← Back to All Reports
DAILY REPORT · THURSDAY 10/01 CLOSE · FIRST SESSION OF OCTOBER AND THE FOURTH QUARTER · A HOT PRICES INDEX, A 24-YEAR HIGH IN THE 10-YEAR, AND THEN A LOWER CLOSE IN YIELD · THE HIGH-YIELD SPREAD WIDENED ANYWAY · A ROTATION INSIDE AN UNREPAIRED WEEK · PAYROLLS SETTLES THE WEEKLY BANDS

Daily Report — 10/01/26 · “The Bond Turned First, and Credit Did Not Follow”

The S&P 500 fell through its weekly lower band in the first ninety minutes of the quarter, on a jump in the factory prices index and a 10-year yield at a 24-year high. Then buyers stepped into the long bond, the 10-year closed lower on the day, and the index closed back inside the band at 7,666.45, up 0.19%. More stocks rose than fell for the first time in four sessions. One thing did not turn with the bond: the high-yield credit spread widened to its widest since March on a day yields fell. This edition grades Wednesday’s report, maps the levels Friday’s payrolls session settles, takes apart the day’s largest option prints from the raw file (the options dashboard was not captured for this session, and that gap is stated wherever it matters), and adds one new structure to the list.

Thursday in one line — the index broke its weekly band, the bond was bought, and the close took it all back

The S&P 500 (SPX) closed at 7,666.45, up 0.19%, after trading 18.63 points under its weekly lower band in the morning. Two reports landed before 09:00 Central and both read hot. Weekly jobless claims came in at 197,000 against 200,000 expected. The September ISM manufacturing index held at 54.5, a ninth month of expansion, and its prices index jumped to 77.9 from 71.1. The 10-year yield ran to 5.34%, a 24-year high. SPX fell to 7,616.97 at 10:11 Central, under the weekly lower band at 7,635.60 and 36 points under the level where dealer hedging stops cushioning moves (7,653).

Then the long bond was bought, with no second piece of news. The long-bond fund TLT traded up to 78.05 just after midday. The 10-year closed at 5.24%, lower than Wednesday’s 5.29%: the first lower close since the run to new highs began. The index followed the bond back up to the minute and closed 30.85 points over the weekly lower and 13.45 over the dealer level. One lower close in yield after a new high is a change in speed, not a turn. It is still the first one.

Under the index the day was a rotation, not a rescue:

After the close NKE reported and was marked near 32.97 against a 35.15 close. The FOM sentiment gauge printed 39.8, a fourth straight “fear” reading and the first up-tick of the four; it is far from the sub-15 readings that have marked washouts.

TAPE SPX low 7,616.97 (10:11 CT), weekly lower 7,635.60, close 7,666.45 (+0.19%) · 10-year high 5.34%, close 5.24% (from 5.29%) · claims 197K vs 200K · ISM 54.5, prices 77.9 from 71.1 · TLT high 78.05, close 77.71 · census 350 up / 298 down; week 193 up / 456 down · RSP +0.47%, SPY +0.18% · SMH 617.81 (+1.45%), XLE 62.70 (+1.95%), XLV 166.20 (-1.32%) · sentiment 39.8

The regime under it — rates eased for one close, and credit moved without them

The Fed is still tightening, and the bond market is still doing more of it than the Fed. The policy rate is 3.75-4.00% after the hike on 09/16. The 10-year real yield (the part of the rate left after expected inflation) is 2.93% against a breakeven inflation rate of 2.36%. When the real yield sits above the breakeven, the market is raising the cost of money on its own, and at a 5.24% 10-year the multiple the rate model allows for the index is about 10.8 times earnings. That is one bearish input for stock multiples. Thursday did not remove it. It slowed it by one close.

The new fact of the day is in credit. Through Wednesday the high-yield spread (the extra yield investors demand to hold junk bonds over Treasuries) had widened as yields rose: one story, counted once. On Thursday the 10-year fell and the spread widened anyway, by 12 basis points to 3.24%, the widest since 03/31 and the eighth wider print in a row from 2.66%. The high-yield bond fund HYG closed at 76.90, zone 76.76 / 77.59 / 78.42; its monthly distribution of about 0.34 was larger than its 0.31 decline, so it was flat on a day TLT rose after its own distribution. A spread that widens when yields rise is about rates. A spread that widens when yields fall is about credit. One session does not prove the second. It is the reason this edition counts credit as its own input for the first time, and why the line that matters is HYG back toward 78.35, not only TLT.

The dollar keeps the metals closed. The dollar index closed at 102.00, above 100 with a wide trend range, which is the framework’s strong-dollar block on new bullish metals positions. The dollar sits about four-fifths of the way up its own zone, inside it, so there is no reading that the dollar itself is stretched; the block governs what gets opened and is not a forecast for the dollar or for gold. GLD (382.76, zone 374.18 / 389.01 / 403.83), SLV (55.02, zone 53.20 / 56.98 / 60.76) and GDX (86.74, zone 84.46 / 91.31 / 98.15) all closed under their weekly lower bands (384.46, 55.98, 88.66). There is no basis for calling a low or a further leg down from those bands: metals have been a coin flip after band breaks, and the dollar has not turned.

Growth is confirmed, and prices are the problem. A factory index at 54.5 with a prices index at 77.9 is an economy that is expanding and passing costs through. That is why the first move of the day was in yields. Crude closed at 92.91, over the top of its daily band (92.38) on an export-ban threat; a threat is not a policy, the weekly band (86.45 to 98.43) holds it, and there is no basis for fading or chasing it.

TAPE Fed 3.75-4.00% · 10-year 5.24% (high 5.34%), 30-year 5.60% · real yield 2.93% vs breakeven 2.36% · fair multiple about 10.8x · high-yield spread 3.24% (+12 bp; 2.66% on 09/21), widest since 03/31 · HYG 76.90, zone 76.76 / 77.59 / 78.42 · dollar index 102.00 · GLD 382.76, SLV 55.02, GDX 86.74 under weekly lowers · crude 92.91 vs daily upper 92.38

The four timeframes — inside the month and the quarter, and the week settles on payrolls

On the two longest timeframes nothing has happened, which is itself the reading. One session into both, SPX sits in the middle of its October monthly band (7,363.93 to 7,939.15) and its fourth-quarter quarterly band 7,089.48 to 8,213.60. No index is near a quarterly edge: SPX is 6.7% under its quarterly upper and 8.1% over its quarterly lower, and SPY, QQQ and IWM are the same. Every sector fund also closed inside its October band. The damage of the last two weeks broke weekly bands, not monthly ones, so it is a swing-timeframe event and not yet a change of regime.

