← Back to All Reports
DAILY REPORT · MONDAY 09/28 CLOSE · DAY ONE OF THE QUARTER-END AND PAYROLLS WEEK · THE 10-YEAR CLOSED AT 5.24% AND THE RATE CHANNEL REACHED CREDIT · THE AI COMPLEX FELL AS A BLOCK WHILE NVDA ROSE ON ITS BUYBACK · SMALL CAPS CLOSED UNDER THEIR MONTHLY BAND · PCE, QUARTER-END, MU, ISM AND PAYROLLS STILL AHEAD

Daily Report — 09/28/26 · “Credit Joined the Rate Trade”

The S&P 500 fell 0.77% to 7,683.69, closing 11 points under its daily band (7,695.01) but still inside its weekly, monthly and quarterly bands. The Iran headline reached stocks through the 10-year yield, which closed at 5.24%, not through oil, which never left its band. Three things broke underneath the modest index loss: the high-yield spread closed above 3% for the first time since April, the small-cap fund IWM closed under its September monthly band, and the AI stocks fell as a group while NVDA rose on a $150B buyback. This edition reads what Monday did to the four timeframes going into Wednesday’s quarter-end, what the options tape did with its protection, and who among the weekend’s commentators the tape agreed with. It grades Friday’s edition and lists the structures the framework would follow, each with its vol, its level and its exit. The options dashboard for Monday was not captured, so the panel reads are replaced by a decomposition of the day’s options file, labelled as such.

Monday in one line — credit joined the rate trade

The S&P 500 (SPX) closed at 7,683.69, down 0.77%, and the loss was a rate event, not an oil event. Before the open, the headline tape was Iran rejecting a Hormuz proposal, oil and yields up, and a 10-year yield reported above 5%. By the close only half of that had shown up in prices. Ten-year note futures sat under their daily band at every read of the day and closed at 104.50 against a lower line of 104.58. Crude never stretched: the front contract settled at 93.51, inside its daily band. The 10-year yield closed at 5.24%, up from about 5.18% on Friday. The breakeven inflation rate did not move (2.34%), so for a fourth session the rise was in real yields, the part of the rate that tightens financial conditions directly.

The index loss was small; three breaks underneath it were not:

The mood gauge felt it more than the index did. The FOM sentiment index fell 11.8 points to 37.9, its first “fear” label since the September 16 Fed meeting, on a day the index lost less than 1%.

TAPE SPX 7,743.41 (Fri) to 7,683.69 (Mon) · 10-year 5.24%, breakeven 2.34% flat · 10-year note futures 104.50 vs daily lower 104.58 · crude 93.51 inside its band · HYG 77.54, high-yield spread 3.02% · IWM 280.02 vs monthly lower 281.09 · AI basket 2 up / 31 down, average -3.42%

The regime under it — real yields did it, and oil stayed out of it

The Fed is still tightening, and the long end is doing more of the Fed’s work every session. The Fed hiked on September 16, and the commentary layer treats another hike on October 28 as live. Real yields sit about 49 basis points above the breakeven rate. At a 5.24% 10-year, the multiple the rate model allows for the index is about 10.8 times earnings, and the high-multiple names are the first to feel that cap. That is the order Monday showed: the AI and software names fell hardest, while staples and health care closed with more risers than fallers.

The dollar keeps the metals closed. The dollar index closed at 101.20, well above 100, with a wide trend range. That is the framework’s strong-dollar block on bullish metals, and it stayed in force the day GLD fell 3.94% to 377.91, a move of nearly four daily standard deviations. The dollar did not have to move for gold to fall. The real-yield rise alone raised the cost of holding an asset with no yield.

Credit has moved from a watch to an input. Friday’s edition called HYG’s weakness duration, the bond fund tracking Treasuries at about a third of their move. From Monday that is no longer the whole story: the spread over Treasuries widened by 36 basis points in five sessions, and the forward zone sheet now reads HYG’s trend as reversed. This is not a hard block on anything; the framework treats credit as a sizing input. It is the first time this month the rate shock has reached the part of the market that funds companies rather than the part that discounts them.

