent near $97 after US strikes on three Iranian tankers. Cleveland and GDPNow CARRIED and flagged. Book down to TWO positions and $125 staked - the smallest it has been - both marked cheap on the same test that closed the others. +$24.28 open, +$738.76 realised, 18/21. SIZING RESCALED TO EQUITY, decided and published today: equity is $1,763.04 with 91.5% - $1,613.76 - sitting in CASH and no NEW position opened since July. The stake ladder was always percentages of a $1,000 bankroll and we stopped updating the base; restated to $50/$100/$125/$175/$225/$275/$350, defined as 2.5-20% of equity and recomputed monthly, cap 20%. THE ENTRY BAR, KELLY FORMULA AND CORRELATION CAP ARE UNTOUCHED - nothing gets bought that would not have been. Published risk: scaling to equity is PRO-CYCLICAL, it raises absolute risk right after a 76% gain, and readers should judge the next drawdown accordingly.
Sep 2, 2026
Daily Macro US
The gate that cost us $29.27 in a day, published above everything that worked: pos-017, sold Tuesday at a volume-weighted 71.69c because the 66c-bid take-profit fired, trades 94.25c this morning after the 10-year settled 4.79% - a FIFTH consecutive higher close and ONE BASIS POINT from the 4.80% touch. Today's 91c bid carries 271 shares where Tuesday's 82c bid carried ten; our 151.5152 shares would realise $137.88 against the $108.61 we took. Foregone $29.27 now, $42.91 if it resolves. WE DO NOT RELITIGATE IT - the rule was written when FV was 60, published daily for a fortnight, honoured the morning it fired. Correct process, bad outcome. SAME GATE PROBLEM, OTHER DIRECTION: pos-010's 72c trim fired a SECOND time in 26 hours at 72.5c; honoured at the 71c bid for +$7.27, two points BELOW our own FV of 73. A half-cent breach re-triggering the same liquidation on consecutive days is a trigger without hysteresis, so the fix is a RE-ARM CONDITION not a wider number - 72c stays put, re-arms only below 68c. Mechanism, not level. METHOD REFINED AGAIN AT OUR EXPENSE: pos-011's top-of-book bid is TEN SHARES against 73.5294 held, so from today every mark is the bid OUR OWN SIZE CLEARS, walked through the book - costs $0.32 and we found it by looking, not by losing. Retired barrier model IMPROVING and STILL RETIRED: 5.00% check narrowed 11.1 -> 7.8pp but the reinstatement rule needs within 5 points for three straight sessions; session one fails. DATA: Cleveland's first SEPTEMBER core PCE nowcast 3.49%, ACCELERATING from 3.40% - the number of the day; ISM manufacturing 54.6%, eighth straight expansion; GDPNow refreshed UP to 4.8%, four-session carry flag OFF; Brent settled $94.65 (+4.6%), WTI $90.22 (+5.2%) after strikes on Larak Island and threats to Kharg Island (~90% of Iran's crude exports); FedWatch 65-68%. FV: pos-013 91->92, pos-010 71->73, pos-004 8->7, pos-011 unchanged at 44. One refusal. AND THE BOOK MOVED EIGHTEEN CENTS - $757.32 to $757.14 on the most hawkish session of the year, because the position that would have captured it was sold Tuesday. Four open, $232 staked, +$41.54 at the size-aware bid, +$715.60 realised, 14/17.
Sep 1, 2026
Daily Macro US
The gate fired and the bid had ten shares behind it: pos-017 CLOSED at a volume-weighted 71.69c against an 82c top-of-book bid, +$83.61 realised on $25 (+334%), every share filled above the published 66c-bid trigger. THE EXECUTION LADDER IS PUBLISHED IN FULL - 10 shares at 82c, 10 at 81c, 3.53 at 73c, 14.82 at 71c, 113.17 at 70c - because it exposes an error of ours: we had marked this position at the MID, which said $132.58 on Friday against a realisable $108.61. We carried $23.96 of paper profit that never existed at our size. FROM TODAY THE BOOK MARKS AT THE BID, a change that lowers our own open P&L by $5.36. SECOND GATE, SECOND BILL: pos-010's trim above 72c fired at 75.5c at 17:00 UTC Monday, five hours after Friday's letter published; we executed half the remainder at 70c today, +$13.64, and the 5.5 points between trigger and fill is the second time in three sessions that a 04:00 ET publish time has cost real money - and our own fair value is 71, so the gate made us sell BELOW fair value and we honoured it anyway. pos-004's sub-8c review fired at 7.10c and concluded HOLD; we refused to move the trigger, having criticised ourselves for moving pos-011's on Saturday, and added a hard auto-close at 5c instead. THE PRE-COMMITMENT FIRED ON SCHEDULE: the 5.00% consistency check failed a FIFTH straight session at 11.1pp, so the barrier model is RETIRED to diagnostic-only exactly as written before we knew today's prices - it narrowed from 20.2 and still cleared the bar, and refusing to argue the improvement is why we wrote it in advance. 10Y settled 4.75%, highest since January 2025, barrier to the 4.80% touch now 5.0bp; JGB 3.00% first time since 1996, Bund 3.3% first time since May 2011, French yields at 2008 highs; Brent above $91 on renewed US-Iran strikes. FedWatch 66% for September. FV: pos-013 90->91, pos-010 67->71, pos-011 42->44, pos-004 unchanged at 8. Two refusals. Book +$48.99 open at the bid on $257 staked, +$708.33 realised, 13/16.
