US close, 12 October
Oil reached the level we published. Yields did not follow.
Brent at US$106.12 is inside the US$105–107 band we have been naming for a week as the point where the inflation channel reopens. The bond market declined to confirm it: both tenors finished flat to lower. Underneath a 0.06% move in the index, technology slipped back into negative gamma and the options market moved its downside marker for semiconductors six per cent lower.
What changed on the day
Where we sit in our risk framework
Divergence
Yields, oil and inflation deteriorate while equities and volatility stay calm.
Structure resilient
Positive gamma and contained credit. The gamma half has failed again.
Controlled de-rating
Twelfth session, and the first in which the failing condition is unambiguous.
Broader deleveraging
Gamma negative across the complex, VIX above 20 and high yield through 325bp.
On Saturday we wrote that the S&P ETF’s gamma was the single input that would decide the stage. It was supplied and it is positive, but the Nasdaq ETF turned negative in the same session, so the test fails anyway. The condition always required all three together, and calling it one input was imprecise. Today’s failure is at least unambiguous rather than unverified.
How the day changed the dashboard
| Indicator | 9 Oct | 12 Oct | Signal | What changed |
|---|---|---|---|---|
| Equities and breadth | ||||
| S&P 500 | 7,811.5 | 7,807 | Watch | −0.06%. Essentially unchanged |
| Nasdaq 100 | 30,883 | 30,665 | Elevated | −0.71%. Twelve times the index decline |
| Equal-weighted S&P 500 | +0.55% | +0.07% | Stable | Still positive. The Dow rose 0.38%, energy 1.44% |
| Semiconductors (SMH) | 603.60 | 589.87 | Severe | −2.11%, through the 590 level. Sector index −2.61% |
| Volatility and dispersion | ||||
| VIX | 14.84 | 15.21 | Stable | +2.5%, and still giving very little warning |
| VVIX (volatility of volatility) | 84.88 | 85.93 | Stable | +1.2%. Modest demand for protection |
| Cboe DSPX dispersion | 36.02 | 36.00 | Elevated | Flat. The gap to the index measure persists |
| Single-stock implied volatility | 38.94 | 39.08 | Elevated | Slightly firmer, at 2.6 times the index measure |
| Dealer positioning (net gamma) | ||||
| SPX | ~+$103.5bn | +$107.6bn | Stable | Firm, and read directly this session rather than derived |
| SPY | — | +$1.5bn | Watch | Supplied, and positive — but only 0.16% above its flip |
| QQQ | +$2.9bn | −$3.8bn | Severe | Negative again, three sessions after repairing |
| SMH | −$199m | −$621m | Severe | Three times worse. Now 3.7% below its flip |
| Macro and rates | ||||
| US 10-year Treasury | 5.244% | 5.237% | Severe | Marginally lower, and not re-accelerating |
| US 30-year Treasury | 5.599% | 5.599% | Severe | Unchanged, despite the move in oil |
| France 10-year | 4.858% | 4.896% | Elevated | +3.8bp. Back to where it was on Thursday |
| Brent crude* | US$104.42 | US$106.12 | Severe | +1.34%. Inside our US$105–107 escalation band |
| Credit | ||||
| ICE BofA US High Yield OAS† | 315bp | 315bp | Elevated | No new observation. Still the 8 October reading |
*Brent is the next ICE Brent futures contract, rolled on each contract's last trading day. †High-yield spreads are shown by observation date; the most recent available remains 8 October, so there is no credit confirmation of the past two sessions. Gamma figures are model-based estimates from third-party options data, taken near rather than at the close.
Options market structure
| Market | Spot | Zero gamma | Spot vs flip | Put wall | Call wall | Net gamma |
|---|---|---|---|---|---|---|
| SPX | 7,807.80 | 7,743.20 | +0.83% | 7,500 | 7,900 | +$107.6bn |
| SPY | 777.50 | 776.29 | +0.16% | 772 | 785 | +$1.5bn |
| QQQ | 746.00 | 748.53 | −0.34% | 745 | 760 | −$3.8bn |
| SMH | 589.87 | 612.34 | −3.67% | 555 | 615 | −$621m |
Spot is the closing level where supplied. For semiconductors the session change implies a close near 590.9 against the 589.87 capture shown, a difference of 0.17% that straddles the 590 level discussed below; we use the capture because the walls and flip are measured against it.
The downside marker moved
Semiconductors did not only fall through 590. The put wall that sat at 590 has migrated to 555, so the largest downside option concentration is now six per cent lower than it was on Friday.
A wall is where hedging demand clusters, and it tends to act as a shelf. When it moves down, the shelf moves with it. That is a more informative change than the 2.11% decline itself, because it describes where the market now expects to defend rather than where it just traded.