TAPE Fed 3.75-4.00% after the 09/16 hike · 10-year real yield 2.83% (09/25) vs breakeven 2.34% · fair multiple about 10.8x at 5.24% · dollar index 101.20 · GLD 377.91 (-3.94%) · high-yield spread 2.66% (09/21) to 3.02% (09/28) · HYG 77.54

The four timeframes into Wednesday’s expiry — the daily band broke, the swing bands held

The index broke its daily band and nothing longer, which makes Monday a test for Tuesday, not a change of trend. SPX closed 11.32 points under its daily lower band (7,695.01). It is inside the weekly band (7,635.60 to 7,851.22), sits almost exactly in the middle of the September monthly band (7,428.73 to 7,943.55), and is inside the quarterly band (6,929.27 to 8,069.45). The swing timeframes outrank the daily one, and they say the week’s frame is intact. The daily break is the kind the framework’s own record says usually reverts: index closes outside the daily band have come back inside the next session 80% of the time and within the week 70% of the time, on only 11 cases, all in one bull regime. It is graded on Tuesday’s close back inside Tuesday’s band, not on anything intraday.

Tuesday’s daily lower band (7,641.01) sits only 5.41 points over the weekly lower (7,635.60). An ordinary down Tuesday therefore tests the weekly band itself, and a close under it would be the first swing-frame break of the week. Silva’s gamma flip at 7,665, the level under which dealers stop damping moves, sat 18.69 points under Monday’s close. The forward zone sheet put the index exactly on its zone middle (7,683.54), with the zone low at 7,555.78 and the high at 7,811.29.

The month and the quarter expire Wednesday, and the lines that bind are on the funds, not on SPX. Two of them matter:

No index is at or beyond its quarterly band, so there is no longest-timeframe signal of a regime change. The one board name outside its quarterly band is the long bond, TLT, at 78.62: under its quarterly lower (82.50) and four cents over the quarterly two-standard-deviation lower (78.58). It will finish the quarter outside its band. The trend on its zone sheet is reversed, and the input that moved is the real yield, so this is a continuing trend, not a setup for a bounce.

TAPE SPX daily 7,695.01 / 7,791.81, weekly 7,635.60 / 7,851.22, monthly 7,428.73 / 7,943.55, quarterly 6,929.27 / 8,069.45 · Tuesday daily 7,641.01 / 7,726.37 · IWM monthly lower 281.09, close 280.02 · QQQ quarter anchor 736.40, close 736.53 · TLT quarterly lower 82.50, two-sigma 78.58, close 78.62

The mega-caps and the roster — one name up on its own event, the rest giving back

NVDA (228.86, +1.68%) closed just over its daily upper band (228.81), inside its weekly band (217.17 to 232.97) and on its zone at 211.48 / 223.12 / 234.76. It printed over its weekly upper intraday and closed back inside. The $150B buyback authorization is capacity to repurchase over time, not executed buying, but it is a standing bid under the stock, and the options tape saw one buyer take about 13,000 March 2027 260 calls in the first hour. No reason to fade an upside break on a corporate event.

TAPE NVDA 228.86 (+1.68%), daily upper 228.81, weekly 217.17 / 232.97, zone 211.48 / 223.12 / 234.76 · Mar-2027 260 calls ~12,970 bought at the ask, 36.4 implied vs 37.7 at the money (0.97x)

META (715.62, -4.79%) made its first close under the weekly lower band (716.06), the line Friday’s edition named as its break line. It sits in the lower half of its zone (649.32 / 727.08 / 804.84). One close is a warning: a second close under 716.06 confirms the price break, and a close back over the zone middle (727.08) repairs it. The flow does not show a bearish setup. Holders are still selling calls against a stock that ran 25% in September, which caps it. Any bear case here is price-only. The day’s catalyst was META’s own enterprise-AI push: it hired MongoDB’s chief executive, and MDB fell 18.46%.

TAPE META 715.62 (-4.79%), weekly lower 716.06, zone 649.32 / 727.08 / 804.84, September monthly two-sigma upper 663.82 still under it · Oct-02 at the money 47.9 vs December 41.2 (1.16x) · MDB 334.68 (-18.46%)

MSFT (509.22, -1.35%) held its breakout on a second close, above the 505 to 508 top of its triangle and above Monday’s daily lower (507.91), after trading as low as 502.22. The zone is 489.97 / 502.86 / 515.75 and the weekly band 500.25 to 532.09. A hold on a closing basis, not a clean one, on a day META’s enterprise push was the headline against it.