The weekly band is where the week is decided, and it is decided Friday. SPX’s weekly band is 7,635.60 to 7,851.22. Thursday traded under it and closed back inside. A weekly-lower touch that recovers has been followed by a higher next session about 65% of the time over the last 24 sessions, against 55% for any session: a lean, measured on a short sample, and graded here on one thing only, a Friday close over 7,635.60. Friday’s daily band is 7,614.13 to 7,718.77, the widest of the week, and its lower edge sits 21 points under the weekly lower. So an ordinary down session on Friday closes the week outside its band, and an ordinary up session leaves it comfortably inside. Friday’s zone is 7,562.23 / 7,682.70 / 7,803.17; the index stalled 16 points under the middle of it on Thursday.

The funds line up the same way, with one exception:

The bond and the rate-sensitive funds are the other half of the map. TLT (77.71, zone 76.94 / 79.24 / 81.54) closed 0.23 under its weekly lower (77.94), but the fund paid a distribution of about 0.31 on Thursday, so on a total-return basis the bond rose. Its monthly band is 74.49 to 81.07 and its quarterly band 72.29 to 83.27. There is no basis for calling either a bounce or a further fall from that band: the yield eased for one close and nothing more. The marker is a Friday close over 77.94. XLF (53.46, zone 52.42 / 54.08 / 55.73) and KRE (69.95, zone 68.02 / 70.63 / 73.23) closed under their weekly lowers (53.75, 70.08) again, on an up day; with rates easing and credit worsening, the two things that move banks point opposite ways, so there is no basis for a call there either. XLV (166.20, zone 165.91 / 169.04 / 172.16), XLP (80.33, zone 79.76 / 81.79 / 83.82) and XLC (109.94, zone 109.11 / 112.03 / 114.95) closed under theirs (167.36, 80.96, 110.59) with no cause named in any input; no basis for a call, and the fact that defensives fell on a day yields eased is weakness the rate story does not explain. XLU (39.68, zone 38.09 / 39.64 / 41.19) is the one rate-sensitive fund still inside its weekly band (38.69 to 40.33).

TAPE SPX weekly 7,635.60 / 7,851.22, October 7,363.93 / 7,939.15, Q4 7,089.48 / 8,213.60, Friday daily 7,614.13 / 7,718.77, zone 7,562.23 / 7,682.70 / 7,803.17 · SPY low 758.89 vs weekly lower 760.54 · IWM low 275.61 vs 275.98 · DIA 508.62 vs 510.31 · TLT 77.71 vs 77.94 (distribution about 0.31) · XLF 53.46 vs 53.75, KRE 69.95 vs 70.08, XLV 166.20 vs 167.36, XLP 80.33 vs 80.96, XLC 109.94 vs 110.59 · XLU 39.68 inside 38.69 / 40.33

The mega-caps and the roster — the weight fell while the breadth rose

Silva’s quarterly board lists only indexes and the long bond, so for single names the timeframes below are Friday’s daily band, the weekly band, the October monthly band and Friday’s zone; the quarter frame for a single stock is its implied-volatility bracket to the end of December, which is labelled where it is used.

AAPL (330.32, -0.81%), zone 328.22 / 335.07 / 341.92, closed under its weekly lower band (332.75) for the second time this week. No cause is named for it: the bond eased, the market rose, and AAPL fell anyway. A single stock that breaks a band without a driver behind it has usually come back inside, on a small sample from one kind of market; the trigger is a close back over 332.75, and the line that says the idea is wrong is a close below 325.75. It is inside its October band (311.95 to 354.09) and over its trend line (322.67); Friday’s daily band is 326.34 to 334.30, and the quarter-end implied-volatility bracket is 290.85 to 375.19. AAPL reports on Thursday 10/29. This is the one new structure on the list below.

TAPE AAPL 330.32, weekly 332.75 / 349.39, October 311.95 / 354.09, Friday daily 326.34 / 334.30, zone 328.22 / 335.07 / 341.92, trend 322.67 · options after structure: flagged as financing and paired legs, not readable as direction (carry moved from an October 300 strike to a November 330 strike; a January 2027 320 call bought against a January 2028 350 call sold, 2,704 lots)

GOOGL (338.24, -1.70%), zone 334.28 / 344.14 / 353.99, and GOOG (334.93, -1.71%), zone 331.56 / 340.69 / 349.81, made their lowest closes of the last five sessions after opening higher on a product launch. Wednesday’s most-bought name became Thursday’s reversal. What the option tape shows is not selling by bears. Holders took income: June 2027 345 calls and November 370 and 400 calls were sold against stock already owned, and long-dated puts were sold as well. That caps the stock into November without being a bet against it. The headline option figures for both classes are too unsided to read as direction once paired legs are removed. Both classes moved together, so the company view is one line: GOOGL’s weekly lower band at 332.07 is the Friday level, with its October band at 313.63 to 374.53 and Friday’s daily band at 332.95 to 343.53. One close out of a range is a watch, not a signal.

NVDA (230.86, +1.09%), zone 213.28 / 225.95 / 238.62, made its highest close in nine sessions, 2.11 under its weekly upper band (232.97). On the up day holders bought December 225 puts and sold November 245 calls: protection and income from people who own it, which is how a run gets hedged, not how it gets shorted. October band 210.13 to 246.63, Friday daily 227.28 to 234.44. META (725.93, +0.10%), zone 661.39 / 733.40 / 805.40, is still oscillating 9.87 over its weekly lower (716.06); October 645.04 to 805.32.

AVGO (343.64, -2.15%), zone 333.18 / 349.13 / 365.08, is the chip name that did not join. It made its lowest close in nine sessions on the day SMH made its highest, and sits 3.95 over its weekly lower (339.69) and 40 points under its trend line. Friday’s daily lower (337.59) is under the weekly lower, so one ordinary down day closes its week outside the band. October 320.72 to 381.66.

