SomerQuant Daily — Options Flow + Dark Pool, 17 September 2026

SomerQuant Daily · Thursday, 17 September 2026 · EOD options flow & dark-pool research · SomerQuant

This edition is published late. The desk was dark while AZTMM.com moved to SomerQuant, and the figures below are the 17 September session exactly as measured, unrevised.

17 September 2026 — EOD read. The gain was made before anyone rang a bell: SPY opened at 763.15, 1.21% above Wednesday’s 754.05 close, and then spent six and a half hours going nowhere, finishing at 762.60, a whisker below its own open. Everything that looks like a strong up day on a daily chart happened overnight, and the cash session itself traded a 3.61-point range, dipping to 759.96 by 10:05 and only tagging its high of 763.57 ten minutes before the close. The one genuinely unusual thing on the tape was in SPY options, where call premium ran $4.27B against $908.7M of puts on the session before the fund went ex-dividend.

SPY opened at 763.15 and put in a 15-minute opening range of 760.62 to 763.41, then broke the low of that range within the half hour to reach the session low of 759.96 at 10:05. From there the path was a slow grind back, with the high of 763.57 not arriving until 15:50, and the official close at 762.60. Volume peaked at 3.5M shares in a single five-minute bar. Fifteen SPY dark-pool prints cleared the regular session on our screen and every one of them was contingent, priced off a derivative leg rather than the tape. The largest was $400.0M at 761.55 at 11:05, and fourteen of the fifteen carry the same 754.07 handle, a price that sits $5.89 below the day’s actual low. That is the reason those marks are drawn hollow and are never read as a level where stock changed hands. None of the five heaviest premium shelves falls inside the day’s range: all five sit below it, at 730 (+$116.4M), 725 (+$73.4M), 720 (+$98.4M), 715 (+$80.0M) and 705 (+$55.6M).

The tide opened at +$50.3M and never once looked back at the zero line. It climbed to its peak of +$207.0M by 10:20, faded through the afternoon to a trough of +$38.4M at 2:30, and finished the 4:00 reading at +$58.3M. All 79 of the five-minute readings sat above zero and none sat below, so the series crossed zero zero times. A whole session on one side of the line is rare in this record, and it is the cleanest corroboration we have that Thursday’s call-side premium was not a single print but a persistent bid.

What happened

Two things ran in opposite directions, and reconciling them is most of the work. SPY call premium came in at $4.27B against $908.7M of puts, a ratio of 4.7 to 1, on a session where the prior day had seen $1.19B of calls against $1.49B of puts. That is a 3.6-fold jump in call premium in a single session. Yet QQQ, on the same tape and up more in percentage terms at +1.73%, still printed more put premium than call premium: $898.3M against $734.1M. An index rally that leaves the Nasdaq proxy put-heavy is not a straightforward risk-on session.

The strike ladder explains the SPY figure without requiring anyone to be bullish. Every one of the ten heaviest SPY strikes by net premium is call-led, and the weight is not near the money. The heaviest is 730 at +$116.4M net, 4.27% below the close, with 720 at +$98.4M and 715 at +$80.0M behind it. Deep in-the-money call premium concentrated one day before an ex-dividend date is a familiar shape: the vendor records 18 September as SPY’s ex-dividend date at $1.888834 a share. We are describing the structure, not the intent behind it.

The dark pool was heavy and overwhelmingly passive. The $100M-and-above universe totalled $27.59B across 101 prints in 49 names, and 65.5% of that notional was ETF rather than single-name. The three largest clusters were all index trackers printing within a hair of the close in the minutes just after the bell: SPY 14 prints at 762.60 for $5.59B, IVV 7 prints at 764.07 for $4.39B, and VOO 7 prints at 701.02 for $3.66B. That is closing-cross plumbing, not a view.

Why it matters

A session that gaps and then refuses to extend tells you something the daily bar hides. Buyers were willing to pay 1.21% more than Wednesday’s close before the open, and then nobody was willing to pay much more than that for the rest of the day. The 763.57 high, set at 15:50 and never exceeded, is the level that matters into Friday, and 759.96 is where the only real selling of the day stopped.

The options picture deserves its caveat stated plainly. A 4.7-to-1 call-to-put premium ratio in SPY would normally read as conviction, but the premium sits 4% to 7% below spot on the session before an ex-dividend date, and QQQ put premium still led. Read the SPY number as structure rather than as sentiment until the open-interest update says otherwise. Volatility agreed with the calm reading: the entire 17 September VIX range of 15.38 to 16.29 sat below the 16 September low of 16.40.

What to watch into Friday’s session

• 763.57 — the session high, set at 15:50 and never exceeded. A close above it would mark the first extension of the gap.
• 759.96 — the session low and the bottom of the only meaningful selling. A break below returns the tape to Wednesday’s range.
• 760.62 to 763.41 — the 15-minute opening range, which contained the close and both afternoon attempts.
• The 730 strike shelf at +$116.4M net, 4.27% below spot: whether it survives Friday’s ex-dividend date or unwinds with it.
• QQQ put premium leadership at $898.3M against $734.1M of calls, which has now persisted through two up sessions.
• The VIX 15.38 low, the lowest reading of the week on our screen.

Names on our radar

Every row below is frozen at publication and scored mechanically in the Accountability Ledger at +5 and +21 sessions — misses stay on the page. Options prints are checked against the next morning’s open-interest update in the following session’s post.

Positioning into Friday

The leverage gauge is empty in both directions. Across the flagged-events screen and the complete $100M-and-above dark-pool universe, leveraged-long trackers, inverse trackers and volatility products each drew exactly zero prints and $0 of notional. A reading of zero on all three is information rather than an absence of it: nobody was expressing this session through the leveraged wrappers, which reads as an ordinary tape rather than a hedged or a crowded one.