TAPE MSFT 509.22, low 502.22, daily lower 507.91, weekly 500.25 / 532.09, zone 489.97 / 502.86 / 515.75 · Dec-18 520 call 29.77 implied vs 29.81 at the money (1.00x)

The rest of the roster, placed on its bands:

TAPE AAPL 338.40 vs monthly upper 335.12 · AMZN 246.15 · TSLA 357.45 vs weekly lower 353.57 · AMD 607.87, INTC 116.03 (monthly +2.57 and +1.14 sigma) · MU 1,053.98 · NFLX 69.23 vs 69.55 · MSTR 157.14 vs 156.28 · TSLA Jun-2029 340 / 630 collar 6,500 opening, structure

The options tape — protection moved out in time, not added

Monday’s options dashboard was not captured, so this section reads the day’s full options file instead. The panels that exist only as images (the dealer book by expiry, the flow map, the zero-day gamma chart) are missing for this session. What follows comes from decomposing every print in the file, and it is labelled as a reconstruction, not a substitute for the panels.

After removing re-reports, stock-substitute contracts and paired legs, the equity options tape was flat. Its directional residue was -$5.1M on $3.37B of sided premium: more calls sold than bought, more puts sold than bought, netting to nothing. The raw headline figure was -$514M, and it flipped sign once the structures came out. Two in-the-money put sales (INTC and SNDK) carried the positive side; without them the residue is about -$59M spread thinly across QCOM, RCL, CRWD, MRVL and SNOW. Of the calls sold, 86.9% were holders cashing or writing against existing gains, against a trailing move of -1.56%. That is income and closing, not the harvest of a run, and the notional of calls sold fell for a third session.

The day’s largest index prints were hedges carried forward in time. At 10:23 Central one institution sold 191,000 QQQ December 660 puts and bought 191,000 March 2027 670 puts in the same second, a matched two-leg roll worth about $12.8B of QQQ put notional. At 10:10 Central the same shape printed in SPY: 118,000 December 660 puts sold against 61,840 March 2027 700 puts bought. Both closed nearly all of their December lines. This is protection kept and extended through the first quarter of next year, bought at a lower skew ratio than the one sold. It is not new fear, and because the legs are matched, the structure carries no directional reading.

The payrolls week is not where the book is positioned. The October 2 expiry netted within a few million dollars across SPX, SPY, QQQ and IWM for a second session. The October protection that was added went into small caps: 52,200 IWM October 16 puts at the 272, 275 and 279 strikes, bought above the ask in one second at the weekly lower band. Friday’s QQQ October 2 725 line reads 34,290 calls against 29,959 puts in open interest, a 1.14x ratio. At that parity it is a box spread or conversion, not a wall. The genuinely one-sided lines are QQQ October 2 730 puts (18.2x puts to calls, sitting on the weekly lower), SPY October 16 750 puts (7.52x) and IWM October 16 280 puts (8.66x): customer protection at the weekly bands, which leaves dealers short that downside if it is tested.

The zero-day book pinned. SPX same-day options traded $132M gross and netted -$0.8M, busiest between the 7,675 and 7,710 strikes, and the index settled at 7,683.69, inside that cluster. The 7,700 same-day put wall expired at the bell and carries no level into Tuesday.

TAPE equity directional residue -$5.1M on $3.37B sided (raw -$514M, sign flipped by structure) · QQQ Dec-18 660 to Mar-2027 670 put roll 191,000 lots, matched · SPY Dec-18 660 / Mar-2027 700 roll 118,000 / 61,840 · IWM Oct-16 272 / 275 / 279 puts 52,200 · QQQ Oct-02 725 parity 1.14x · monetizing share 86.9%

Volatility — the payrolls week moved into the front

The index term structure went from contango to flat in one session. SPX at-the-money vol for the October 2 expiry rose from 11.3 to 14.4, a 27% jump in a day, while December barely moved (13.5 to 14.1). Front over back went from 0.84x to 1.02x. The event week (PCE, quarter-end, ISM, payrolls) is now priced in the front, which means an index calendar that sells October 2 no longer collects a premium for doing so. The October 16 tail is the richest index contract measured: the 10-delta proxy put prices at 2.03 times its at-the-money vol, over the 2.0x line where the buyer is paying up. The December body is closer to fair (1.26x at 5% out of the money). Own the December body, and if anything is sold, sell the October wing.

Across single names the flat tape is gone. Thirteen of 40 names now price the October 2 front at or above 1.10x a back month at least 60 days out, against 2 of 48 on Friday. The rich fronts sit in three places, and most of them are not safe to sell:

Where vol is sticking is telling. Against the September 11 baseline, TLT vol is up 81% and HYG vol up 73%, while index vol rose only in the event-week front. Bond and credit vol lead equity vol, which is the Cem Karsan read of the same day.