MSFT (512.80, -0.02%), zone 492.35 / 506.83 / 521.30, has held five closes over its 505 to 508 breakout without reaching the 520 area that would complete the move; weekly 500.25 to 532.09, October 470.94 to 554.86, Friday daily 506.04 to 519.56. AMZN (248.23, -0.37%), zone 241.26 / 248.86 / 256.45, is mid-band on every frame (weekly 241.15 to 258.19, October 225.41 to 272.89) and under its trend line at 256.68. TSLA (354.11, -0.20%), zone 344.94 / 363.96 / 382.98, has closed within two points of its weekly lower (353.57) for three sessions; Friday’s deliveries figure and Friday’s close decide it, and October is 314.96 to 394.66.

MU (1,097.39, +3.03%), zone 934.42 / 1,057.72 / 1,181.01, made its highest close in nine sessions on the session after its report, after trading down 3.9% in the first hour. The move was 0.41 of what its options had priced. A holder then sold a December 1,200 / 1,500 call spread and kept October 1,150 calls, 10,092 lots: capping the run and keeping the near-term upside. Weekly 984.11 to 1,180.45, October 911.47 to 1,218.75. AMD (615.73, +0.65%), zone 533.69 / 612.83 / 691.96, and INTC (120.00, -0.19%), zone 100.83 / 119.04 / 137.25, are inside every band after vertical runs (AMD weekly 593.18 to 668.08, October 536.09 to 687.43; INTC weekly 113.99 to 132.01, October 98.58 to 141.88). Their option stories are in the unusual-trades section: AMD’s is the cleanest figure on the tape and INTC’s largest print carries no information at all.

NFLX (67.85, -2.49%), zone 64.64 / 70.26 / 75.88, closed under its weekly lower band (68.77) for the first time, to a new low for the move. Wednesday’s edition called its three flat closes “a pause, not a floor,” and that held. There is no basis for a call from the band on one close with no driver named. The raw option file shows heavy put volume in NFLX; it is part of an afternoon exercise trade described below and says nothing about the stock. October 61.69 to 77.47.

SPCX (148.07, -1.85%) is mid-band on its weekly (140.80 to 156.56) and October (134.80 to 166.96) frames; SPCX ZONE GAP. MSTR (160.50, +4.84%) reversed after seven lower closes, inside its weekly (145.85 to 171.37) and October (128.27 to 177.91) bands; MSTR ZONE GAP. Holders rolled calls up and out on the bounce (October 160 calls sold, October 23 172.5 calls bought), which is income taken by owners, not a short.

TAPE GOOGL 338.24 / GOOG 334.93, weekly lower 332.07 · NVDA 230.86 vs weekly upper 232.97 · META 725.93 vs weekly lower 716.06 · AVGO 343.64 vs weekly lower 339.69, trend 384.27 · MSFT 512.80, five closes over 505-508 · TSLA 354.11 vs weekly lower 353.57 · MU 1,097.39 (+3.03%), December 1,200 / 1,500 call spread sold, 10,092 lots · NFLX 67.85 vs weekly lower 68.77 · MSTR 160.50 (+4.84%) · of ten mega-cap lines, two rose

The sector map — a rotation, and what it did not repair

Thursday was broad by count and narrow by repair. Seven of the twelve sector groups in the census had more risers than fallers. But only three of them, energy, utilities and technology, finished with more names at five-session highs than at five-session lows, and only technology and utilities are higher on the week at the median. The equal-weight fund RSP and the cap-weighted SPY are level on the week (-1.00% and -0.95%); the median census stock is down about 2%. The rally was real and it did not undo the week.

The two baskets parted. The AI hardware basket averaged +1.74% and the speculative basket (nuclear, miners, drones, quantum) -0.71%: money went to the names that report revenue and away from the names that share their theme. The exhaustion screen, the one bearish setup the flow data can see (a vertical run, heavy call chasing and heavy call selling together), tagged two names on Thursday, S and ENTG. Neither is an index leader. Sixty-three names carry the sponsorship tag, which describes a trend that is being supported and is not a buy signal.

TAPE sector groups up/down: technology 74/26, energy 29/16, utilities 10/0, financials 24/16, industrials 53/29, staples 8/15, materials 7/25, health care 4/43 · week: financials 3/37, materials 0/32 · five-session lows 239 vs highs 153 · RSP -1.00% and SPY -0.95% on the week · AI basket +1.74%, speculative basket -0.71% · setup tags: 63 sponsorship, 2 exhaustion (S, ENTG)

The options tape — flat after structure, and the dashboard is missing

A gap first. The options dashboard was not captured for Thursday, so the panels this section normally reads (dealer positioning by expiration, same-day gamma by strike, the flow timeline) are not available, and nothing below claims a dealer wall or a pin. What follows is built from the raw end-of-day option file, with financing trades, paired legs and stock substitutes removed before any side is read. The only dealer levels cited are Silva’s published ones: 7,653 for Thursday and 7,657 for Friday as the level where dealer hedging changes character, with 7,500 and 7,800 as the outer anchors.

After structure, the tape had no direction. The file traded $13.55B of premium. Its raw net was -$542.6M, which looks like heavy put buying and is not: $1.11B of it was a strip of deep in-the-money puts bought across forty names in one afternoon hour, in sizes keyed to open interest, which is an exercise trade by market makers and carries no view. Once that strip, the paired legs and the loans are removed, the file’s directional net was -$24.8M: flat. For single stocks the net after structure was -$61.7M on $3.30B of sided premium, and three names account for more than all of it: GOOGL, NVDA and GOOG, each a case of holders selling calls on stock they own, and each too unsided to cite as direction. Without them the equity tape was +$3.7M.

The index tape is structure, not a view. The index net after structure was +$40.7M, and it is three things: one block of 42,000 QQQ November 775 calls, one second of long-dated SPY puts being sold, and leftover legs of SPX box spreads. Box spreads (a four-leg package that is a loan, not a position) were 41.7% of all SPX premium, priced at 4.4% to 4.8% a year, the same range as Wednesday. The SPX 7,000 and 8,000 strikes are loans on five expirations and are not levels. The one index line with real one-sided buying was IWM, where puts were bought on four expirations on a day it closed up. That is hedging by people who own small caps, not a forecast.

