Monday’s oil bought 10 points of September pricing. Tuesday’s bought nothing at all, and we would rather like to know why.
⚓ Weathervane: The tightening is arriving and nothing is transmitting it. The whole curve rose through July, the thirty-year to a nineteen-year high, and the S&P still closed the month up. What actually reprices now is attention rather than risk: premiums leave oil, volatility and equities the moment headlines stop arriving, whether or not anything on the ground has changed. Multiples answer to the news cycle, and the bill for that arrives in one session rather than gradually.
Ahoy there, Trader! ⚓️
It’s Phil…
If Brent at $90 could not move September pricing, what is left for CPI to move?
Here is the version of Tuesday most desks will have written up. Oil rose on Middle East tension, equities slipped, everyone awaits the inflation number. True, and it explains nothing, which is roughly where we always start.
Walk it properly and it gets stranger. On Monday, Brent rose about 4.3%. The whole US curve rose with it and CME September hike pricing went from roughly 42% to roughly 52%. Ten points. The chain was legible: dearer barrels, more inflation next quarter, more reason for a Fed chair who has tied September to inflation readings to go.
On Tuesday, Brent went further. It printed $90.03, the highest of the cycle, before settling near $88.91. Same input, larger dose. September pricing finished at roughly 49.9% against 50.1% for a hold, 2 points the wrong way. The 2-year sat near 4.2%. The dollar moved 0.005. Volatility fell.
So the chain that worked on Monday was not connected on Tuesday, and nobody unplugged it in public.
The textbook here is unusually confident, which is why the gap is worth sitting with. Energy is an inflation input. A sustained crude repricing should lift breakeven expectations, and those should show up first at the front of the curve, because that is where policy lives. The tape gave us the opposite: movement at the very back, the 30-year adding 3 basis points to 5.2790% near a 20-year high, and nothing at the front. That is a term-premium story. The market is charging more to lend for 30 years whilst declining to change its view on what happens in 5 weeks.
There is a third detail we cannot make fit, and would rather flag than tidy away. Tuesday was the session in which 6 people were killed on a cargo ship in the Bab el-Mandeb, the first commercial-shipping deaths of this war. Crude gave back its high that same session. The absence of a deal was worth more to the barrel than a fatal attack. We have watched this for 29 cycles and it still reads backwards.
Which leaves us where the common view and the tape part company. Most desks treat this morning’s print as decisive, and the pricing agrees: 50/50 is a market saying it does not know. But an instrument that just shrugged at a cycle-high barrel is not obviously one that will jump at a 0.1% monthly headline. It might. We would only note that the last 48 hours cut the other way, and that if CPI lands in line and September still moves 8 points, we do not understand what moves it.
Phil’s Musing
The bit I keep coming back to is that we were right about the barrel and learnt nothing. Six sessions of saying the premium comes back without anything being signed, and it came back to the tick, and the front end just sat there. I would rather have a mechanism I could rely on than a call I got lucky on, and today I have the second one.
Happy trading,
Phil
Less Brain, More Gain
…and may your trades be smoother than a cashmere codpiece
P.S. – Phils Footnote I set out expecting to write about oil and ended up writing about a wire that came loose. That is usually the sign a week is more interesting than it looks. I do not know which of Monday and Tuesday was mispriced, and I have deliberately not picked one, because the last fortnight has been a long lesson in what happens when I decide a single session is a rule. Ask me again after 08:30.
🗂 Desk Notes | Wednesday, August 12, 2026
Raw briefing export. Observations, not trades.
0. Close-to-open read
SESSION BRIDGE: prior session (Tuesday) full reaction: S&P 500 -0.32% to 7,728.20, Nasdaq Composite -0.60% to 26,445.45, Dow -0.34% to 53,791.85, Russell 2000 +0.32% to 3,027.12, VIX 15.30 -1.16%, Brent intraday $90.03 then close near $88.91, overnight tail added CoreWeave +12% and Super Micro +8.79%; live premarket ES +0.19% / NQ +0.53%, VIX 15.30 unchanged; threshold: none.