TAPE SPX at the money Oct-02 14.4 (from 11.3), Dec-18 14.1, front over back 1.02x (0.84x Friday) · SPX Oct-16 10-delta proxy put 24.2 vs 11.9 at the money (2.03x) · 13 of 40 names at or above 1.10x front over back · TLT +81%, HYG +73% vs the 09/11 baseline

The commentary — who the tape agreed with

The macro voices were right on Monday; the rotation voices were half right. Karsan (bond and currency vol will outrun equity vol), Ilya Spivak (gold breaking as yields surge) and Rob’s Child (junk spreads widening) all described the day the tape printed. Rob’s Child’s separate “the top is in for yields” did not hold on the day, and neither did a Spivak clip, recorded before the break, that read gold’s stability as a sign the rate risk was priced. Silva’s “The Losers Might Be Next” and the Maverick’s “buy the losers, short the winners” both called for a rotation out of September’s winners. The first half printed: the winners fell 2.6% to 5.7%. The second half did not: September’s losers fell too, and IWM broke its monthly band. Silva still reads the regime as bull-strong, with the Nasdaq Composite near its highs and over its rising averages, and he pegs SPY’s Tuesday range at about 760.50 to 770.

The Maverick’s one print claim does not survive as stated. He described buyers of QQQ October 16 730 puts selling the 700s or 710s. Checked structure before side, that exact spread printed on Friday, not Monday: 5,000 of each at 10:05 Central on the 25th, a genuine vertical by the fill test (one leg at the ask, one at the bid, not a matched box), with the 730 open interest rising by 12,013. On Monday the same shape printed at other strikes and expiries. The largest was 16,250 October 30 720 / 700 put spreads at 13:44 Central, about $6.8M. The heaviest Monday flow in the October 16 730 puts ran the other way, sold at the bid and rolled down. He also said a 710 put was being sold; the only large 710 print was a 7,000-lot block bought at the ask with no partner leg. His description of September’s split (losers down 9% to 14%, AI winners up 20% to 30%) checks out against the census. His calls for a relief bounce in gold, IWM and TLT this week, a Nasdaq-futures flush to 30,000, and a VIX breakout by October 12 are graded at their own deadlines, not absorbed as a mood.

The NVDA “$19M bet on the buyback floor” checks out as one buyer. About 10,000 March 2027 260 calls were swept at the offer in 35 seconds at 09:21 Central, with no partner leg on any strike. The line totaled about $18.5M on the day at 0.97 times at-the-money vol, so the buyer did not pay up for the wing. It is one institution’s outright long, a description of sponsorship, not a forward signal. James (InvestAnswers) had bitcoin’s highest weekly close since January and its biggest ETF-inflow week since last October’s flush. He is watching 81,000 to 82,000 as support. Bitcoin held above it, and above the 82,800 and 80,000 lines Cowen and Krown named.

TAPE QQQ Oct-16 730 / 700 put spread 5,000 x 5,000 on 09/25 10:05 CT, a vertical · QQQ Oct-30 720 / 700 16,250 x 16,250 on 09/28 13:44 CT · QQQ Oct-16 710 puts 7,000 bought at the ask, no partner · NVDA Mar-2027 260 calls ~$18.5M at 0.97x · bitcoin ~83,375

Earnings — MU and NKE carry the week’s single-name risk

No roster name reported Monday; the week’s tests are MU on Wednesday and NKE on Thursday, both after the close. MU’s October 2 front at 94.8 implied against 57.8 in December prices the report at about 7.9%, a little under its measured median of 9.1%, and five of its last eight reports moved 8% or more. The tail beyond 8% is not fully priced. NKE’s front prices about 9.1%, roughly 1.4 times its 6.6% median. CCL (before Tuesday’s open) is priced at about 1.6 times its median. The recent reaction regime is mixed and leaning muted: of the September large-cap reports, beats were sold on MDB, LULU and ORCL and rewarded on SNOW, ADBE and COST. On Monday the market bought NVDA’s buyback and sold META’s enterprise launch. Options flow before a report prices how big the move will be, never which way.