What was bought with time on it points one way: a bond bounce. The largest opening call purchase of the day was 75,000 TLT September 2027 86 calls, $12.2M, mostly at the offer, beside Wednesday’s November 83 call line, which was held (open interest 96,741). The largest fund print was an XLU November 40 / 44 call spread of 88,866 lots a leg, with no quotes shown, so its side cannot be read. Against that, protection was added in credit and banks: 57,750 HYG February 2027 75 puts and 44,000 October 76 puts bought, 26,500 KRE November 60 puts, 15,000 XLF September 2027 47 puts. The option market is long a rebound in bonds and long protection on lenders at the same time. That is the same split the credit spread printed, expressed in positions. None of it is a signal to follow: TLT’s own option figure is dominated by financing legs and is not readable as direction.

TAPE file gross $13.55B; raw net -$542.6M; net after structure -$24.8M (Wednesday +$572.1M) · deep in-the-money put strip $1.11B, 13:30-15:15 CT · single stocks -$61.7M on $3.30B sided; +$3.7M without GOOGL, NVDA, GOOG · index +$40.7M, structure · SPX box spreads 41.7% of SPX premium at 4.4-4.8% · TLT Sep-2027 86 calls 75,000 lots; Nov-30 83 call open interest 96,741 · XLU Nov-20 40 / 44 call spread 88,866 lots, unsided · HYG Feb-2027 75 puts 57,750 lots · holders selling calls on a run: 82.3% of call-selling premium

Volatility — the whole premium sat in one contract

The option market priced payrolls as a one-day event and the week after it as the calmest stretch on the curve. SPX at-the-money implied volatility closed at 16.0 for Friday’s expiration, up 15% in a session, against 12.4 for the following Friday, 12.7 for October 16 and 14.2 for December. Friday’s contract priced 1.13x December; next Friday’s prices 0.87x. On Friday’s contract, 27 of 39 names carried a premium of a tenth or more over a contract sixty days out. One expiration later only 3 of 40 names sit at or above 1.10x (NKE, EWZ and XLF), and the median name prices 0.92x. So from Monday there is no rich near-dated contract left to sell: a calendar spread sold next week would be selling the low point of the curve.

Where the surface does pay is the strike, not the date. A put 5% under the index for October 16 prices 1.54x its at-the-money volatility; the same distance in December prices 1.24x. The far tail (about 9% under) for October 16 prices 1.86x, under the 2.0x line where the framework says a tail is overpaid for the first time in three editions, because the near-the-money contract rose and the tail did not. The people hedging the index on Thursday traded it that way: a 20,000-lot SPY October 30 put spread bought the 755 strike at 1.04x and sold the 725 strike at 1.30x. Buy the contract near the money, sell the one far from it.

Bonds and credit are where volatility is still building. TLT’s Friday contract priced 1.52x its December (1.45x on Wednesday, 1.28x on Monday), and its one-year implied volatility is at the 99th percentile of its own history. HYG’s October contract prices 1.53x its December. IWM’s Friday contract was the richest of the indexes at 1.37x. In equities the stress is priced as one day; in rates and credit it is priced as ongoing.

TAPE SPX at-the-money vol 16.0 (10/02), 12.4 (10/09), 12.7 (10/16), 14.2 (12/18) · 5% put Oct-16 1.54x, Dec-18 1.24x; far tail Oct-16 1.86x (2.02x Wednesday) · SPY Oct-30 755 / 725 put spread 1.04x / 1.30x · TLT front 1.52x December; one-year vol 99th percentile · HYG October 1.53x December · IWM front 1.37x · next-week front: 3 of 40 names at or above 1.10x, median 0.92x

The commentary — everyone agrees on the diagnosis and splits on the next two weeks

Twenty-seven commentary pieces were read for this session, and on rates they say one thing. The long end is a supply problem that the policy rate cannot fix. Thursday showed it in one session: the odds of an October hike fell from about 70% to the mid-20s, the 2-year yield fell about 10 basis points by one desk’s figures, and the 10-year gave back only 5 and the 30-year about 3. The Fed moves the front of the curve; the back answers to issuance. That is this site’s own view, stated by other people, so it is reported here as agreement and not argued against. Darius Dale supplies the arithmetic (his figures, not checked here) and dates a Treasury-and-Fed response to early 2027. No Silva commentary or slide deck was published for this session, which is a gap; his regime label from Wednesday night stands as the last one on file.

Where they split is timing. Three desks put a turn in rates inside the next two weeks: Silva’s 2018 comparison (a pause in yields around Monday 10/05), Arete (a peak in the 10-year by Thursday 10/15) and Tim Knight (a bond bounce starting Friday, which he would then sell). Benjamin Cowen puts the top four to six weeks out, between Thursday 10/29 and Thursday 11/12. They cannot all be right. Each is graded on whether the turn arrives in its window, never on a price.

The Maverick named prints this time, and the prints are real. His video cited a large TLT call purchase for 2027, crowded TLT calls for Friday, and MU’s 1,100 calls. Checked against the raw file:

His morning call for Thursday, that the market would either fade AI hardware on MU’s report or revive the largest platform stocks, missed on both branches: AI hardware was bought and the platforms were sold. He is a standing bear; his calls are scored at their deadlines and not followed. What survives from his piece is a positioning fact that now spans three sessions: someone keeps buying upside in the long bond and holding it.

On the AI trade the desks split three ways, and Thursday sided with the bulls on price and the sceptics on breadth. The bulls (Arete, Geeks of Finance, Dividend Data) read MU’s first-hour drop as a shakeout. The sceptics (Dale, Knight, FX Evolution) say chips are the only thing holding the index up. The timing bears (the Maverick, Brent Kochuba of SpotGamma) say the trade is in trouble until earnings around 10/16. The tape: SMH at its high of the week, MU +3.03%, against GOOGL, AVGO and AAPL lower. The flow data gives the timing bears nothing (two exhaustion tags, neither on a chip leader) and gives the bulls less than they claim (most of the chip option tape is holders selling calls and hedging, with TSM the one name where calls were simply bought).