Threshold note, third consecutive session. The §8.4.1 test is calibrated on S&P futures. Tuesday’s equity move was 0.32%, comfortably silent. The session’s actual event was a 2.6% intraday round trip in Brent through the highest print of the cycle, plus the first commercial-shipping fatalities of the war. The equity gate has now failed to register the day’s real move on 3 consecutive sessions. Logged again for MIP 11.0 as a candidate widening of §8.4.1 to include a commodity leg. The carry-over fires on that basis rather than on the equity number.
Carry-over (§10.6). Monday’s 5% crude session bought 10 points of September hike pricing. Tuesday’s session took Brent higher still, to $90.03, and September pricing fell roughly 2 points. Same input, larger dose, opposite sign. The carry-over is not that oil rose; it is that the channel which carried oil into the front end on Monday was not carrying anything by Tuesday.
1. The mechanism read (full)
What moved.
- Front end, 2-year. Sits near 4.2% against Monday’s 4.241% close. Roughly unchanged to marginally lower on a session in which crude printed a cycle high. This is the number that matters today and it is a live variable; the exact Tuesday constant-maturity close needs verifying before publication.
- Long end, 10-year. Closed 4.699% against 4.705% Monday, having touched 4.7334% in early trade. Round trip, net flat.
- 30-year. Rose more than 3 basis points to 5.2790%, from 5.251%. Near a 20-year high. This is the only leg of the curve that actually moved, and it moved on term premium, not policy.
- Dollar, DXY. 99.812, up 0.005, which is as close to literally unchanged as the index gets. Commentary elsewhere describes the index as forming a near-term bottom. Not confirmed by the tape.
- Volatility, VIX. 15.30, down 1.16% from 15.45. Vol fell into a CPI print, on a session containing a cycle-high crude print and the first fatal attack on commercial shipping of the war.
What it implies.
The curve steepened at the very back and did nothing at the front. That combination is a term-premium repricing, not a policy repricing. Monday was the reverse: the whole curve rose together and September pricing moved 10 points. Tuesday delivered a larger version of Monday’s input and produced none of Monday’s response at the front, some of it at the back, and a lower VIX.
The equity composition supports the same read. A 0.32% S&P decline against a 0.32% Russell gain, with 1,256 small-cap advancers and Alphabet down 3.61%, is not a macro de-risking. It is concentrated hedging in the names that carry index weight, ahead of a scheduled event. Breadth was fine. Vol was lower. Nobody was selling the economy; a handful of desks were buying insurance on the top 10 lines of the index.
The one artery.
The crude-to-front-end transmission ran on Monday and was dead by Tuesday. Brent at $90.03 could not move September pricing. That is the mechanism, and it is the thing that makes today’s CPI genuinely hard to handicap: the instrument that is supposed to absorb the print has just demonstrated that it will ignore a larger input than the print is likely to be.
The honest caveat, stated up front. This is 1 session. The failure mode this ledger has logged 3 times in a fortnight is converting a single session into a mechanism. Tuesday is evidence that Monday’s channel is not reliable. It is not yet evidence that the channel is broken.