TAPE MU Oct-02 94.8 vs Dec-18 57.8 (1.64x), implied event about 7.9% vs 9.1% median · NKE front 2.26x, implied about 9.1% vs 6.6% median · CCL about 1.6x its 4.6% median

Timing — the major turn printed; the drawn bounce is next

The Savino September chart called a major turn at the Friday-to-Monday boundary, and it came. Friday was the high point and Monday closed lower, matching the standard version of the chart. The inverted twin missed the direction for the second turn in a row. The chart’s shape into the boundary also matched: a flat top, a down day, a flat day, an up day short of the top, then the turn. What the chart draws next is a low at the weekend boundary and an up leg from Tuesday into Thursday October 1. Tuesday’s close confirms or misses that low. The up leg is graded at Thursday’s close, on whether it rose, never on how far. Timing charts give timing and shape, not price targets; the size of any move comes from the bands above. There is no bond timing chart on file this week. Silva’s window for a top in the 10-year yield runs to Monday October 5, and the yield was still rising on Monday.

TAPE Savino September: major turn 09/25-09/28 printed (Friday high, Monday lower close) · drawn low at the boundary, up leg 09/29 to 10/01, graded Thursday · inverse twin wrong on direction twice running · 10-year 5.24%

Unusual trades — the structures worth naming

1. QQQ and SPY: December index puts carried out to March

The largest prints of the day were two institutions rolling protection, not buying fear. In QQQ, 191,000 December 660 puts were sold at the bid and 191,000 March 2027 670 puts bought to the ask in the same second, closing 95% of the December line. SPY did the same thing at a 1.9-to-1 ratio, 118,000 December 660 puts against 61,840 March 700 puts. Each fund sold the far wing at a higher skew ratio (QQQ 1.29x, SPY 1.65x) and bought nearer the money at a lower one, which is buying the body and selling the wing. QQQ closed at 736.53, on its zone middle (702.25 / 731.43 / 760.60). SPY closed at 765.61, zone 754.48 / 765.94 / 777.40. The new strikes sit about 9% under both, between the September monthly and quarterly lower bands. Matched legs carry no direction; the information is in the timing. The book’s protection now reaches past the election and the first quarter.

TAPE QQQ Dec-18 660P 191,000 sold 7.87 (26.0 implied) / Mar-2027 670P 191,000 bought 18.27 (24.7 implied), 10:23 CT, about $199M debit · SPY Dec-18 660P 118,000 sold 3.30 / Mar-2027 700P 61,840 bought 12.14, 10:10 CT

2. NVDA: long-dated calls through the offer on the buyback morning

One buyer took about 12,970 March 2027 260 calls and 2,750 of the 255s in the first 45 minutes, with no partner leg. The 260s priced at 36.4 implied against 37.7 at the money (0.97x), so the buyer paid no premium for the upside strike. NVDA’s one-year vol sits at the 2nd percentile of its own history, the one condition in which a single long-dated call is the allowed expression. NVDA closed at 228.86, in the upper part of its zone (211.48 / 223.12 / 234.76). The 260 strike is 13.6% above the close. It is one institution’s outright long on a buyback day, a description of sponsorship, not a forecast. Overwriters sold December and January calls beside it, which is why the name’s net options figure was small.

TAPE NVDA Mar-2027 260C ~12,970 at 14.25-14.32, 09:20-10:03 CT, open interest 26,930 before the day · Mar-2027 255C 2,750 at 15.25 · Nov-20 250C 6,703 sold at the bid

3. MDT: a $418M conversion ladder with no direction in it

MDT ranked fifth by raw options premium on the day and carried no direction at all. From 10:31 to 14:44 Central, October 16 130 puts and 130 calls traded in equal sizes in the same second, 2,000 to 15,250 contracts a pair, priced at parity. That is a conversion ladder, stock plus a synthetic, most likely financing or a dividend trade. MDT closed at 89.50, in the middle of its zone (87.34 / 90.70 / 94.06). The stock’s own block tape showed a large intraday sell figure on heavy volume the same session, most likely the stock legs of the conversions. A screen that sorts by premium put MDT near the top of the board for no reason.

TAPE MDT Oct-16 130P / 130C in equal pairs, 10,000 at 11:38 CT (43.20 / 0.60), 15,250 at 12:32 CT (41.65 / 0.05); put less call 41.60 vs strike less spot 41.58 · line volume 96,500 vs 47,578 open interest

4. MU, TSLA and AMD: about $0.9B of stock moved into four-day options in 19 seconds

At 12:11 Central one institution built the same structure in three names within 19 seconds. It traded October 2 calls and puts at the same strike in equal size, priced at parity: MU at 935, TSLA at 320, AMD at 542.5. These are synthetic stock positions (conversions or reversals) worth about $294M of MU, $294M of TSLA and $339M of AMD, all unsided. The direction of the synthetic cannot be known. The one informative choice is the expiry: October 2 spans MU’s Wednesday report, so whoever holds the MU synthetic carries the event. MU closed at 1,053.98, zone 916.34 / 1,032.55 / 1,148.76; TSLA at 357.45, zone 351.92 / 367.30 / 382.68; AMD at 607.87, zone 507.11 / 593.23 / 679.34.