Aug 31, 2026
Daily Macro US
The twelve-point move we published a warning about and then missed: Warsh told Jackson Hole the Fed may have work to do, the September hike contract went 30.5c to 52.5c in three hours and pos-010 went 56.5c to 68.5c, all at 10:00 ET on Friday with this letter six hours old. FULL GATE AUDIT with hourly evidence - pos-013's 90c trim did NOT fire (high 88.9c), pos-010's 72c trim missed by 3.5 cents, pos-011's 40c review DID fire on a weekend print and is conducted here two days late, concluding hold. pos-017 barrier now 7.0bp and the model says 76.8, but the 5.00% check failed a FOURTH straight session at 20.2pp so FV stays FROZEN at 61 - a 15.8-point self-imposed understatement, with a pre-committed replacement rule published today before we know tomorrow's prices. BLS benchmark revision -79,000, benign. FV up: pos-010 55->67, pos-011 31->42, pos-013 89->90. Two corrections: we were wrong that gamma hides the bid, and both of Friday's session counts were wrong. No trades. Book +$134.04 open, +$611.08 realised, 11/14.
Aug 28, 2026
Daily Macro US
The feed came back and took our excuse with it: yesterday we blamed the failing 5.00% consistency check partly on a carried market leg, and today both legs are live, the market prints 17.5c against a 35.1 model, and the gap is WIDER at 17.6pp. pos-017 fair value stays FROZEN at 61 against an arithmetic 66.8, so our own rule now understates the book by 5.8 points. Claims 203k beat ~208k, Brent settled UP 0.77%, Cleveland core nowcast holds 3.40%. FV up: pos-013 88->89, pos-010 54->55, pos-011 30->31. New 04:06 ET publishing time means Warsh and the BLS benchmark revision are PREVIEWS, and the pos-013 trim at 90c may fire unwatched
Aug 27, 2026
Daily Macro US
Our own rule just cost us four points and we kept it: the 10Y settle at 4.664% narrows the pos-017 barrier to 13.6bp and the model says 65, but the Aug 25 standing rule freezes fair value at 61 because the 5.00% check still fails - and got worse. July PCE 3.7%/core 3.3% beat consensus and our Cleveland anchor beat consensus too. GDPNow 4.0% -> 4.6%. Real spending flat. Polymarket unreachable all session, every position price CARRIED
Aug 26, 2026
Daily Macro US
The war premium unwinds and the whole book feels it at once: Iran-Oman agree a framework for a temporary Hormuz corridor, Brent falls 10% in three sessions to $84.90, the 10Y settles at 4.639% and our pos-017 fair value drops eleven points from 72 to 61. We publish a correction - yesterday's Brent and 10Y figures were intraday marks, not settles. FV cut on four of five positions, no trades
Aug 25, 2026
Daily Macro US
The consistency check we used yesterday to defend our own fair value does not reproduce: the same model gives 40 on the 10Y 5.00% leg, not 16, against a 17.5c market. We freeze pos-017 FV at 72 instead of taking the 73 the arithmetic offers, and refuse a nominal 22pp gap as model failure rather than edge. Fed-hike-2026 56.5c, GDPNow Q3 4.0%, no trades
Aug 24, 2026
Daily Macro US
Friday's flagged oddity resolved and we were not on it: the US-Canada tariff collapse takes Fed-hike-2026 49.5c -> 55.5c. We refused to add at 49.5c under our own 10pp bar and we publish what that cost. pos-017 FV 70 -> 72, arithmetic not narrative
Aug 21, 2026
Daily Macro US
The buyback bid lasted one session: the 10Y is back to 4.691% and we are reversing our own fair-value cut, 60 → 70, one day after making it. pos-017 +$75.76 — and we are not raising the gate we lowered
Aug 20, 2026
Daily Macro US
Treasury doubles its long-end buybacks and our best bet loses its engine: pos-017 fair value cut 74 → 60, now marked above our own FV — while pos-011 walks to the 30 we published. Four missed sessions disclosed
Aug 13, 2026
Daily Macro US
The trade we refused won: July CPI printed 3.4% and cost us $41.67 — we are keeping the bar anyway. BoJ hits 72c after we passed at 42.5. Oil near $90 puts pos-017 11bp from home at +$78
Aug 10, 2026
Daily Macro US
We missed Friday's -23K payroll print, and we were wrong about our own best position: pos-017 fair value corrected 55 -> 65 as Iran's Hormuz conditions re-arm oil; refusing that 44.5c Fed-September ticket saved 10 points
Aug 6, 2026
Daily Macro US
ADP misses at +44K and the touch bet slips again — fair value cut a second day (62→55); we refused two trades, including a 12pp Fed-September gap we could not verify at source
Aug 5, 2026
Daily Macro US
Dovish drift tests the book before NFP: soft JOLTS + Iran calm pull the 10Y to 4.6%, our touch bet gives back 11c — held into tomorrow's decisive jobs print; no-cut bet unmoved