TAPE TLT Sep-2027 86 calls 75,000 lots, 61,000 at the offer, 0.96x at-the-money vol · TLT Oct-02 78 calls: open interest 30,655 vs 8,500 puts; captured lots 8,064 at the offer, 8,003 at the bid · MU Oct-02 1,100 calls: volume 82,630, 49% at the offer, 43% at the bid · timing windows: Silva 10/05, Arete by 10/15, Cowen 10/29 to 11/12

Earnings — Wednesday’s frame was wrong within a session

Wednesday’s edition said beats are not being paid. Thursday paid two of them. MU closed up 3.03% after falling almost 4% in the first hour, and ACN closed up 15.78%, more than twice what its options had priced and its second-largest reaction in nine reports. Of the last ten large reports, five have now been paid (ADBE, COST, CCL, MU, ACN) and five sold or ignored (LULU, ORCL, KMX, CAG, JBL). That is a mixed record, not a pattern in either direction. MKC fell 4.87% on a beat the same day, so the split is real. The lesson is about the frame, not the names: the options tape prices the size of an earnings move and has no record of pricing its direction, and this edition makes no direction call on any report.

NKE (35.15, -0.71%) reported after the close; NKE ZONE GAP. Its options priced a move of about 9.3%; it was marked near 32.97 after hours, a move of about 6%, inside what was priced. Its weekly implied-volatility lower band is 32.06, and that is the level Friday’s close is read against. Next week: STZ reports Tuesday 10/06 after the close, APLD Wednesday 10/07 after the close, PEP Thursday 10/08 before the open, DAL Friday 10/09 before the open. The banks begin Tuesday 10/13, the consumer price report is Wednesday 10/14, TSM reports Thursday 10/15, and the monthly expiration is Friday 10/16.

Timing — the drawn bounce came two sessions late

The timing projections are read for three things only: when a turn is drawn, which way the next leg is drawn, and whether the shape matches. They carry no price information; size comes from the expected-move bands.

Both versions of that chart end at Friday’s payrolls session, and no October chart or bond chart is on file, which is a gap. Size for Friday is in the bands above: SPX daily 7,614.13 to 7,718.77, weekly 7,635.60 to 7,851.22.

Unusual trades — the structures worth naming

1. INTC: a $572M call print that was stock changing hands

The largest single-name option print of the day carries no information about INTC. At 13:57 and 14:02 Central, 61,528 October 23 calls at the 27 and 30 strikes traded with no bid or ask shown, $572.1M of premium, 82% of INTC’s option volume by value. The strikes are about 90 points under a 120-dollar stock, so the calls are all intrinsic value: stock in another wrapper. In the same minute the share tape printed INTC’s largest block of the day, $475.2M at 120.75, which is the larger call block’s share count to the share. A position held through deep in-the-money calls was exchanged for shares. The option leg and the stock leg cancel. INTC closed at 120.00, zone 100.83 / 119.04 / 137.25, inside its weekly band (113.99 to 132.01).

TAPE INTC Oct-23 30 calls 18,178 lots at 90.88 (13:57 CT) and Oct-23 27 calls 43,350 lots at 93.87 (14:02 CT), no quotes, 0.2% time value · share block $475.2M at 120.75, 14:02 CT · INTC option figure after structure: 87% unsided, not readable

2. The afternoon put strip: $1.11B that looks like fear and is an exercise trade

Anyone reading the raw file would have counted a billion dollars of put buying in LULU, SLV, TLT and NFLX. None of it is a view. Between 13:30 and 15:15 Central, 664 prints of deep in-the-money puts traded across about forty unrelated names, $1.11B in all, with 97% to 99% of each price being intrinsic value. In 183 of those prints the size is exactly 1.5 times the open interest on the line, a signature the prior session’s file does not show at all. That is how a market-maker exercise trade is sized: with short rates near 4.5%, deep in-the-money puts are worth exercising early. The inference comes from the size pattern; the tape does not name the owner. After those legs are removed, LULU’s raw -$77.7M becomes +$0.6M, SLV’s -$144.3M becomes -$6.2M and NFLX’s -$86.5M becomes -$1.2M. NFLX closed at 67.85, zone 64.64 / 70.26 / 75.88; its break of the weekly band is a price fact with no option story behind it.

TAPE deep in-the-money put strip $1.11B, $888.6M of it in the 14:00 CT hour · by name: SLV $187.9M, LULU $168.0M, TLT $116.3M, NFLX $86.8M · 183 prints sized at 1.5x open interest · net after structure: LULU +$0.6M, SLV -$6.2M, NFLX -$1.2M (all flagged as financing artifacts)

3. TLT: Wednesday’s calls were held, and 75,000 more were bought a year out

The buyer of upside in the long bond stayed and went further out. Wednesday’s November 83 call line stood at 96,741 contracts of open interest on Thursday, so it was neither closed nor added to. Beside it, 75,000 September 2027 86 calls were opened in four blocks between 11:09 and 11:28 Central, $12.2M, with no other leg in the same second. They are single options made entirely of time value, not stock substitutes; the deep in-the-money 60-strike calls that traded beside them are stock in another wrapper and are not counted. They priced at 14.5 implied against 15.1 at the money for that expiration, 0.96x: an upside wing bought under the cost of the middle of the curve, while TLT’s one-year volatility is at the 99th percentile of its own history. The 86 strike is above TLT’s whole fourth-quarter band (72.29 to 83.27), so it pays at expiration only on a move the quarterly band treats as unusual, which fits a one-year option. It is not a call that the low is in. TLT closed at 77.71, zone 76.94 / 79.24 / 81.54, under its weekly lower (77.94) before its distribution is counted. The position pays if TLT turns; it does not make it turn.

TAPE TLT Sep-2027 86 calls 1.59 to 1.63, 75,000 lots (61,000 at the offer), 11:09-11:28 CT, opening, open interest 4,190, 14.5 implied vs 15.1 at-the-money · Nov-30 83 call open interest 96,741 · TLT option figure after structure: flagged as a financing artifact, not readable as direction

4. TSM: a second day of long-dated calls, and the first day’s were kept

TSM is the one chip name where calls were simply bought. On Wednesday a buyer took 11,326 November 460 calls at the offer on a line with almost no open interest. Thursday’s open interest on that line read 11,861: the block was held. Then 4,000 June 2027 480 calls were bought at the offer in one print ($21.4M), with no call sold and no put against them, and 2,410 March 2027 520 calls in pieces through the day. The June call priced at 37.4 implied, about 1.11x the December at-the-money, so the buyer paid the going rate for the long end and no event premium. After structure TSM’s net was +$25.2M with no flag, one of two clean single-name figures on the tape. It is two sessions of one buyer ahead of the 10/15 report, so it is described as held and added to, not counted as a campaign. TSM closed at 459.20, zone 420.32 / 448.73 / 477.14, its highest close of the week.