TAPE 12:11:02 CT AMD Oct-02 542.5C 5,611 at 65.00 / 542.5P at 0.85 · 12:11:11 TSLA Oct-02 320C 8,185 at 40.60 / 320P at 0.28 · 12:11:21 MU Oct-02 935C 2,795 at 123.00 / 935P at 5.20 · open interest zero, opening

5. XLU: long-dated puts sold on a day every utility fell

Someone agreed to own the utilities fund up to three years out and was paid to wait. About 30,350 puts were sold to open across three tickets: January 2029 34s, January 2028 39s and June 2027 36s. They are single legs at three strikes and three expiries, not a matched package, so the structure carries the direction of a put sale. That is roughly 3.0 million XLU shares of obligation, 0.6% to 13% under the close. The puts sold at about 0.95 to 0.97 times the at-the-money vol, near the money at a flat ratio, not a far wing. XLU closed at 39.25, near the lower third of its zone (38.24 / 39.78 / 41.32) and back under its September two-standard-deviation lower band. Selling puts is willingness to own lower, the opposite of a bearish read. With the long end still rising there is no case for a bounce yet, but the patient buyer is visible.

TAPE XLU Jan-2029 34P 17,850 opening at 1.22 (17.9 implied) · Jan-2028 39P 5,000 sold to the bid at 2.40 (17.5 implied) · Jun-2027 36P 7,500 sold to the bid at 0.89 · utilities 0 up, 10 down

6. RCL: a put package the size of the whole line, the afternoon before CCL reported

In three minutes before 14:20 Central, RCL traded 7,000 December 220 puts, 5,750 December 250 puts and 3,000 January 2027 310 calls. Each ticket was 92% to 99% of that line’s entire open interest. That is the signature of an existing position being closed or rolled, not of new protection. Tuesday’s open-interest update settles which: both December lines falling by about the traded size means closed, rising means new. RCL closed at 242.59. RCL ZONE GAP (no row on the zone document); its census iVol weekly bracket is 222.83 / 262.57, with the close in the middle. It is event-risk management ahead of a peer’s report, and put buying is not a bearish read either way.

TAPE RCL Dec-18 220P 7,000 at 9.60 (open interest 7,069) · Dec-18 250P 5,750 at 22.40 to the ask (open interest 5,866) · Jan-2027 310C 3,000 at 5.90 (open interest 3,250) · 220 over 250 skew 1.02x

Scorecard — grading Friday’s edition on Monday’s close

Friday’s edition: B. Its forward mechanism held again, its one dated timing call hit, and both of its Monday band conditions resolved the way it said was more likely. Its one miss was META.

The miss. Friday’s edition called META’s give-back “a line crossed, not a reversal” and said the zone middle would be reached first. On Monday META fell 4.79% to 715.62, straight through the zone middle and under the weekly lower (716.06) the edition had named as its break line. The line was right; the framing that it would take time to reach was wrong. It is the third edition in a row to under-read how fast a stretched leader gives back.

Friday’s Top Trades, marked at Monday’s closes, were flat together (about +$2). The SPY hedge rose about $156, the IWM call fell about $121 after its entry level printed into a band break, MSFT fell about $60, GOOGL rose about $32, and NVO fell about $5. The one entry trigger that fired, IWM’s tag of the monthly lower, fired into a monthly close under it, the opposite of the holding band the entry assumed. That is the lesson for this list. At Monday’s prices the SPY diagonal no longer fits under the structure cap, so it is re-struck below.

The last five editions: B-, C+, B, B-, B.

Projections, followed up — the branch weights and the open theses

The branch weights to Friday October 2 are carried as registered: absorb and grind about 55%, the rate shock transmits into the index about 25%, an early inflection about 20%. After five closes, absorb still leads by the grading rule: no SPX close under about 7,550, and none of the four early-inflection legs (a 10-year under 5.00%, the dollar under 100.05, TLT over 81.35, HYG over 78.35) on any close. But the transmission branch gained two confirmations on Monday. IWM closed under its monthly lower, and credit moved its way, with HYG under its zone low and the high-yield spread through 3%. The trigger that would settle it, a close near 7,550, sits about 134 points under Monday’s close, about 1.2 weekly standard deviations. The weights are recorded, not re-weighted: they are graded as they were written.