Aug 4, 2026
Daily Macro US
Iran de-escalation cools the book: oil -5%, 10Y back to 4.68% — our touch bet holds on dual-engine design; ECB market flips 16→87.5c, the non-trade we're glad we skipped
Aug 3, 2026
Daily Macro US
Crescendo delivered: GDP bet wins +$114 (GDPNow nailed it), hawkish 9-3 Fed lifts the book to a record — but we MISSED the BoJ trade our own thesis called, for lack of event-day coverage
Jul 27, 2026
Daily Macro US
First Fed-complex profit banked: pos-010 trimmed at 66.5c (+$21) as hike odds hit 38% and oil rolls over — 10Y at 4.71% leaves our touch bet 9bp from home, into FOMC week
Jul 22, 2026
Daily Macro US
Brent $92 drives the 10Y to within 21bp of our touch bet — book hits a fresh high; we decline to re-chase the BoJ gate at our own re-open line, three sessions from the crescendo
Jul 21, 2026
Daily Macro US
Hormuz gets a 20% toll and the BoJ gate closes empty: our two-week thesis repriced 8→19c before the gate opened — right call, no trade, rule intact
Jul 20, 2026
Daily Macro US
The scan that mostly said no: Friday's automated GDP entry opens +35% while the CPI wing trade dies by its own condition and the yen hedge runs away — a study in process
Jul 17, 2026
Daily Macro US
The war premium is back: Brent $84.5 re-arms the hawkish book — zero-cuts firms to 84c on real volume, claims 208K starve the cut case; 10Y-touch lags its own thesis again
Jul 15, 2026
Daily Macro US
China prints 4.3% candor — pos-016 loses $25 and the regime lesson gets published; BoJ gate corrected to Jul 30-31 (our error); Fed-July completes its 10→50→7c whipsaw
Jul 14, 2026
Daily Macro US
The referee said 3.5% — everyone wrong, us most of all: flagship loses $75, full post-mortem. July-hike market whipsaws 50c→9c in a morning; amended pos-010 rule holds through the round trip; oil +9% keeps the 10Y ticket alive
Jul 13, 2026
Daily Macro US
Eve of the gauntlet: flagship at 22.5c into tomorrow's CPI + Warsh double-header — China bracket's mystery seller surfaces (Iran Q2 impact), BoJ gate case strengthens overnight
Jul 12, 2026
Daily Macro US
Sunday brief: weekend surge takes hike-2026 to one tick below our trim line — every Tuesday branch pre-committed before CPI + Warsh testimony hit the same morning
Jul 10, 2026
Daily Macro US
Market walks toward our 3.9% bracket (15→20c) as Cleveland prints 3.92 a fifth straight day — 10Y touches 4.60, hike-2026 gap re-widens, deploy gate armed for CPI Tuesday
Jul 9, 2026
Daily Macro US
The CPI crowd round-trips in 24 hours — contagion call vindicated as 3.9 firms to 17c; split-committee minutes fade July hike (our exit caught the top); new 10Y touch position at 16.5c
Jul 8, 2026
Daily Macro US
Fastest win yet: Fed-July trade exits +68% in 22 hours as the market converges to FedWatch — pos-010 goes green at 57.5c; CPI market reorganizes around our 3.9% bracket, Cleveland unmoved at 3.92
Jul 7, 2026
Daily Macro US
Widest gap we've published: Cleveland nowcasts June CPI at 3.92% — the 3.9% bracket trades at 15c. Entered $75, plus Fed-July and China GDP; core-CPI correction logged
Jul 6, 2026
Daily Macro US
First loss on the book: IRGC corridor warnings crash Hormuz to 14.5c, exit rule fires at -$55 — zero-cuts add executed at 77.7c; CPI Jul 14 is the week's gate
Jul 3, 2026
Daily Macro US
Correction: NFP June was +57K, not ~110K — July hike odds collapse to 22%, zero-cuts cracks 13pp to 78c; Dow record 52.9K, CPI Jul 14 decisive
Jul 2, 2026
Jul 2, 2026
NFP June ~110K misses 150K deploy threshold — Brent erases war premium, pos-012 under pressure; CPI July 14 is the next trigger
Jul 1, 2026
Jul 1, 2026
Tankan +22 smashes forecasts, 10Y back to 4.47% — NFP tomorrow is the deploy trigger for pos-010's $275 free capital
Jun 30, 2026
Jun 30, 2026
Dow first above 52K as SCOTUS locks in Fed independence — Doha talks today, pos-010 gap widens to 15pp, NFP Thursday
Jun 29, 2026
Jun 29, 2026
Nasdaq -4.6% on AI bubble fears, Hormuz MOU fractures to 35% — NFP Thursday is the week's fulcrum
Jun 25, 2026
Jun 25, 2026
PCE May hits 4.0% YoY as Hormuz flows 4.8M b/d — adding to pos-012, Micron reverses AH gains, S&P -1%
Jun 24, 2026
Jun 24, 2026
Micron earnings and Fed stress tests after close — WTI $72, PCE May tomorrow, Hormuz 47%
Jun 23, 2026
Jun 23, 2026
Taking +$277 early profit on inflation NO (pos-001) as oil deflation closes the 4.5% path — Hormuz 49%, PCE May Friday
Jun 22, 2026
Daily Macro US
Chips lead Monday bounce to 7,500 — closing pos-009 at exit target, Hormuz 41%, PCE May this week
