Delta Gamma Vault Weekly Journal
Oct 11, 2026
Records, and the leadership quietly changed hands
The S&P and Nasdaq both set closing records. Underneath, semiconductors were sold, utilities and staples led, and the ten year pushed above 5.3% for the first time in more than two decades. The index went up. Almost nothing about what was driving it stayed the same.
| +1.15%S&P on the week | 5.3%+10-year, 2-decade high | 19%October hike odds | 7.4%30-yr mortgage rate |
The S&P rose from 7,722.72 to 7,811.54, up 1.15%. The Dow added 0.9% and the Nasdaq 0.6%. Both the S&P and the Nasdaq set new closing records along the way.
And yet the week's sector table does not look like a record week at all.
Utilities, staples, energy and real estate led. Technology was flat and communication services fell. For a month the story has been a narrow tape carried by megacap tech. This week the index rose with tech contributing nothing.
That is either healthy broadening or defensive rotation at a high, and the honest answer is that one week does not tell you which. What it does tell you is that the explanation that worked in September stopped working in October.
How the week unfolded
Two records, an AI scare, and a bond market that would not cooperate.
| Mon 10/5Records |
The Nasdaq Composite closed at a record 27,477.31, up 1.05%, blowing past its September 22 high. The Nasdaq 100 set a record of its own. Nvidia and Broadcom both added 2% and the technology sector posted a record close. The ten year touched 5.32%, levels last seen in 2002, and the thirty year breached 5.7%, a twenty-four-year high. Records on one side of the screen, a breaking bond market on the other. |
| Tue 10/6Above the walls |
Both indices traded above their call walls for the first time in weeks, which meant the dampening thinned out and what sat overhead was combo levels rather than a wall. Speculators had added 114,848 contracts to the five-year Treasury net short, taking it to 995,701. The ten-year net short reached 900,615. Nearly a million contracts on one side of the trade. |
| Wed 10/7Minutes |
The September meeting minutes landed at 2:00. They covered a meeting held before the payroll miss that broke the hawkish regime, which made them stale on arrival. The NQ put wall jumped 2,200 points to NDX 30,200, the largest single relocation of the stretch. I wrote that one session is noise until it holds and did not build the bias on it. |
| Thu 10/8AI scare |
A Financial Times report put OpenAI's annualized revenue roughly $20 billion below earlier estimates. Semiconductors sold off — Nvidia down about 1.8%, AMD about 4%, Micron about 4.3%, with Intel, Seagate and Western Digital all down nine to twelve percent on the week. Trump posted that the US will not attack Iran before the November 3 midterms, while keeping the blockade in force. Oil had surged earlier in the session on reports the White House had asked the Pentagon for strike options. The put wall was back at NDX 28,000. One session, exactly as the rule said. |
| Fri 10/9Rebound |
A later Bloomberg report said OpenAI expects revenue of $70 billion or more by year end, which partly reversed the chip concern. Oil fell on the Trump post. Brent finished near $104.72, up about 2.4% on the week. Refined fuel rose faster than crude, which lifted the refiners — Marathon up about 7.4%, Valero more than 6.5%, Phillips 66 about 5%. |
Two maps, two completely different weeks
The most useful thing on my screen this week was not a level. It was how differently the two structures behaved.
Eight observations per index across five sessions. ES relocated twice, both times stepping the call wall upward in an orderly ratchet. NQ relocated six times, including a put wall that alternated between two strikes on four consecutive days.
The ES map was well behaved. The call wall ratcheted 7800 to 7850 to 7900 and then held, which confirmed it. The put wall sat at SPX 7500 for all five sessions without moving a single point. That is a structure you can trade against.
The NQ map was close to unusable. The put wall alternated 28,000, 28,000, 30,200, 28,000, 30,200. The call wall ratcheted up to NDX 31,500 and then collapsed to 29,475, below spot, on Friday.
An alternating wall is not relocation. It is two strikes carrying near-identical open interest, with small daily changes flipping which one registers as largest.
Same pattern I identified for the ES call wall oscillating between SPX 7800 and 8100 around September expiration. The useful response is to treat both strikes as real floors rather than describing each flip as a 2,200-point move.
The rule I published on September 24 — that a single-session relocation is noise until it holds a second session — prevented two bad reads this week, on Wednesday and again on Thursday. I also amended it on Thursday to allow for a consistent one-direction drift, and then on Friday the ES call wall confirmed under the original rule without needing the amendment. The amendment may still be right. It was not needed this time, and I would rather say that than take credit for it.
The liquidity question, properly answered
A member asked about Fed injections and bond buybacks. The real answer is more interesting than the framework one I gave midweek.
There are two separate programs and they are pulling in opposite directions.
The Fed is paused
The New York Fed's reserve management purchases — bill buying intended to keep bank reserves from getting scarce — have been suspended for two consecutive monthly windows. The current schedule covers September 15 through October 14 with approximately $15.6 billion of reinvestment purchases and no reserve management purchases at all.