The open lines Monday touched:

TAPE branch weights 55 / 25 / 20 carried · transmission trigger about 7,550 vs close 7,683.69 · early-inflection legs 0 of 4 · GLD 377.91 vs 392.16 · QQQ 736.53 vs 725 · NFLX 69.23 vs 69.25 · EOSE 3.07 vs 3.34 · NEM 116.05 vs 115.81

The week ahead — the conditions, day by day, with levels

For rates, the 10-year’s new zone high is 5.31% (the breakout line) and 5.00% is the topping line. TLT’s weekly lower (77.94) is the next test. The high-yield spread at 3.02% has room to its 2026 high of 3.46% from March.

Bottom line

Monday was a rate day that reached credit. The 10-year closed at 5.24% on real yields, not inflation expectations, and oil never stretched. The index gave back less than 1% and broke only its daily band, while the high-yield spread went through 3%, small caps closed under their monthly band, gold fell nearly four daily standard deviations, and the AI stocks fell as a group while NVDA’s buyback held the cap-weighted total up. The options tape did not panic. Its largest prints carried index protection from December into March, and its directional residue was flat. The swing frame is intact, and the week’s tests are all still ahead: Tuesday’s close against the weekly band, then PCE and the quarter-end Wednesday, ISM Thursday and payrolls Friday.

So, in order:

Top Trades to Follow

These are institutional-flow structures the framework would follow. Each is priced off the surface it trades on, with the window it is owned into and the line that retires it. Every entry went through the entry check (vol, level, exit): READY means all three points are met, WATCH means one is missing, and the missing one is named. The check measures entry quality, not the odds of winning; its own retrospective found no edge as a picker. Every long leg is sixty days or longer. Every maximum loss sits under the per-structure cap of 1,342 dollars at Monday’s closing marks, and holds only if the live debit is under that cap. This is institutional flow to follow, never advice. With 13 of 40 names now pricing the front at or above 1.10x the back, the list carries two term structures. It carries only two, because most of the rich fronts sit on earnings dates or band-break lows, where nothing is sold.

DIAGONAL SPY Dec-18 750 put / short Oct-16 735 put diagonal — the index hedge, re-struck under the cap. WATCH: at Monday’s close the index reversion-buy setup was active on SPY, and nothing is sold into a band-break low. The diagonal is entered only after Tuesday’s close settles it, either back inside Tuesday’s daily band or, on a close under the weekly lower 760.54, owned as the December body without the wing. The December 750 body prices 16.1 implied, 1.09x the December at-the-money (fair, owned). The October 16 735 wing prices 17.2 against an October 16 at-the-money of 13.1, 1.31x (the richer contract, sold); the SPX October tail at 2.03x confirms that is where the premium is dear. Long-dated SPY vol sits in its high third, so a spread, which a diagonal is. Debit about 11.38, maximum loss about 1,138 dollars. Owned into payrolls and the October window (the wing dies Fri 10/16) and the December window; retired on a SPY close over its zone high 777.40 or at the December expiry. It registers no index direction.
CALENDAR HOOD Dec-18 / Oct-02 120 call calendar — sell the event-week front, own the December body. WATCH: HOOD at 116.46 sits inside every band, 1.2% under its monthly upper, at no measured buy level. That is the missing point. After a vertical run a naked long call is not allowed, and a calendar is the sanctioned shape. The October 2 120 call prices 75.9 implied against the December 120 call at 61.3, 1.24x front over back. Long-dated HOOD vol sits in its low third (20th percentile), so the back leg is the fair contract to own. Debit about 10.09, maximum loss about 1,009 dollars. The front dies Fri 10/02 (payrolls and month-end), with at most one roll; the December body is closed by Tue 10/27, ahead of the late-October earnings window; retired on a HOOD close under its implied-vol weekly lower 109.70.
CALL IWM Dec-18 280 call, one lot — READY on the three points, with the driver named. Vol: long-dated IWM vol is 20.9, the 7th percentile of its own history, so a single long-dated call is allowed and keeps the tail; the 280 call prices 19.0 implied, about 0.98x the December at-the-money. Level: 280.02, the first close under the monthly lower 281.09, the level Thursday’s edition named. Exit: registered, a close below 274.14. Debit about 11.35, maximum loss about 1,135 dollars. The named driver, the 10-year yield, is still rising, and the monthly break is a continuing trend. This is a level to own the small-cap rebound at, not a call that the low is in. Owned into the December window through the election and the refunding (11/03 and 11/04), harvested at the zone high 290.15, retired on a close below 274.14.
SPREAD NVO Jan-15-2027 40 / 45 call spread, one lot — READY on all three points. Long-dated NVO vol is 39.0, the 3rd percentile of its own history. The 40 call prices 35.6 implied, and the front is the cheaper contract (about 29.7 against 35.6 in January, 0.83x), so no calendar. Price 38.71 sits under its monthly implied-vol lower (41.95). The exit is registered: a monthly close under 34.15. Debit about 1.48, maximum loss about 148 dollars. Owned into the January window and closed before its final sixty days (by Mon 11/16), harvested at the 45 strike or twice the debit, retired on that monthly close.
SPREAD MSFT Dec-18 520 / 540 call spread — capped, because holders are capping it. WATCH: the breakout held a second close, but price sits 1.8% above the weekly lower 500.25, not at a level to buy anyway. Long-dated MSFT vol sits in its low third. The 520 call prices 29.8 implied, 1.00x the December at-the-money, and the October front is not rich (29.9 against 29.8), so no calendar. Debit about 7.90, maximum loss about 790 dollars. Owned into the December window, closed by Tue 10/27 before the late-October report; retired on a close under 500.25 before entry and, once owned, under the weekly two-sigma lower 484.33.
SPREAD GOOGL Dec-18 340 / 370 call spread — WATCH: the plan enters only after a close with HYG at or above 78.35 and TLT at or above 80.33. Monday closed at 77.54 and 78.62, both further away than Friday. Long-dated GOOGL vol sits in its low third. The 340 call prices 33.4 implied, 0.99x the December at-the-money, and the October 16 front is the cheaper contract (30.1 against 33.4), so no calendar. Debit about 12.60, maximum loss about 1,260 dollars. Owned into the December window but closed by Tue 10/27, before the late-October report and the Fed meeting; retired on a close under 327.74.