TAPE TSM Jun-2027 480 calls 53.51 at the offer, 4,000 lots, 09:26 CT, opening, 37.4 implied · Mar-2027 520 calls 2,410 lots at the offer · Nov-06 460 call open interest 22 to 11,861 · net after structure +$25.2M (calls bought, no flag)

5. AMD and FSLR: long-dated puts sold in size

Someone was paid to promise to buy AMD two years from now, much lower. At 13:29 Central, 500 AMD December 2028 600 puts and 750 December 2028 400 puts were sold in the same second, both at the bid. Both legs on the bid side means two sales, not a spread. The seller collects premium and takes on the obligation to buy AMD at an effective 443.75 and 341.40. They priced at about 1.02x to 1.03x the at-the-money volatility, so the seller was paid the ordinary rate, not a fear premium. AMD’s net after structure was +$24.3M with only 4% of it unsided, the cleanest large figure of the day. AMD closed at 615.73, zone 533.69 / 612.83 / 691.96. In FSLR, 10,400 September 2027 150 puts were sold at the bid ($21.3M) on a day the stock fell 1.74%. A long-dated put sale is a statement about a price someone is willing to own at, a year or two out. It says nothing about next week.

TAPE AMD Dec-2028 600 puts 156.25 and 400 puts 58.60, both at the bid, 500 and 750 lots, 13:29 CT, 55.8 / 55.4 implied · AMD net after structure +$24.3M, 4% unsided · FSLR Sep-2027 150 puts 20.60 / 20.05 at the bid, 10,400 lots, 54.0 implied (0.99x)

6. MRVL: a 15-month straddle, sized so it has no direction

The MRVL print that looked like put buying is a straddle, a position in volatility. A January 2028 280 call and a January 2028 280 put traded together three times, 10,126 calls and 18,517 puts, $223.0M, with no quotes shown and no open interest before the trade. There are 1.83 puts for each call. With the strike 5% over the stock, that is the ratio at which the pair has no net exposure to the stock’s direction. It was priced at 66 implied, exactly the at-the-money level for that expiration. Whether it was bought or sold is not on the tape. What the sided tape does show is holders selling December 300 and October 295 calls into a three-day, 6.4% run. MRVL closed at 268.08, inside its weekly implied-volatility band (240.15 to 283.73).

TAPE MRVL Jan-2028 280 calls 80.60 / 81.40, 10,126 lots; Jan-2028 280 puts 77.30 / 75.70, 18,517 lots; 10:03, 13:02, 13:04 CT, opening, no quotes · 66 to 67 implied (1.00x) · MRVL option figure after structure: 79% unsided, not readable

7. EWZ: three call strikes sold against one bought, 134,200 lots a leg

The largest package by contract count was a call ladder on Brazil. At 14:18 Central, EWZ November 39, 41 and 43 calls were sold and November 45 calls bought, 134,200 contracts each, for a net credit near $49.7M. It is a four-leg package, so it is structure and is left out of the directional figure. Either an existing 45-strike call sale was moved down, or a holder sold calls at three strikes with one owned above them; Friday’s open interest will say which. In both readings it is a holder taking income, and the same session’s block tape in the fund was tagged at the offer. It caps EWZ between 39 and 43 into November. EWZ’s near-dated options are priced far above its later ones (66 implied for October 9 against 45 for December), so the seller was paid well; the event behind that premium is not named in any source read here. EWZ closed at 37.14, zone 35.96 / 37.14 / 38.32.

TAPE EWZ Nov-20 39 / 41 / 43 calls 2.03 / 1.35 / 0.85 at the bid, 45 calls 0.53 at the offer, 134,200 lots each, 14:18 CT · implied 50.1 / 49.2 / 48.1 / 47.8 (0.96x to 1.00x the November at-the-money) · a four-leg package, excluded from the directional net

8. SPX: the financing book rolled forward, with no stress in the rate

The SPX 7,000 and 8,000 strikes are loans, on five expirations now. A December 7,000 / 8,000 box spread traded 4,700 times, the October box was partly rolled to January, and a new November 27 box was opened. A box spread is four options that together pay a fixed amount at expiration, so it is a way to lend or borrow cash. The rates implied were 4.4% to 4.8% a year, the same as Wednesday, which says there was no funding strain on the first day of the quarter. Call and put open interest at those strikes is nearly equal (1.06 and 0.92 times), the mark of a loan and the opposite of a hedge. Nothing at 7,000 or 8,000 is a level. A separate 1,000-lot trade bought July 2027 7,000 puts and sold July 2027 9,000 calls in the same second: a collar on a long portfolio, not a bet on Friday.

TAPE SPX Dec-18 7,000 / 8,000 box spread 990.40, 4,700 lots, 4.54% · Oct-16 998.10, 4.63% · Jan-2027 986.20, 4.82% · Nov-27 993.20 and 993.02 · box spreads $1.53B, 41.7% of SPX premium · Jul-2027 7,000 put / 9,000 call collar 1,000 lots, 20.4 / 13.0 implied

9. GDX: 44,811 puts moved between November and January

One hedge on the gold miners was rearranged, not added to. GDX November 85 puts and January 2027 80 puts traded 44,811 contracts each in the same second, both in the middle of the quote, so the direction of the move cannot be read: either November protection was extended to January, or January protection bought on Wednesday was pulled closer. The size is 90% of the November line. Both legs priced at about 1.00x the at-the-money volatility. GDX closed at 86.74, zone 84.46 / 91.31 / 98.15, under its weekly lower (88.66), where there is no basis for a call either way while the strong-dollar block is on.

TAPE GDX Nov-20 85 puts 3.88 and Jan-2027 80 puts 4.15, 44,811 lots each, 12:50 CT, unsided at mid, 40.8 / 41.4 implied · open interest 49,963 / 46,068 (January line 14,969 on Wednesday)

Scorecard — grading Wednesday’s edition on Thursday’s close

Wednesday’s edition: B-, provisional on one session. Its level map was right and its conditions kept every losing idea off the list. It was wrong on earnings and it left out a dividend. Of 35 lines, 10 resolved (6 of 10 right), 8 did not trigger and 17 are still pending, most of them at Friday’s close.