Jun 19, 2026
Daily Macro US
MOU signed, blockade lifted, tankers moving — pos-012 at 46% despite physical Hormuz progress
Jun 18, 2026
Daily Macro US
Polymarket catches up — Fed hike odds 57% (+20pp overnight), WTI $75, Iran signing tomorrow
Jun 17, 2026
Daily Macro US
Warsh holds, 9 of 18 hike dots — S&P -1.21% to 7,420, market sells off on hawkish dot plot
Jun 16, 2026
Daily Macro US
BoJ hiked to 1.0% (highest since 1995) — FOMC Day 1, Warsh presser tomorrow is everything
Jun 15, 2026
Daily Macro US
Iran deal confirmed — WTI $82, Fed hike odds collapse to 38%, Warsh presser tomorrow is everything
Jun 12, 2026
Daily Macro US
Iran deal imminent — WTI crashes to $86, Fed hike odds slide to 50.5%, FOMC 4 days
Jun 11, 2026
Daily Macro US
ECB hiked +25bp to 2.25% — S&P falls to 7,267, FOMC June 16-17 is the week's pivot
Jun 10, 2026
Daily Macro US
CPI 4.2% resolves three positions +$208.61 — Fed hike odds hit 72%, ECB tomorrow
Jun 9, 2026
Daily Macro US
CPI in 12 hours — three positions resolve tomorrow at 8:30am ET
Jun 8, 2026
Daily Macro US
CPI prints tomorrow — pos-005 closed +$48.53, BoJ 97%
Jun 5, 2026
Daily Macro US
NFP +172K — unemployment holds 4.3%, pos-005 resolves YES
Jun 4, 2026
Daily Macro US
Oil pulls back $88.90 — ECB tomorrow
Jun 3, 2026
Daily Macro US
Iran escalation — WTI surges $93-97
Jun 2, 2026
Daily Macro US
ECB position exited — HICP invalidation
Jun 1, 2026
Daily Macro US
Oil stabilising — 35pp gap widens
May 31, 2026
Daily Macro US
Month-end rebalance — 9 positions
May 30, 2026
Daily Macro US
Cleveland nowcast 3.3-3.5% confirmed
May 29, 2026
Daily Macro US
Fed hold 97% — rate cuts off table
May 28, 2026
Daily Macro US
CPI MoM gap widens to 28pp
May 27, 2026
Daily Macro US
Memorial Day — markets thin
May 26, 2026
Daily Macro US
PCE 3.3% — inflation thesis intact
May 25, 2026
Daily Macro US
SPF forecast: unemployment capped 4.5%
May 24, 2026
Daily Macro US
GDPNow 4.3% — recession risk low
May 23, 2026
Daily Macro US
US Recession 2026 entered at 28%
May 22, 2026
Daily Macro US
CPI YoY 4.3% + 4.4% positions entered
May 21, 2026
Daily Macro US
Unemployment + CPI MoM positions entered
May 20, 2026
Daily Macro US
Portfolio opened — 4 positions entered

Global Rates

Sep 9, 2026
Global Rates
THE GATE'S BILL IS FINAL: $42.91. The 10-year printed 4.80% on Sep 8 (Treasury par curve, primary source) and Polymarket's 4.8%-touch contract resolved YES at 22:21 UTC. We sold it Sep 1 at a volume-weighted 71.69c on a 66c-bid gate published since Aug 20 - 151.5152 shares, $108.61 taken, +$83.61 realised. At resolution those shares pay $151.52. Correct process, public trigger, clean fills, and a $42.91 invoice on a $25 ticket. WORSE: the barrier model we retired on Sep 1 priced this contract at 76.8 against a 67.5 market on Aug 31 and it resolved at 100 - the model we threw away beat the market on the last question it was asked. It stays retired on a pre-commitment written before this was knowable, but the retirement was not costless and we say so. BET CLOSED: pos-011 sold at a volume-weighted 40.11c, +$4.49 on $25, NOT on a gate. Re-deriving its fair value we found the published method is FV = market_4.0 x (our P(>=1 hike) / the market's hike mass) - the price of the thing being valued is on both sides, and it says nothing about whether the Fed hikes ONCE or TWICE, which is the whole contract. Priced properly off a per-meeting path (62.8 / 23.6 / 61.2, resolution rule verified this session - the December meeting counts), P(exactly one hike) is 39.4% at zero correlation and 35.0 / 30.5 / 25.4 / 18.8 as correlation rises, against a 40.2c bid. THE SIGN DOES NOT FLIP, so we acted. NO NEW TRADE: the same anchor makes the Fed's September hike worth 62.8% against a 53c ask, +9.8pp - under the 10-point bar, and on a dashboard we could not corroborate because CME FedWatch was unreachable and press citations of it span 52-58.7%. Pre-commitment written for tomorrow: >=60% from a citable source and we take it; 52-59% and it is refused permanently. NEW STANDING RULE: every published fair value must name which inputs are exogenous to the market being priced; if none are, there is no fair value, only a price. That test would have caught all three models we have now retired. pos-004 is the whole book - YES 9.75, flat a third session, sixth without an explanation, 3.8 points cheap on the bid, both gates untriggered and neither hostage to a print later today. One open, $100 staked, +$20.20 open, +$743.25 realised, +$763.45 total - identical to yesterday to the cent. 19/22.