2. Forward catalyst slate
| When | What | Why it is on the slate |
|---|---|---|
| Wed 12 Aug 08:30 ET | July CPI. Consensus 3.4% headline y/y, 2.5% core; WSJ survey of 15 banks at 0.12% and 0.22% m/m. Kalshi under 55% for above 3.3%. | The Part 185 falsifier resolves on this print, close to close. September pricing sits at 49.9% hike / 50.1% hold. |
| Wed 12 Aug PM | Cisco and Cerebras report. | Cerebras is early in its public life; guidance quality matters more than the print. |
| Thu 13 Aug | July PPI. | Second leg of the inflation pair. A soft CPI followed by a firm PPI is the awkward outcome nobody is positioned for. |
| Thu 13 Aug | Applied Materials, options pricing a 10.39% post-earnings move. Tencent also reports. | Widest implied move on the calendar. Feeds the chipflation and China supply-side threads. |
| Thu 13 Aug | Korea amplifier falsifier window closes. | Resolved early. See ledger Part B. |
| Fri 14 Aug | Part 184 falsifier window closes (2-year versus crude). | Needs a 2nd qualifying Brent session of more than 2% or it resolves inconclusive. |
| Sun 17 Aug | Alibaba earnings; Qwen3.8-Max open-weight release still pending. Hashdex ETF trading ends. | China open-weight thread. |
| Late Aug | Pentagon 21-day production deadline resolves. | Defence industrial base thread; unchanged. |
| Sept 15-16 | FOMC. Procedural CLARITY vote 15 September. | The meeting everything above is being priced into. |
3. Divergence flags
- The headline flag. Crude made a cycle high and the front end did not respond. Brent $90.03, September pricing down roughly 2 points. Monday’s identical-but-smaller input moved it 10 points up. One of those 2 sessions is mispriced and there is no way to tell which from here.
- Physical escalation coincided with the giveback. The first commercial-shipping deaths of the war, a double-tap missile strike killing 6 in the Bab el-Mandeb, plus a US Navy helicopter disabling a blockade runner with Hellfire missiles. Brent’s response to all of it was to retreat from $90.03 toward $87.51. This is cycle 29 of the anti-pricing pattern and the cleanest instance yet: the absence of a deal was worth more than a fatal attack.
- Vol fell into a scheduled event. VIX down 1.16% to 15.30 on the eve of a print that both sides of a 50/50 rate market claim is decisive. Either the print is genuinely expected to be a non-event, or hedging has migrated entirely into single-name and index options and the headline gauge is no longer where the protection is.
- The equity split is an exact mirror and probably means nothing arithmetically, but the composition is real. Megacap tech down, 1,256 small caps up. Rate-sensitive domestic names outperforming on a day the inflation input rose is backwards on the textbook and worth watching if it repeats.
- Money is arriving into bitcoin without price moving. $853.5m of spot ETF inflows in the week to 7 August, the best since April, against 1.79m coins of overhead supply between $62,000 and $65,000 and $34.4m of long liquidations Tuesday. Flow is not clearing the shelf.
- AI capital formation, third observation. CoreWeave rose about 12% after hours on a $104bn backlog against a $48.1bn market value. Intel’s raise, upsized from $15bn to $20bn, closed lower again. That takes the AI-capital-raise watch thread to 3 observations, one of them positive. The stated threshold before this counts as a pattern remains 4. It is not there yet and today is not the day to say otherwise.
4. Regime status and sensitivity
⚓ Weathervane: carried unchanged. Rewritten 3 August. Next rewrite candidate remains Monday 17 August. Tuesday is mixed for it: the banner’s attention-not-risk clause is confirmed on crude (a non-event outpricing a fatal attack) and challenged on rates (the same attention input produced no repricing at all). The banner survives; it does not strengthen.
⚓ Regime Flags: duration respec counter, Korea, chipflation and vol regime all detailed in ledger Part C. No flag hardens today.
Sensitivity read (§17.3): HIGH-TIGHT, hold at yellow. No change recommended, and today is the clearest vindication the setting has had. On 11 August the tight setting refused to mark the Part 181 crude leg on a WTI quote and an assumed spread, explicitly because inferring a Brent print to bank a HIT is the kind of convenient arithmetic this ledger exists to prevent. Roughly 12 hours later Reuters reported a confirmed $90.03 print and both open crude shots resolved cleanly on their stated terms. The refusal cost nothing and bought a clean mark. Against that, the same setting today forces a MISS on the Korea shot at the first qualifying session, and refuses to count CoreWeave as a 4th AI-capital observation in our favour. Held in both directions.
Public tell: not triggered. “Hoist the mainsail” stays holstered.
Desk Notes are observations, never trades.