Global Rates
JGB 10-year
3.00%
First time since 1996. Read as normalisation rather than crisis by most desks, but it is the single largest anchor moving under the global long end and it moved without a Bank of Japan meeting.
Bund 10-year
3.3%
First time since May 2011. French yields highest since November 2008, Dutch at fifteen-year highs, Italian and Spanish at multi-year highs. ECB minutes have officials seeing a further increase as likely necessary.
US 10-year
4.75% (Aug 31 SETTLE)
Treasury par yield curve, primary source. Highest since January 2025. The 2-year was flat at 4.34% and the 30-year settled 5.25% - a bear steepening, which is term premium rather than policy, and the opposite of Friday's shape.
pos-017 CLOSED
sold at 71.69c VWAP, +$83.61
The 66c-bid take-profit fired at an 82c bid. 151.5152 shares sold: 10 at 82c, 10 at 81c, 3.53 at 73c, 14.82 at 71c, 113.17 at 70c. Every share above the trigger. The mid said 87.5c and the mid was not available.
Central bank board
ECB 98.55c / BoJ 88.5c / Fed 56.5c
All three September meetings priced for tightening, all LIVE. ECB +0.2 on the day, BoJ +0.5, Fed +5.0. FedWatch has the US at 66%, a 9.5-point premium to the prediction market - under our 10-point bar and refused for a second session.
Brent crude
$91.28 LIVE
Renewed US-Iran strikes: Larak Island hit, Iranian attacks on the UAE and Jordan, Kharg Island threatened, a supertanker burning in Hormuz after two mine strikes. We could NOT verify an Aug 31 settle at source, so no percentage change is attached to this level.
The global long end broke together overnight and it did more damage to one of our positions than any American data point has all year. The Japanese 10-year touched 3.00% for the first time since 1996 - a level no bond desk staffed today has traded through in a professional career. The Bund reached 3.3%, the first time since May 2011. French yields hit their highest since November 2008, Dutch yields fifteen-year highs, Italian and Spanish yields multi-year highs. The American 10-year settled 4.75% on Aug 31 - the Treasury's own daily par yield curve, primary source - its highest since January 2025, leaving five basis points to the 4.80% level our pos-017 was written on. That combination did something a single-market move could not: it took the Polymarket contract for a 4.8% touch from 67.5c to 87.5c in eighteen hours, and it pushed the top of the order book to an 82c bid. Our published gate on that position was to sell on a 66c bid. It fired, and we sold, and the honest fill was 71.69c volume-weighted because there were ten shares at 82 and ten more at 81 and then nothing until 70. That is the whole letter, and the rest of the global board is context for it. Renewed hostilities are the proximate cause: US forces struck Iranian rocket launchers on Larak Island, Tehran hit targets in the UAE and Jordan, Trump extended his threats to Kharg Island - Iran's principal export terminal - and a supertanker caught fire in the Strait of Hormuz after striking two naval mines. Brent trades $91.28 live. On the policy board the direction is uniform: ECB September tightening prices 98.55c, the Bank of Japan 88.5c, and the Fed's September meeting 56.5c for a 25 basis point INCREASE against CME FedWatch at 66%. Three major central banks, three markets, one direction. We hold no position in any of them, and the reason is correlation rather than conviction - a point that is only worth making on days when the correlated trade looks as good as it does this morning.
Today's Market Moves
10Y Touches 4.8% Before 2027 (pos-017)
67.5%→87.5%+20.0pp
CLOSED TODAY. Bid 82 / ask 93, and behind that 82c bid sat ten shares. Hourly: 67.5c at 12:00 UTC Monday, 76.5c at 14:00, 78c at 21:00, 85.5c at 02:00, peaking 88.5c at 05:00 this morning. We can reconstruct the mid but not the historical bid, and we say so rather than implying we watched the trigger print.
10Y Touches 5.00% Before 2027 (check leg)
21.5%→33.5%+12.0pp
Bid 28 / ask 39. Our barrier model on the same inputs - 25.0bp barrier, 83 sessions, 36.44bp horizon sigma - returns 44.6 after the haircut. Gap 11.1 points, a FIFTH consecutive failure, and the first narrowing in the sequence (14.1, 17.0, 17.6, 20.2, 11.1). The pre-commitment written Aug 31 retires the model regardless.