Sep 8, 2026
Global Rates
The backtest we pre-committed to on Sunday killed the trade it was meant to justify, and reversed its sign. The August CPI MoM 0.3% bucket screened +5.6pp against the ask on an ASSUMED standard deviation of 0.05. The Cleveland Fed publishes an evaluation of its own nowcasting model - EC 2023-06, Knotek and Zaman, Table 1, real-time data Sep 2000 to Dec 2022 - and the RMSE for headline CPI at the last-day-of-target-month horizon is 0.164 percentage points. WE WERE 3.3x TOO CONFIDENT. On the measured number the bucket is 12.6pp RICH, not 5.6pp cheap. We are not banking the reversal either: 0.164 is unconditional across a sample including the 2021-22 surge, and at that dispersion our model flattens to 10.0/15.1/22.4/23.3/29.2 against a market peaked at 2.6/6.0/33.0/48.5/13.5 - and a flat model always calls the tails cheap, which is the exact shape the barrier model failed in. THE SIGN FLIPS BETWEEN sd 0.08 AND 0.10 (+5.7/+1.6/-2.5/-7.9/-12.6) and we cannot pin dispersion that finely. A position whose DIRECTION depends on an input we cannot measure is not a position, so the pre-commitment fires as written and the CPI-BRACKET CATEGORY IS RETIRED to diagnostic-only - the category that produced three of this book's biggest early winners, retired on evidence we went looking for ourselves. Reinstatement needs a CONDITIONAL dispersion estimate reproducing the market's peaked shape within 5 points across three consecutive releases. LIMITATION PUBLISHED: we promised a bias check and the paper reports RMSE but not mean error, so half that promise is unfulfilled. CORRECTION AT THE TOP: MONDAY SEP 7 WAS LABOR DAY and yesterday's letter treated a shut US session as a normal trading day, running a daily-settle gate check against a day with no settle - one letter after adopting that rule. Treasury curve has no Sep 7 row. The trades STAND (Polymarket runs continuously, published bids, verified depth) but the description was wrong; and the Cleveland 09/04 stamp we had flagged for three sessions was correct all along - Monday was not a business day. pos-017 keeps billing: the 4.8% touch trades 98.40 against our 71.69c fill, $38.81 foregone at the bid and $42.91 at resolution, two basis points from printing. FV: pos-011 41 unchanged (and its cross-check is GONE, deleted by our own Sep 7 close), pos-004 6 unchanged. No trades, no gate fired. Two open, $125 staked, +$24.69 open, +$738.76 realised, 18/21. FIFTH SCREEN IN EIGHT SESSIONS, NO TRADE, PUBLISHED AS THE FINDING: today we hunted CONSTRAINT VIOLATIONS rather than forecasts - 56 exclusive ladders where a bid sum above 100 locks a profit with no model needed. Two hits, both net negative. Fed October sums 101.4 (a real 1.4pp edge on $864K depth, but a $2.45 prize across five simultaneous legs); China inflation 2026 sums 104.2 against a ~3.4pp nine-leg half-spread cost on $9.2K of total liquidity - the arb exists BECAUSE nobody tightens those quotes and collecting it means crossing every spread that creates it. THE BINDING CONSTRAINT IS OUR OWN SIZE, not the opportunity set: at $125-350 a position, trades must clear execution costs we have now measured twice (pos-017 surrendered 16 points to a ten-share book; pos-011's top bid holds 41 shares against 73.5 owned). We ACCEPT LOW DEPLOYMENT AS THE ANSWER - a $1,763 book with a 10pp bar in efficiently priced macro will sit in cash most of the time, and 91.5% cash is what that looks like, not a malfunction. The tiered entry bar proposed Sep 7 is SHELVED UNBUILT for the same reason.
Sep 7, 2026
Global Rates
Nine days ago this book held five positions expressing one view; this morning it holds two, and not because the view was wrong - the market came to meet us and went past. THE TRADE WE REFUSED CLOSED ITS OWN GAP: Fed September was 59.5c against a 65-68% FedWatch on Sep 2 and we declined at 8.5pp under our published 10pp bar, writing that 'nine is basically ten' is exactly the sentence a bar exists to prevent. Today it is 48.5c against a 49.4% FedWatch - a 0.9pp gap closed by the MARKET FALLING ELEVEN POINTS to meet the futures. The bar saved an eleven-point drawdown into a nine-day expiry and WE DID NOT EARN IT: Waller told an audience Thursday he could back a hold if disinflation continues and that August CPI decides his vote, and we did not forecast Waller. Both remaining hawkish positions closed this morning. pos-013 IN FULL at 92.8c for +$8.03 on Sunday's pre-commitment - the meeting ladder prices cuts at 0.50% Sep / 3.70% Oct / 8.15% Dec, capping no-cut-in-2026 at 91.85 under PERFECT CORRELATION against a 92.85 market, with $3.80 of upside left against $49.03 at risk. pos-010 at 70c for +$3.41 with NO GATE behind it - FV cut 70->65 on two agreeing routes, and having closed pos-013 at one point rich we could not hold this at 5.5. Counterweight published: the stub had BETTER convexity, 2.3:1 vs 12.9:1. THE STRUCTURAL FAILURE IS FIXED NOT CONFESSED A FIFTH TIME - every gate since Aug 31 was honoured late because it fired intraday with no run scheduled, so from today EVERY PRICE-TRIGGERED GATE EVALUATES ON THE DAILY SETTLE. THE BOARD'S SHAPE IS THE REAL NEWS: Sep 48.5c, Oct 28.5c, Dec 43.5c - the market POSTPONED the hike rather than abandoning it, so we sold a mispriced expression of a view we still hold. Brent near $97, highest since July, after US strikes destroyed one of three Iranian tankers; Tehran promises a restricted maritime zone beyond Hormuz. 10Y 4.78% Sep 4 settle. Cleveland and GDPNow CARRIED and flagged. Two open, $125 staked - the smallest this book has been - +$24.28 open, +$738.76 realised, 18/21. SIZING RESCALED TO EQUITY, decided and published today: equity is $1,763.04 with 91.5% - $1,613.76 - sitting in CASH and no NEW position opened since July. The stake ladder was always percentages of a $1,000 bankroll and we stopped updating the base; restated to $50/$100/$125/$175/$225/$275/$350, defined as 2.5-20% of equity and recomputed monthly, cap 20%. THE ENTRY BAR, KELLY FORMULA AND CORRELATION CAP ARE UNTOUCHED - nothing gets bought that would not have been. Published risk: scaling to equity is PRO-CYCLICAL, it raises absolute risk right after a 76% gain, and readers should judge the next drawdown accordingly.