Wednesday’s Top Trades, marked at Thursday’s official closes. Nothing new triggered and no exit line printed. The SPY December 750 put against the sold October 735 put marked 11.98, unchanged. The IWM December 280 call marked 10.88 from 10.01, about 87 dollars better and still 47 dollars under its entry. The NVO January 40 / 45 call spread marked 1.20 from 1.30, about 28 dollars under its entry. The three watch entries did not trigger, and each untriggered condition was the right side of Thursday: GOOGL fell, MSFT opened over its zone high and faded, and the bond did not close over its line for XLU. Monday’s HOOD calendar marked 9.17 against a 10.09 debit, about 92 dollars lower, and is retired at Friday’s expiration of its near leg.

The last five editions grade B, B-, B, B- and B-. The pattern repeats: the conditions and the stand-aside rules are right or harmless, the small testable calls score, and the edition loses where it leans on a frame (this time earnings) or leaves a line under-specified (this time the dividend).

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

The three branches registered on 09/23 resolve at Friday’s close. They were: the market absorbs the rate shock and grinds (about 55%), the shock transmits into the index (about 25%, triggered by an SPX close under about 7,550), and an early turn in rates and the dollar (about 20%). Through eight closes the first branch leads: there has been no close under 7,550, and Thursday’s low was 67 points over it. The early-turn branch has none of its four legs: the 10-year is 5.24% against under 5.00%, the dollar index 102.00 against under 100.05, TLT 77.71 against over 81.35, HYG 76.90 against over 78.35. The weights are carried as registered and graded on Friday, not adjusted after the fact.

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

Friday 10/02 is payrolls, a weekly expiration, and the settle of every weekly band. The report printed before the open: payrolls +29,000 against 84,000 to 90,000 expected, unemployment 4.2% against 4.1%, hourly earnings +0.1% on the month, and the two prior months revised down by 60,000. That is a weak number. The framework’s morning check on it reads unclear (two inputs say relief, one says growth scare, two are neutral), which means stand aside: no fading the move and no chasing it, in either direction. Nothing from Friday’s session is graded in this edition.

Next week (10/05 to 10/09) the weekly bands re-anchor on Friday’s close; Silva’s new weekly board is not published yet, so the standing frame until it lands is the October band, SPX 7,363.93 to 7,939.15. Monday 10/05 brings the ISM services report, and it is the day Silva’s comparison looks for yields to pause. STZ, APLD, PEP and DAL report Tuesday through Friday. The option market prices that week as the quietest on the curve. The first repair signal for breadth is a session where five-session highs outnumber five-session lows (Thursday: 153 against 239). The week after holds the bank reports (Tuesday 10/13), the consumer price report (Wednesday 10/14), TSM (Thursday 10/15) and the monthly expiration (Friday 10/16). The Fed meets Wednesday 10/28 and AAPL reports Thursday 10/29.

Bottom line

Thursday opened the quarter with the scare the tape had been set up for: a prices shock, a 24-year high in the 10-year, the index through its weekly band. Then the bond was bought with no new information, yields closed lower, and the index took the whole move back. More stocks rose than fell, led by chips, software and energy. That is a real change from Wednesday. It is also one session inside a week in which seven of ten stocks are still lower, health care and the defensives broke down on a day rates eased, and eight of the ten largest stocks fell. And the one number that did not follow the bond is the credit spread, which widened to its widest since March while yields fell.

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, CARRIED means it was entered on an earlier edition and is being held to its line. The check measures entry quality, not the odds of winning; its own review found no edge as a picker. Every long leg is sixty days or longer. Every maximum loss sits under the per-structure cap of 1,354 dollars at Thursday’s official closing prices, and holds only if the live debit is under that cap. This is institutional flow to follow, never advice. Once Friday’s contract expires the curve is flat: 3 of 40 names sit at or above 1.10x on next week’s contract against one sixty days out, so nothing here sells a near-dated contract for a time premium. The two index structures take their edge from the strike (the October put is the dear one, the December put is not), and the AAPL structure uses a sold call to fit the cap and says so.