Fed September 25bp increase
51.5%→56.5%+5.0pp
Bid 56 / ask 57 on $15.1M; no-change trades 42.5c. CME FedWatch reads 66% as of Aug 31, the second consecutive fresh reading after nine stale sessions, so futures sit 9.5 points above the prediction market. Below our 10-point bar and a sixth correlated hawkish-rates expression. Refused.
ECB September 25bp increase
98.35%→98.55%+0.2pp
Bid 98.5 / ask 98.6 on $107K. Effectively resolved as a market. ECB minutes reportedly have officials seeing a further increase as likely necessary, and the Bund at 3.3% is the curve agreeing. No edge at 98.55 and nothing to do.
Bank of Japan September 25bp increase
88.0%→88.5%+0.5pp
Bid 88 / ask 89 on $131K; no-change 11.5c. The JGB 10-year touching 3.00% for the first time since 1996 is the market pricing the hike in cash before the meeting. Refused on correlation for a fifth session, which is the honest reason - we like the trade and we will not add a sixth rates-tightening bet.
Fed Rate Hike in 2026 (pos-010)
67.5%→71.5%+4.0pp
HALF THE REMAINDER TRIMMED at the 70c CLOB bid, where the book shows 5,681 shares. The 72c trim gate fired at 75.5c at 17:00 UTC Monday during the US session and we were five hours published. Gamma reports bestBid 71; we executed at the level with confirmed depth and said which.
Screening Table
| # | Market | Expiry | Market Price | Fair Value | Gap (pp) | Direction | Volume | Confidence |
|---|---|---|---|---|---|---|---|---|
| 1 | 10Y Touches 4.8% Before 2027 (pos-017) | Dec 31 | 87.5% | retired% | 0pp | SOLD $25 YES at 71.69c VWAP on the 66c-bid gate - realised +$83.61, and the barrier model that priced it is retired the same morning | $$60.8K on the leg | 5/10 |
| 2 | Fed September 25bp increase | Sep 16 | 56.5% | 66% | +9.5pp | NO POSITION - a 9.5pp gap to FedWatch, under our 10pp bar and a sixth correlated rates bet; refused for a second session | $$15.1M on the leg | 3/10 |
| 3 | Bank of Japan September 25bp increase | Sep 18 | 88.5% | 88% | -0.5pp | NO POSITION - the JGB at 3.00% has priced this in cash; no edge and we would not add correlation for one anyway | $$131K on the leg | 3/10 |
| 4 | 10Y Touches 5.00% Before 2027 | Dec 31 | 33.5% | retired% | 0pp | NO POSITION - this leg is the diagnostic that just retired our barrier model; a nominal 11.1pp gap we decline for a fifth session | $$84.9K on the leg | 5/10 |
| 5 | ECB September 25bp increase | Sep 10 | 98.55% | 98% | -0.55pp | NO POSITION - effectively resolved; the Bund at a fifteen-year high is the same trade in cash | $$107K on the leg | 3/10 |
Top 5 Opportunities
1
10Y Touches 4.8% Before 2027 - a synchronised global selloff, and an order book with ten shares on top — YES
↑ BUY YES0pp
Market price
87.5%
Fair value
0%
Gap: 0pp
This position was a bet that the global long end had further to sell off, bought at 16.5c in July. Overnight it was proved right by three markets at once: the JGB 10-year at 3.00% for the first time since 1996, the Bund at 3.3% for the first time since May 2011, and the US 10-year settling 4.75%, its highest since January 2025 and five basis points from the touch level. The contract went 67.5c to 87.5c and the top-of-book bid to 82c, through our published 66c-bid take-profit. We sold all 151.5152 shares. The fill was 71.69c volume-weighted, because behind the 82c bid there were ten shares, behind the 81c bid ten more, and then nothing until 70c. Proceeds $108.61 against a $25 stake, realised +$83.61. Every share filled above the trigger, which is the difference between an execution and an excuse. What this taught us is worth more than the money. We had been marking this position at the mid; Friday's mid said $132.58 and the realisable value was $108.61. Twenty-four dollars of paper profit on a twenty-five dollar ticket, carried for weeks, invisible until we sold. From today this book marks at the bid.
▵ Bull case
- The 10-year is five basis points from the touch. If it prints, the contract pays $1 and we will have sold the last thirty cents of it - and we will say so.
- A synchronised move across JGBs, Bunds, OATs and Treasuries is a far more durable driver than any single domestic print, and none of it has obviously finished.