Sep 2, 2026
Global Rates
The board stopped being a board and became a single trade: BoJ September repriced 88.5c to 98.35c in ONE SESSION - the largest single-day move this board has shown - after the JGB 10-year touched 3.00% for the first time since 1996. ECB September 98.95c. Fed September 59.5c against a FedWatch reported 65-68%. Three of the world's four largest central banks priced to TIGHTEN inside a fortnight, and the cash markets got there first. THE US 10-YEAR SETTLED 4.79%, a fifth consecutive higher close and ONE BASIS POINT from the 4.80% touch - so this letter leads with a position we sold on Tuesday. pos-017 went at a volume-weighted 71.69c when the 66c-bid gate fired into a book with ten shares on top; it trades 94.25c now with a 91c bid 271 shares deep. Foregone $29.27 at today's bid, $42.91 at resolution, published ABOVE what worked. We do not relitigate it - written when FV was 60, published daily for a fortnight, honoured the morning it fired - but we concede the design flaw: a fixed take-profit that never tracks fair value sells winners early by construction. SAME PROBLEM, OTHER SIDE: pos-010's 72c trim fired a SECOND time in 26 hours at 72.5c, honoured at 71c for +$7.27, two points BELOW our own FV of 73; fix is a RE-ARM CONDITION not a wider number - 72c stays, re-arms only below 68c. METHOD REFINED AGAIN: pos-011's top bid is TEN SHARES, so marks are now the bid OUR SIZE CLEARS, walked through the book, costing $0.32. Retired model IMPROVED to a 7.8pp gap from 11.1 and STAYS RETIRED - session one of the three-session five-point test fails, and we note we have every incentive to reinstate a tool that would have told us to hold pos-017. Brent settled $94.65 (+4.6%), WTI $90.22 (+5.2%) after strikes on Larak Island and threats to Kharg Island (~90% of Iran's crude exports); yesterday's unverifiable-settle flag comes OFF. Three refusals. Book +$41.54 at the size-aware bid, +$715.60 realised, $757.14 total, 14/17 - down eighteen cents on the most hawkish session of the year.
Sep 1, 2026
Global Rates
The global long end broke together and it closed a position: the JGB 10-year touched 3.00% for the FIRST TIME SINCE 1996, the Bund hit 3.3% for the first time since May 2011, French yields reached 2008 highs and the US 10-year settled 4.75% (Treasury par curve), its highest since January 2025 - taking our 4.8%-touch contract from 67.5c to an 82c bid, through the 66c-bid take-profit we had published every day for a fortnight. pos-017 SOLD, all 151.5152 shares, volume-weighted 71.69c, +$83.61 realised on $25 (+334%). THE FILL LADDER IS PUBLISHED - 10 shares at 82c, 10 at 81c, 3.53 at 73c, 14.82 at 71c, 113.17 at 70c - because it exposes our own error: we marked at the MID, Friday's mid said $132.58, and $23.96 of that never existed at our size. THE BOOK NOW MARKS AT THE BID, lowering our published open P&L by $5.36. pos-010's 72c trim fired at 75.5c during Monday's US session while we were five hours published; half the remainder sold at 70c, +$13.64, and the 5.5-point gap between trigger and fill is a bill - our fair value is 71, so the gate made us sell BELOW fair value and we honoured it anyway. THE PRE-COMMITMENT FIRED: the 5.00% consistency check failed a FIFTH straight session at 11.1pp, so the barrier model is RETIRED to diagnostic-only exactly as written before today's prices were known - it narrowed from 20.2 and still cleared the bar, and refusing to argue the improvement is why it was pre-committed. Reinstatement now needs three consecutive sessions within 5 points. Central bank board all LIVE and all one direction: ECB Sep 98.55c, BoJ Sep 88.5c, Fed Sep 56.5c against FedWatch 66%. Renewed US-Iran strikes - Larak Island, attacks on the UAE and Jordan, Kharg Island threatened, a supertanker burning in Hormuz - put Brent at $91.28 LIVE, with no percentage change attached because we could not verify an Aug 31 settle. Three refusals on bar and correlation. Book +$48.99 open at the bid, +$708.33 realised, 13/16.
Aug 31, 2026
Global Rates
Warsh repriced the whole board in three hours and we were already published: the September Fed hike contract went 30.5c to 52.5c between 14:00 and 17:00 UTC Friday and trades 51.5c now, the 2-year settled +14bp at 4.34% against the 10-year's +6bp to 4.73%, and CME FedWatch is readable again at about 57% - the first fresh reading in nine sessions, so the stale 31.6 we have carried since Aug 20 is RETIRED. ECB Sep 98.35c, BoJ Sep 88.0c. pos-017's barrier is now 7.0bp and the model says 76.8, but the 5.00% consistency check failed a FOURTH straight session at 20.2pp - its widest - so fair value stays FROZEN at 61, a 15.8-point self-imposed understatement, with a replacement rule pre-committed today before we know tomorrow's prices. CORRECTION: we were wrong that the gamma feed hides the bid, on the one field pos-017's exit gate depends on - the bid is 65 against a 66c trigger. Bund/JGB/gilt CARRIED and quoted qualitatively. Three refusals on correlation and model failure. No trades. Book +$134.04 open, +$611.08 realised, 11/14.