Bank reserves sat at $3.04 trillion as of September 9, above the year-to-date average of $3.01 trillion and up from $2.85 trillion at the end of last year. The pause reads as policymakers being comfortable with that level rather than as stress.
The street is split. Wells Fargo and Bank of America expect purchases to resume in October as bill issuance rises. Barclays looks for roughly $10 billion in October rising to $20 billion in November. Citigroup argues reserves have entered a lightly abundant regime and expects the pause to run through year end.
The Treasury is buying
Separately, and more actively, Treasury doubled the maximum size of its liquidity support buybacks in the long end, from $2 billion to at least $4 billion per operation, covering the 10-to-20-year and 20-to-30-year sectors. That change took effect September 9 and runs through November 4.
Last week had two of them. October 6 in the two-to-three-year bucket at $4 billion, and October 8 in the twenty-to-thirty-year bucket at four billion or more. The week before had October 1 in the ten-to-twenty-year bucket.
Treasury bought back long-dated paper at double the usual size on two separate days last week, and the ten year still pushed above 5.3% to a two-decade high while the thirty year reached 5.64%.
That is the signal worth carrying. When an active buyer at the long end cannot hold yields down, the pressure is coming from supply and from inflation expectations rather than from a liquidity shortage. Buybacks improve market functioning. They do not change the direction of rates.
So the honest framing for members: the Fed is neutral and paused, Treasury is adding liquidity at the long end, and the long end sold off regardless. Repo operations and buybacks are plumbing. They tell you about market functioning, not about direction, and a spike in standing repo usage would be a stress signal rather than a bullish one.
Marking the week
Tuesday: both indices above their call walls, structure thin overhead. No pocket triggered cleanly.
Wednesday: flagged the put wall jump as unconfirmed and did not trade it. Correct — it reversed the next day.
Thursday: the ES lower pocket triggered on losing the 7822 volatility trigger and the overnight low at 7804. Both went, and the prior-day low printed 7783, three points short of the first target at 7780. The NQ lower pocket triggered on losing 31118 and paid two targets, 31000 and 30855, with the low at 30790.
Friday: the double top I flagged at 7858.50 and 7859.50 was never challenged — the prior-day high came in at 7851 and price went the other way.
Two corrections also belong here. I had the October FOMC as next week when it is the 27th and 28th, and on two mornings my news sourcing came up empty, which I flagged in the plans rather than filling the gap with something confident.
The week ahead
The densest calendar since the September FOMC.
Monday is a half market — bond desks closed, equities open. Wednesday and Thursday carry almost everything that matters.
Wednesday is the one
September CPI at 8:30. Consensus is around 3.7% headline against 3.4% in August, with core near 2.5% against 2.4%. That is an expected acceleration, and it lands into a market that has priced October hike odds down to roughly 19% from 51% a week earlier.
Those two things do not sit comfortably together. If headline accelerates as expected and core firms, the October contract has a long way to travel back up. December already prices around 70%, so the market is betting on timing rather than on no hike at all.
The Beige Book follows at 2:00, and the New York Fed's current no-purchase window ends the same day, which means the next monthly schedule and any decision on resuming reserve management purchases arrives around then.
Thursday is four releases at once
PPI, retail sales, Philadelphia Fed and Empire State all print at 8:30. Then a Treasury buyback in the ten-to-twenty-year bucket at 1:40, TSMC earnings, and Warsh speaking. Any one of those would be a day. Together they make the 8:30 bar effectively untradeable until the dust settles.
Earnings season starts
JPMorgan, Goldman, Citi and Wells Fargo report Tuesday. Bank of America, Morgan Stanley and BlackRock on Wednesday. With the long end at two-decade highs, net interest income and credit quality matter more than the headline numbers, and bank stocks have been pressured by exactly that.
What I am watching in the structure
Whether the ES put wall at SPX 7500 finally moves. It has not budged in five sessions. That is the most stable reference on either board, and a move there would be genuine information rather than noise.
Whether the NQ maps settle down. Six relocations in a week makes that structure unreliable. Until the put wall stops alternating, ES is the cleaner instrument and I will be weighting it accordingly.
Whether the rotation sticks. If utilities and staples lead a second week while tech lags, that is a defensive bid rather than a broadening, and it would change how I read every record high from here.
And oil. Trump has effectively put a political floor under the war timeline through November 3 while keeping the blockade in place. That removes escalation headlines for three weeks but leaves supply constrained. Brent still finished the week up 2.4% at around $104.72.
The index set records while its leadership changed hands, the Treasury bought back long bonds at double size and yields rose anyway, and the Nasdaq's options map stopped being reliable enough to trade.
Level maps, triggers and tiered targets for both ES and NQ post before every open in the Vault.
Nothing here is investment advice. Futures trading involves substantial risk of loss.
Stay connected with news and updates!
Join our mailing list to receive the latest news and updates from our team.
Don't worry, your information will not be shared.
We hate SPAM. We will never sell your information, for any reason.