- Renewed US-Iran hostilities put Brent above $91 with Kharg Island under explicit threat. An oil shock is the cleanest possible route from 4.75% to 4.80%.
- ECB minutes point to a further increase, the BoJ is 88.5c to hike, and FedWatch has the Fed at 66%. Three central banks tightening into the same curve.
▿ Bear case
- We sold into a mid of 87.5c and got 71.69c. On a liquid leg that gap would not exist; on this one it was sixteen points, and we published a fair value all summer that implicitly assumed we could transact near the mid.
- The barrier model that justified holding this since July is being retired today for failing its own consistency check five consecutive times. Being right here does not make it reliable.
- We were late. The bid crossed the trigger during Monday's US session and we executed the following morning at whatever was left.
- Closing removes a genuinely convex position from a book that now has nothing left with real upside - four positions, none of them cheap, three of them expressions of the same hawkish-rates view.
2
Fed September - a 9.5-point gap to FedWatch that we are refusing for the second session — NO POSITION
↑ BUY YES+9.5pp
Market price
56.5%
Fair value
66%
Gap: +9.5pp
CME FedWatch reads 66% for a 25 basis point September increase as of Aug 31 - the second consecutive fresh reading after a nine-session stretch where we could not verify it at source at all. Polymarket's leg trades 56.5c with a one-cent spread on $15.1M of volume. That is a 9.5-point gap on one of the deepest macro contracts available, in the direction our whole book already leans, three days before an employment report and a fortnight before the meeting. We are not taking it, and the reasons are the same two as Friday. It is below our published 10-point entry bar - by half a point, which is exactly the sort of margin a bar exists to survive. And it would be a sixth expression of hawkish-Fed risk in a book that, after today's close and trim, has four positions and $257 of stake. The correlation cap only costs something on days when the correlated trade looks good, and this morning it looks very good indeed.
▵ Bull case
- 9.5 points on a $15.1M contract with a one-cent spread is as clean a gap as this market produces, and prediction markets have historically converged toward futures-implied odds into FOMC dates.
- Warsh said on the record at Jackson Hole that inflation is still too high and the committee may have work to do. That is the highest-quality input available on this question.
- The global tightening board agrees: ECB 98.55c, BoJ 88.5c, and long-end yields at multi-decade highs in three currencies.
- Brent above $91 on renewed Iran strikes makes the September inflation print harder, not easier.
▿ Bear case
- 9.5 is below 10, and moving a published bar because we like a trade is how a process dies. We would rather miss this than explain that.
- It would be the sixth correlated hawkish-rates bet in a four-position book. That is not a portfolio, it is one idea with six tickets.
- The August employment report lands Friday. Consensus is +60,000 with unemployment ticking to 4.2%, and July payrolls actually FELL 23,000. A weak print between now and the meeting reprices this leg violently.
- FedWatch is read from a third-party report rather than directly from CME's tool, which we could not parse at source this morning. We are comparing our market price to a number we did not verify ourselves.
3
Bank of Japan September - the JGB at 3% has already paid for the hike — NO POSITION
↑ BUY YES-0.5pp
Market price
88.5%
Fair value
88%
Gap: -0.5pp
The Japanese 10-year touched 3.00% for the first time since 1996. That is the cash market pricing a policy change with more conviction than the prediction market, which has the September meeting at 88.5c for a 25 basis point increase against 11.5c for no change. Our fair value is 88 and the ask is 89, so there is nothing here even before the portfolio constraint. What is worth saying is what the 3% print means beyond Japan. Japanese institutions have been the marginal buyer of long-dated foreign paper for two decades; a domestic 10-year at 3% removes the reason to own hedged Bunds or Treasuries at any spread they currently offer. That repatriation channel is a large part of why the Bund printed 3.3% and the US 10-year 4.75% in the same eighteen hours, and it is the mechanism that made our pos-017 close possible. We refuse this leg for the fifth consecutive session on correlation, and we keep saying it plainly: we are not refusing because we dislike the trade.
▵ Bull case
- The JGB 10-year at 3.00% for the first time since 1996 is the market doing the Bank of Japan's work in advance.
- Brent above $91 on Iran escalation is an import-price shock for an energy importer with no domestic production.
- ECB minutes and Warsh's keynote both point the same way; the BoJ moving alone would be the anomaly.