Aug 28, 2026
Global Rates
Both feeds came back and the restoration cost us our excuse: the 5.00% consistency check now fails on TWO live legs by 17.6pp, wider than the 17pp we partly blamed on a carried leg, so pos-017 fair value stays FROZEN at 61 against an arithmetic 66.8. The Bund hit a 15-YEAR HIGH at 3.2648% and Ireland a 12.5-year high on the same morning we refused to raise the mark. JGB +3.7bp to 2.9320%, gilt 5.0491%, all live after two carried sessions. ECB Sep 97.6c, BoJ Sep 87.5c, Fed Sep 30.5c - within a point of our stale FedWatch 31.6, agreeing for the first time in eight sessions
Aug 27, 2026
Global Rates
A four-point upgrade blocked by our own rule: the 10Y settle at 4.664% narrows the pos-017 barrier to 13.6bp and the model says 65, but the Aug 25 rule freezes FV at 61 while the 5.00% check keeps failing. Polymarket unreachable all session so the entire central-bank board is carried; Bund, JGB and gilt carried for a second day; FedWatch stale for a seventh. The 5.00% gap grew to 17pp purely because our side updated and the market's did not
Aug 26, 2026
Global Rates
One Hormuz headline, two central banks moving opposite ways: ECB Sep hike UP to 96.75c on disinflationary news, BoJ Sep DOWN to 86.0c as cheaper crude guts the imported-inflation case. Brent $84.90, -10% in three sessions. Our pos-017 fair value cut 72 -> 61 after we correct yesterday's 10Y from an intraday 4.7050% to the 4.639% settle. Bund, JGB and gilt levels carried - feed failed at source
Aug 25, 2026
Global Rates
Four sovereign curves, four moves under 2bp — and the only thing that actually changed is our arithmetic. The 5.00% consistency check that validated pos-017 yesterday returns 40 today, not 16. FV frozen at 72, the 22pp gap refused. ECB Sep 95.5c, BoJ Sep 87.5c unchanged, Fed Sep 33.5c with our carried FV withdrawn
Aug 24, 2026
Global Rates
The three-way divergence closes from the American end: Fed Sep 27.5c -> 32.5c on a North American trade war, while ECB Sep hits 94.5c and BoJ Sep 87.5c. Bund near a 2011 high, JGB at a 1996 high, gilt above 5% — and our 5.00% consistency check says the 4.8% fair value is honest
Aug 21, 2026
Global Rates
Same oil shock, three answers: ECB September priced at 92c, BoJ at 84.5c, the Fed at 27.5c — Bund at a 2011 high, JGB at a 1996 high, and a US intervention that held the curve for one day
Aug 20, 2026
Global Rates
Two long ends hit multi-decade highs, one got rescued: the US Treasury doubled its buybacks and the 10Y fell while Brent rose — our oil-drives-term-premium mechanism failed a clean test, FV 74 → 60
Aug 13, 2026
Global Rates
Tokyo repriced without us: the BoJ hike we called at 40c trades at 72c, retired as a completed miss — while attacks on Red Sea shipping put Brent near $90 and the long end back to 4.69% on a benign CPI
Aug 10, 2026
Global Rates
The unwind splits: dovish front ends on -23K US payrolls, re-arming term premium on Iran's five Hormuz preconditions — we correct our own pos-017 fair value back up, 55 -> 65
Aug 6, 2026
Global Rates
Hormuz route deal takes WTI -11% and ADP misses at +44K — the synchronized unwind hits our premium ticket again while the structural one holds; a 12pp Fed gap refused for lack of verification
Aug 5, 2026
Global Rates
Synchronized unwind: soft JOLTS + Iran calm ease global long ends, Fed Sep hike sub-50% — the correlated book cools as one; no-cut anchor holds, NFP the hinge
Aug 4, 2026
Global Rates
De-escalation unwinds the war premium: oil -5%, yields ease, yen firms — the dual-engine touch bet holds; ECB market flips 16→87.5c, proving the 10pp bar both ways
Aug 3, 2026
Global Rates
The hawkish trio: 9-3 Fed, hawkish BoJ hold, 10Y at 4.75% — GDP bet wins, but the BoJ trade our thesis called slipped away for lack of event-day coverage
Jul 27, 2026
Global Rates
Oil's round-trip meets a live FOMC: Brent +27% then gapping lower on paused strikes — we trim our hike bet into the strength; 10Y at 4.71% and a 72-hour FOMC/GDP/BoJ cluster
Jul 22, 2026
Global Rates
Brent $92 is the whole story: term premium exporting worldwide, ceasefire markets confirm the war premium persists — we decline to re-chase the BoJ at our own re-open line
Jul 21, 2026
Global Rates
Policy becomes the supply shock: a 20% Hormuz toll reprices global inflation — and our BoJ thesis comes true at 19c, one gate too late
Jul 20, 2026
Global Rates
Process week opens: the yen hedge that ran away (63→74.5c unverified), the CPI trade killed by its own condition, and a fortnight bending toward Jul 29-31
Jul 17, 2026
Global Rates
The decoupling lasted eleven days: Brent $84.5 redistributes inflation risk across every curve — Japan CPI Tuesday could jump our gate; gold bracket dies without us
Jul 15, 2026
Global Rates
Beijing prints candor: 4.3% grades our bracket at zero — regime lesson filed; BoJ gate corrected to Jul 30-31 as Iran's oil bid decouples from US disinflation
Jul 14, 2026
Global Rates
One Washington number redraws every curve: dollar dumped, July hike vaporized — yet the long end rallied through it on a burning Strait. Tomorrow at dawn: China grades pos-016, BoJ statement executes the Sep gate