Not listed, and why:

Sources

Expected moves & levels: the Silva daily board for 09/29 (forward) and 09/28 (graded), the ZONE DOCUMENT, Zone Visual, zones and range-and-trend sheets for 09/29 and 09/28; the weekly board for 09/28 to 10/02; the September monthly and Q3 quarterly boards with the JPM collar levels for Q3; the census implied-volatility blocks and the computed implied-volatility bracket for 09/29; the FOM sentiment gauge for 09/28 (37.9, fear). Tradytics: the Live Options Flow and Darkpool Market Summary end-of-day files for 09/28, with the archived 09/18 to 09/25 option files for the commentary checks. The options dashboard for 09/28 was not captured and is declared missing; its panel reads are replaced by a decomposition of the options file. Recon pipeline: the 2026-09-28 census (650 per-ticker reports; 116 read in full for this edition, all 650 scanned for breadth; SPX, NDX, RUT, PCG, FICO, SN, JEF, MTN, IDT and KSPI have no census file and are declared) and the twelve sector chunks for the bottom-up sector map; the 09/18 to 09/25 census for the multi-day closes. Ledgers and research: the flow, dark-pool, monetization, setup-scan, expected-move, implied-volatility, long-dated-volatility and volatility-tracker ledgers for 09/28; the census screen and both basket monitors; the entry check, reversion-buy state and loss-budget state; the thesis, timing and expected-move analysis ledgers; the earnings-reaction history. Intraday: the pre-open brief and the 09:15, 10:15, 12:45 and 14:30 briefs for 09/28, the close level frame and kill-line alerts, the FOM and Maverick morning briefs. Commentary: every 09/28 polished transcript on the shelf, including Mike Silva’s “The Losers Might Be Next” with five pages of his deck, the Maverick’s “October Outlook”, Cem Karsan, Ilya Spivak (two), Darius Dale, SpotGamma, Rob’s Child, Mike Butler, James (InvestAnswers) with his 14 Patreon posts, Benjamin Cowen, Krown and the tastylive NVDA segment; the external-analyst feed. Timing: the Savino September projection and its inverse twin (no bond timing chart on file). Operator book: the Robinhood close reconcile for 09/28 and the Fidelity roster (used for the framework’s own records only; nothing about the book appears above). Market data: option day bars and equity closes through the Robinhood connector, read-only; FRED series for the 10-year, real yield, breakeven, 30-year, 2-year and high-yield spread. No news web search was used for this edition.