- 88.5c on a contract that pays $1 is a 13% return over seventeen days if it resolves as priced.
▿ Bear case
- Our fair value is 88 and the ask is 89. There is no edge to take before any portfolio argument is made.
- It would be another tightening bet in a book that already carries three of them, and correlation is the risk we have named as our largest for a month.
- The BoJ has repeatedly declined to validate market pricing at meetings where the market was this confident, and 11.5c is the honest price of that history.
- $131K of volume on the leg is thin enough that our own size would move it, and today taught us exactly what thin books do to a fill.
4
10Y Touches 5.00% Before 2027 - the diagnostic that retired our own model — NO POSITION
↑ BUY YES0pp
Market price
33.5%
Fair value
0%
Gap: 0pp
On Aug 25 we published a consistency check: run the same barrier model on the 5.00% leg and see whether it agrees with the market. It did not, and it kept not agreeing - 14.1 points, then 17.0, then 17.6, then 20.2. On Aug 31, before we knew today's prices, we pre-committed: a fifth consecutive failure retires the barrier model from published fair value, leaving it as a diagnostic only. Today the model returns 44.6 against a 33.5c market. That is 11.1 points - a fifth consecutive failure, and the first NARROWING in the sequence. The narrowing is precisely the argument we wrote the pre-commitment to prevent ourselves from making, so we are not making it. The model is retired. It is worth being clear about what the model got wrong. Fixed-volatility barrier models over-price far touches by construction, because they ignore the policy reaction that a move of that size would itself provoke: a 10-year at 5.00% brings a Treasury response, a buyback response and a demand response that the arithmetic has no way to represent. The market has understood that consistently for two weeks and we have argued with it five times. The market won.
▵ Bull case
- A nominal 11.1-point gap is above our 10-point entry bar on paper, and it has narrowed for the first time.
- The 10-year settled 4.75% and the global long end is at multi-decade highs. The path to 5.00% is more credible today than at any point this year.
- The buyback mechanism that produced our 0.9045 haircut expires Nov 4 while the barrier runs to Dec 31 - a two-month window with less official support.
- Brent above $91 with Kharg Island threatened is a genuine tail-inflation scenario, and this leg is the cheapest expression of it.
▿ Bear case
- This gap IS our failing consistency check. Treating it as an opportunity would mean betting on the model in the same session we retired it for being unreliable.
- Five consecutive failures in one direction is evidence about the model, not about the market.
- It would be a further correlated rates bet in a book we have just spent a session de-correlating.
- The spread is 11 points wide - bid 28, ask 39 - and today demonstrated exactly what a wide book does to a real fill.
5
ECB September - resolved in all but name, and the Bund is the same trade in cash — NO POSITION
↑ BUY YES-0.55pp
Market price
98.55%
Fair value
98%
Gap: -0.55pp
98.55c with a one-tick spread is a market that has stopped being a question. ECB minutes reportedly have officials seeing a further increase as likely necessary, and the cash market has already voted: the Bund 10-year at 3.3% is the highest since May 2011, French yields the highest since November 2008, Dutch at fifteen-year highs, Italian and Spanish at multi-year highs. What is interesting is not the probability but the shape of what it confirms. Six weeks ago this leg was an inflation trade. It is now a term-premium and fiscal-supply trade - European yields are rising across the periphery and the core together, which is not how a market prices a central bank fighting demand-side inflation. That distinction matters for how we would read a European surprise, and it is the reason we track this board at all despite never trading it. There is 1.45 cents of upside here for a fortnight of capital, and we have four positions and better things to not do with the money.
▵ Bull case
- ECB minutes point to a further increase as likely necessary - close to explicit guidance from a committee that rarely gives it.
- The Bund at a fifteen-year high and peripheral spreads rising with the core is the curve agreeing in cash.
- Brent above $91 is a euro-area energy shock with no domestic offset.
- At 98.55c a correct resolution still returns 1.5% over ten days, which annualises to something a money-market desk would look at.
▿ Bear case
- 1.45 cents of upside against a 100-cent downside is a risk shape we would refuse anywhere else, and the ECB has surprised before.
- $107K of volume on the leg means our own ticket moves the price against us - the exact problem today's pos-017 fill exposed.
- Another tightening bet in a book already carrying three.
- We have no informational edge on the ECB whatsoever, and pricing a 98.55c contract requires being right about the 1.45c, not the 98.55c.