A four-step hierarchy
The S&P sits 0.83% above its flip with US$107.6bn behind it. Its ETF is 0.16% above its own. The Nasdaq ETF is 0.34% below. Semiconductors are 3.67% below.
Read top to bottom, that is the market: stable where it is broadest, fragile where it is most concentrated. The index is not holding because nothing is wrong; it is holding because what is wrong is confined.
Oil arrived, rates did not
Brent at US$106.12 is inside the band we published as the point where the oil-to-inflation-to-yields chain reopens. The chain's second link did not respond: the 10-year eased slightly and the 30-year was unchanged.
That divergence is the most useful thing in the session. Either the bond market is discounting the oil move as supply-driven and temporary, or it is slow. We would treat US$107 on Brent together with the 10-year back above 5.30% as the combination that matters, rather than either alone.
Credit has gone quiet
There has been no new high-yield observation since 8 October, so the series still reads 315bp and tells us nothing about the past two sessions. The sequence before it ran 303, 309, 315.
That matters more than usual now, because credit is the condition that would distinguish a sector problem from a financing one. The next print is the most consequential number on our board, and we are waiting on it rather than guessing.
Where the stage-two test stands
| Condition for stage two | Status | Where it stands |
|---|---|---|
| No credit reading above 320bp across three observations | Met | 303, 309 and 315bp, though none newer than 8 October |
| At least one observation below 310bp | Met | 303bp and 309bp both qualify |
| Positive gamma across the S&P, its ETF and the Nasdaq ETF | Failed | Two positive; the Nasdaq ETF is −US$3.8bn |
| VIX below 18 | Met | 15.21, up slightly but far inside |
For the first time in this sequence the failing condition is plainly failing rather than unverified, which at least makes the position easy to state. Three sessions ago the Nasdaq ETF was the structure that had repaired; it has now broken twice in five sessions, and a level reclaimed and lost twice is better described as a boundary the market is oscillating around than as a repair.
We would also flag that the credit conditions are being carried on readings up to five days old. They are met on the evidence we have, and we are not treating a stale pass as equivalent to a fresh one. If the next observation prints above 320bp, two of the four conditions fail at once.
AI compute fundamentals overlay
Unchanged from the 7 October reading: frontier coverage 2.03× against our 1.20× level, down 12.6% from 3 October. The overlay remains supportive, on watch.
| Generation | Rate | Coverage | Read |
|---|---|---|---|
| Oldest frontier | US$2.85/hr | 1.19× | Sitting on our 1.20× level |
| Middle generation | US$5.29/hr | 2.03× | The driver of the decline since 3 October |
| Newest generation | US$7.56/hr | 2.38× | Softer, though only 3.2% below its 3 October level |
Three sessions of semiconductor weakness now sit alongside a compute reading that has eased, options positioning three times more negative than Friday, and a downside marker six per cent lower. We have argued throughout that a fall in rental prices is not a fall in demand, and the second benchmark we follow still does not corroborate the decline. That argument is intact on the fundamentals and increasingly lonely against the price.
Tonight's board
- QQQ put wall / flip745 / 748.5
- SMH repair zone612–615
- SMH new put wall555
- SPY flip / put wall776 / 772
- SPX gamma flip7,743
- BrentUS$107
- US 10-year5.30%
The combination that matters
Any one of these alone is noise.
- Brent aboveUS$107
- with the 10-year above5.30%
- and high yield above325bp
- and the S&P ETF below776
How we see the next one to five days
The stage-four sequence is unchanged. We want these together rather than one at a time:
These weightings are our current assessment and will change with the data.
What this means for our outlook
The index did almost nothing and the structure beneath it moved a long way. Semiconductors fell 2.11% through the 590 level that had been their downside marker, and that marker has since moved to 555 — the options market has relocated where it expects to defend, six per cent lower. The Nasdaq ETF slipped back into negative gamma three sessions after repairing. Set against that, the S&P held, the Dow rose, energy gained and the equal-weighted index was positive.
Oil reaching US$106.12 puts it inside the band we named a week ago as the point at which the inflation channel reopens. The interesting part is what did not happen: the 10-year eased and the 30-year was unchanged. The bond market is either reading this as supply-driven and temporary, or it is slow, and we do not know which. We would not treat oil at these levels as decisive on its own, and have said what would make it so.
Stage three holds, and for once the reason is simple: the gamma condition plainly fails. We said on Saturday that the S&P ETF's reading was the single input that would decide it. That reading arrived and was positive, but the Nasdaq ETF turned negative in the same session, so the test fails regardless. The condition always required all three together, and calling it one input was imprecise of us.
What we are most conscious of is what we cannot see. There has been no new credit observation since 8 October, so the two conditions that currently pass are passing on week-old data, and credit is precisely the series that would separate a contained sector problem from a financing one. We would rather say that plainly than let three green marks on page four imply more comfort than the evidence supports.