What positioning there was ran through index trackers at the close and through deep in-the-money SPY calls into an ex-dividend date. Neither is a directional statement. The genuinely directional prints on the board were single-name and modest against the index flow: technology took $6.14B of the $9.51B single-name notional, with STX the largest single name at $1.20B across two prints.

The boards — data appendix

Thursday’s $100M-and-above board of $27.59B sits between the two extremes of the short record. It is 30.7% below Tuesday’s $39.81B, which remains the widest session we have logged, and 48.9% above Monday’s $18.52B. Print count fell from 127 to 101 and distinct names fell from 85 to 49, so the notional was carried by fewer names doing more. ETF share rose to 65.5% from 62.9%, the highest of the four sessions on record, which is consistent with a board dominated by closing-cross tracker prints rather than single-name accumulation.

Every one of the ten strikes shown carries more call premium than put premium, which is not the usual shape. The heaviest is the 730 strike at +$116.4M net, 4.27% below the 762.60 close, built from $123.4M of call premium against $7.0M of puts. The nearest shelf leaning the other way is not in the top ten at all: within the same band it is the 760 strike at −$12.2M net, 0.34% below spot, where $113.1M of put premium edges out $100.9M of calls. Across the full ±8% band, 42 of 122 strikes are put-led, so the call lead is concentrated rather than universal.

The universe below is every off-exchange print of $100M or more on the 17 September tape, retrieved complete by pagination, with canceled rows dropped and exact duplicates removed: 105 raw rows became 101 after dropping 3 canceled prints and 1 duplicate.

Relative size

Relative size is each print’s share count divided by that name’s 30-day average daily volume.

These two screens disagree by construction and the disagreement is worth naming. The boards universe above is a $100M notional screen across all sessions and issue types; the relative-size tape is a $10M screen limited to regular-hours common stock. GBTG tops the relative-size table at 1.58× on $30.3M and never appears on the $100M board, while EIX appears on both at $111.0M and 0.29×. The headline board figures in this edition come from the $100M screen.

Cluster watch

The three clusters carry $13.64B between them, just under half the whole board, and all three are index trackers printing within a few cents of their closing marks in the twenty-three minutes after the bell. Clustered prints are not re-narrated as standalone blocks elsewhere in this edition.

Notable data points

1. SPY closed at 762.60, up 1.13% on Wednesday’s 754.05, having opened at 763.15 and closed 0.07% below that open.
2. The whole cash session fitted inside 3.61 points, from 759.96 at 10:05 to 763.57 at 15:50.
3. SPY call premium of $4.27B was 3.6 times the prior session’s $1.19B; put premium fell to $908.7M from $1.49B.
4. QQQ rose 1.73% to 716.92 and still printed more put premium ($898.3M) than call premium ($734.1M).
5. Net options premium held above zero for all 79 five-minute readings, peaking at +$207.0M at 10:20.
6. The $100M-and-above dark-pool board totalled $27.59B across 101 prints in 49 names, 65.5% of it ETF notional.
7. All fifteen SPY dark-pool prints on the regular-session screen were contingent, fourteen of them at 754.07, which is $5.89 below the day’s low.
8. The entire 17 September VIX range of 15.38 to 16.29 sat below the 16 September session low of 16.40.

Method note

Flow and dark-pool data sourced from Unusual Whales. MPI score and regime classifier are our internal composite; daily synthesis is AI-assisted from those inputs.

This edition was built after the session it covers. Every pull was bounded to the 17 September trading date rather than to the time of writing, and each returned payload was checked to carry 17 September timestamps before use. The index feed was current for this session: its as-of date is 17 September and the SPY spot it carries, 762.60, matches the session close, so the MPI and regime cells are the session’s own and carry a single as-of label. No figure in this edition has been revised after the fact.

The flagged-events screen is the 50 largest events by premium within the 17 September regular session at a $1M minimum, and the loudest-line figure is stated within that disclosed depth. Volume-over-OI activity is unconfirmed as new positioning until the next morning’s open-interest update; confirmations are published in the following session’s post.

Relative size is a print’s share count divided by that name’s 30-day average volume as reported with the print, computed by SomerQuant; it is not any vendor’s relative-size or relative-strength score. Option premium by strike is aggregated across all expiries. The session map is drawn from 5-minute bars; the opening range is the first fifteen minutes; dark-pool prints marked contingent carry the vendor’s contingent or derivative-priced sale condition. Board baselines are trailing-session averages computed from our own archived pulls; derived indicators are computed by SomerQuant from licensed market data.

The VIX close the vendor returned for this session is the integer 15, which sits below the same session’s low of 15.38 and is therefore unusable. The strip and the text carry the session range of 15.38 to 16.29 instead, and no percentage change is computed from a truncated close. The prior session’s low of 16.40 is used only as a range comparison, not as a divisor.

The boards strip covers four sessions rather than five because the 16 September edition was not yet on the site when this recovery edition was assembled. For the same reason this edition carries no open-interest confirm section: that section reads the prior session’s radar manifest, and no 16 September manifest existed to read.

This is research, not advice. Nothing here is a recommendation to buy, sell, or hold any security.

AZTMM HLDGS LLC is not a registered broker-dealer, investment adviser, or FINRA member. All content is retrospective research published for general circulation — not personalized advice, not trade signals. Options involve substantial risk, including losses that may exceed the initial investment. Full disclaimer.

New here? Start Here · Pulse Lab · Accountability Ledger · Trading Academy · Subscribe for daily updates