DIAGONAL AAPL Jan-15-2027 335 call / short Oct-30 340 call diagonal — new. READY on all three points. AAPL at 330.32 closed 2.43 under its weekly lower band 332.75 with no cause named, inside its October band and over its trend line; a single stock in that position has usually come back inside, on a small sample. The January 335 call prices about 25.0 implied, 0.98x the December at-the-money (the body, owned). The October 30 340 call prices about 25.5 implied, 0.97x its own expiration’s at-the-money, so it is sold for financing and to fit the cap, not because it is rich; the long call alone costs 1,758 dollars and fails the cap. Long-dated AAPL vol sits in the middle third of its own history, which calls for a spread or a diagonal and never a single call. Debit about 11.63, maximum loss about 1,163 dollars. Owned into the earnings report on Thu 10/29, whose premium sits in the sold call (it expires Fri 10/30, the session after); the sold call is bought back if AAPL closes over 340 before the report; retired on a close below 325.75.
DIAGONAL SPY Dec-18 750 put / short Oct-16 735 put diagonal — the index hedge. CARRIED: entered at about 11.52 on Tuesday’s close, it marks 11.98, unchanged on the day. The December 750 put prices about 15.75 implied, 1.08x the December at-the-money (fair, owned). An October 16 put 5% under SPY prices 18.3 against an October 16 at-the-money of 13.9, 1.31x (the dearer contract, sold). Long-dated SPY vol is at the 3rd percentile of its own history. For a new entry the debit is about 11.98, maximum loss about 1,198 dollars, valid only while SPY closes over its weekly lower 760.54; under it nothing new is sold. Owned into the bank reports, the consumer price report and the monthly expiration (the short put dies Fri 10/16), then the December window; retired on a SPY close over its zone high 776.56 or at the December expiration. It registers no index direction.
DIAGONAL QQQ Dec-18 700 put / short Oct-16 690 put diagonal — the same hedge in the other wrapper. WATCH: an alternative to the SPY diagonal, never an addition to it; entered only if that one is not held. An October 16 put 5% under QQQ prices 23.6 against an at-the-money of 18.8, 1.26x (sold); the December 700 put priced about 24.2 implied at Friday morning’s quote, about 1.15x the January at-the-money (owned). Long-dated QQQ vol sits in its low third. Debit about 11.67, maximum loss about 1,167 dollars. QQQ at 742.03 is 13.01 over its weekly lower 729.02, not at a band-break low. Owned into the monthly expiration (the short put dies Fri 10/16) and the late-October reports; retired on a QQQ close over its zone high 766.07 or at the December expiration.
CALL IWM Dec-18 280 call, one lot — CARRIED to its line and not re-offered. It marks 10.88 from 10.01, still about 47 dollars under its 11.35 entry. The December at-the-money prices 19.9 implied, and long-dated IWM vol is at the 13th percentile of its own history. IWM at 279.02 touched its weekly lower 275.98 and closed 3.04 over it. No new entry until TLT closes back over 77.94: the cause, not the name. Owned into the December window through the election (11/03); harvested at its zone high 287.87; retired on a close below 274.14.
SPREAD NVO Jan-15-2027 40 / 45 call spread, one lot — CARRIED to its line, losing, and not re-offered. It marks 1.20 from 1.30 on a fourth lower close (NVO 37.40). Long-dated NVO vol is at the 2nd percentile of its own history, near 37.7 implied. 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 a monthly close under 34.15, first gradeable at the close on Fri 10/30.
SPREAD MSFT Dec-18 520 / 540 call spread — WATCH for a fifth edition; the missing point is the level. MSFT at 512.80 has held five closes over its 505 to 508 breakout but has not come back to it. The spread marks 8.22 from 8.80. The 520 call prices about 28.9 implied, 0.94x the December at-the-money, with long-dated MSFT vol in its low third. Entered on a session that trades down to the 505 to 508 shelf (Friday’s daily lower is 506.04) and closes back over it; not chased over the zone high 521.30. Maximum loss about 822 dollars. Owned into the December window, closed by Tue 10/27 before the report on the 29th; retired before entry on a close under 500.25.
SPREAD GOOGL Dec-18 340 / 370 call spread — WATCH on the rate and credit legs: the plan enters only after a close with HYG at or above 78.35 and TLT at or above 80.33. Thursday closed at 76.90 and 77.71. GOOGL at 338.24 made a five-session low while holders sold calls against it; its weekly lower 332.07 is the Friday level. The 340 call prices about 33.1 implied, 0.99x the December at-the-money, with long-dated GOOGL vol in its middle third, which calls for a spread. Debit about 11.33, maximum loss about 1,133 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.
SPREAD XLU Jan-15-2027 40 / 43 call spread — the rates-relief idea. WATCH, conditional on the cause: entered only on a TLT close back over 77.94 with XLU closing over 39.71. XLU at 39.68 is the one rate-sensitive fund inside its weekly band, and the day’s largest fund option print was a November 40 / 44 call spread in XLU whose side cannot be read. The 40 call prices about 18.1 implied, 0.95x the January at-the-money; XLU’s near-dated contracts stay about 1.10x its later ones past payrolls, so the shape is a spread, never a single call. Debit about 1.00, maximum loss about 100 dollars a spread. Owned into the January window and closed before its final sixty days (by Mon 11/16); harvested at the 43 strike or twice the debit; retired on a close under 37.87. Without the TLT close it is not a trade.

Not listed, and why:

Sources

Expected moves & levels: the Silva daily board for 10/02 (forward) and 10/01, 09/30 and 09/29 (graded); the ZONE DOCUMENT and Zone Visual for 10/02 and the zone rows for 10/01; the weekly board for 09/28 to 10/02; the October monthly board and the fourth-quarter board (an image, read and checked value by value); the census implied-volatility blocks and the computed implied-volatility bracket for 10/02; the FOM sentiment gauge for 10/01 (39.8). Tradytics: the Live Options Flow and Darkpool Market Summary end-of-day files for 10/01, with the archived 09/29 and 09/30 files for comparisons. The options dashboard for 10/01 was not captured: every dealer, timeline and panel statement in this edition is a reconstruction from the option file or one of Silva’s published levels, and is labelled as such. Recon pipeline: the 2026-10-01 census (650 per-ticker reports, all scanned for breadth; SPX, NDX, RUT, ENTG and PCG have no census file and are declared; CTVA is left out of the breadth counts because of a corporate separation), the twelve sector chunks for the bottom-up sector map, and the census closes for 09/25 through 09/30 for the five-session structures. Ledgers and research: the flow, dark-pool, monetization, setup-scan, expected-move, implied-volatility, long-dated-volatility and volatility-tracker ledgers for 10/01; the census screen and both basket monitors; the entry check and the structure-cap check; the thesis, timing and expected-move analysis ledgers; the band-break base rates; the oversold-list study of 10/01; the September 24 trade-selection review. Intraday: the five briefs and level frames for 10/01; the pre-open and 09:15 briefs for 10/02, as forward context only; Silva’s morning briefs for 10/01 and 10/02 (dealer levels) and the Maverick’s morning briefs for both days. Commentary: 27 pieces dated 10/01 and 10/02, read in full: Arete Trading, Benjamin Cowen, Darius Dale of 42 Macro (three), Dividend Data, ET Tradytics (two), FX Evolution, Geeks of Finance, Ilya Spivak, James of InvestAnswers, Krown, the Maverick (the 10/01 video), Mike Jones, Rob’s Child, SpotGamma (Brent Kochuba), Tim Knight (two), Trading Apologist, Tyler S, Volsignals, and tastylive (five). No Silva commentary or slide deck was published for 10/01. Timing: the Savino September projection and its inverse (both end 10/02; no October or bond chart on file). Market data: official option closing prices for 10/01 and Friday-morning quotes for every structure on the list, read from the broker’s market-data feed. Web: Reuters on jobless claims (via finance.yahoo.com) and on the ISM report and the move in yields (via longbridge.com); CNBC, NPR and Benzinga on the September payrolls report (cnbc.com, npr.org, benzinga.com); the St. Louis Fed’s FRED database for the high-yield spread, the 10-year real yield, the breakeven rate and the Fed’s balance sheet (fred.stlouisfed.org). Prior editions: the 09/30 daily report and its working file, graded above.