Jul 13, 2026
Global Rates
Asia leans dollar into the gauntlet — BoJ growth-upgrade reports draft our gate language; China bracket sold to 63c on Iran-impact fears, regime test Wednesday
Jul 12, 2026
Global Rates
Sunday brief: hawkish weekend radiates from Washington into BoJ week — 30 hours until CPI + Warsh, then BoJ + China grade half the book
Jul 10, 2026
Global Rates
Everything converges on a 30-hour window: US 10Y touches 4.60 into BoJ week, gold's $4,300 bracket collapses through our pass, China GDP drifts unexplained
Jul 9, 2026
Global Rates
Split-committee minutes drift global yields; BoJ market crawls 6→9c six days from the gate — crowd round-trips keep paying the model-anchored book
Jul 8, 2026
Global Rates
Hawkish FOMC minutes radiate globally: carry gap widens on the yen into BoJ Jul 15 — US CPI bracket migration puts the oil-crash disinflation narrative on trial
Jul 7, 2026
Global Rates
Yen 162.83 — Bessent urges BoJ hikes as $73B of intervention fails; BoJ market halves on no volume, China GDP entered at 72.5c
Jul 6, 2026
Global Rates
Oil decouples from Hormuz: Brent at 4-month lows despite IRGC standoff as OPEC+ adds supply — yen 1% from four-decade low, BoJ Jul 15 the gate
Jul 3, 2026
Global Rates
Yen at 161 with intervention warnings — BoJ Sep hike re-based to 11% (+14pp gap); Brent third straight decline, Hormuz 20 transits vs 60-call bar
Jul 2, 2026
Jul 2, 2026
Brent erases war premium, BoJ's Nagahama eyes year-end hike — KOSPI below 8,000 as soft NFP weakens dollar
Jul 1, 2026
Jul 1, 2026
Tankan +22 smashes forecasts, EUR CPI falls to 2.8% — BoJ entry approaching, NFP tomorrow, Doha stalled
Jun 30, 2026
Jun 30, 2026
Q2 closes with Dow above 52K — SCOTUS locks in Fed independence, Doha talks ongoing, NFP Thursday is the week's fulcrum
Jun 29, 2026
Jun 29, 2026
Nasdaq -4.6% week on AI capex bubble — Hormuz at 35% as IRGC fires on vessel; NFP Thursday + Tankan Tuesday ahead
Jun 25, 2026
Jun 25, 2026
PCE 4.0% YoY and Hormuz 4.8M b/d — 10Y hits 4.53%, USD/JPY pushes 162, pos-012 add confirmed
Jun 24, 2026
Jun 24, 2026
Nikkei -3.2% on hike fears but Kospi +3% as Micron eve — Brent $76.50, BoJ Sep 50%, PCE May tomorrow
Jun 23, 2026
Jun 23, 2026
Brent slides to $77 as Hormuz accelerates to 49% — yen at 161, BoJ Sep hike watch crosses 50%
Jun 22, 2026
Global Rates
USD/JPY hits 161 — BoJ Sep watch live as Hormuz slips to 41%, Bürgenstock talks resume
Jun 19, 2026
Global Rates
MOU in effect — blockade lifted, tankers moving, oil -10% on week. 60-day clock starts now.
Jun 18, 2026
Global Rates
Iran MOU signs tomorrow — WTI $75, BoJ Sep watch open, Poly hike 57% as dot plot sinks in
Jun 17, 2026
Global Rates
Warsh 9-dot hawkish shock — S&P -1.21% to 7,420, Poly Fed hike 37% lags CME 66%
Jun 16, 2026
Global Rates
BoJ hiked to 1.0% (highest since 1995, 7-1) — Warsh's first FOMC presser tomorrow is the global macro pivot
Jun 15, 2026
Global Rates
BoJ hikes today (98%) + FOMC starts today + Iran deal confirmed — the biggest macro week of 2026
Jun 12, 2026
Global Rates
BoJ 4 days (98%) + FOMC 4 days + Iran deal imminent — convergence week begins
Jun 11, 2026
Global Rates
ECB hiked to 2.25% (confirmed) — BoJ June 16 in 5 days (97.7%), FOMC simultaneously
Jun 10, 2026
Global Rates
ECB tomorrow (91%), BoJ in 6 days (96%) — US hike odds 72% reinforces global rates-higher
Jun 9, 2026
Global Rates
ECB decision tomorrow (91%), BoJ in 7 days (96%) — CPI cross-asset watch at 8:30am
Jun 8, 2026
Global Rates
ECB in 3 days (92%), BoJ in 8 days (97%) — both near-certainties
Jun 5, 2026
Global Rates
ECB June 11 is the key date — BoJ 43pp gap intact post-NFP
Jun 4, 2026
Global Rates
ECB hikes tomorrow — BoJ 43pp gap
Jun 3, 2026
Global Rates
Iran escalation — BoJ gap narrows to 43pp
Jun 2, 2026
Global Rates
ECB position exited on HICP trigger
Jun 1, 2026
Global Rates
BoJ June 16-17 — 46pp gap intact
May 31, 2026
Global Rates
USD/JPY 155 — yen weakness persists
May 30, 2026
Global Rates
ECB +25bp near-certain for June 5
May 29, 2026
Global Rates
BoJ hold vote 6-3 — dissent unlikely
May 28, 2026
Global Rates
ECB June hike: top opportunity replaces BoJ
May 27, 2026
Global Rates
Japan Q1 GDP -0.5% — hike case weak
May 26, 2026
Global Rates
BoJ gap 44pp — highest since entry
May 25, 2026
Global Rates
Reuters poll: 65% expect June hike
May 24, 2026
Global Rates
WTI $88 — import cost argument fades
May 23, 2026
Global Rates
BoJ meeting June 16-17 confirmed
May 22, 2026
Global Rates
BoJ +25bp June entered at 88%
May 21, 2026
Global Rates
ECB HICP 3.1% — June hike certain

MacroPilot AI is an analysis tool — not investment advice. Estimated probabilities are opinions and may be wrong. All trades carry risk of capital loss. MacroPilot AI © 2026

Get the daily analysis

Every day: 10+ macro markets screened, top opportunities ranked. Straight to your inbox — free.

Academic analysis only. Not investment advice.

You're subscribed!

Check your inbox to confirm.