2026-09-17 · midday

Nothing new broke. The same two structural rails fire that fired this morning - 13 long puts above the $260 band, and gross leverage now 1.667x - and every runbook row in Parts 2, 3 and 4 is clear on price. Gross leverage has now risen at three consecutive reads while the account barely moved: NLV is $167 LOWER than this morning and gross position value is $8,324 higher, so the ratio climbed on marks alone with nothing traded. Micron and Intel are still the day's story and still have no written rule between them: MU $978.15 up 5.57%, INTC $109.49 up 8.35%, both named and nothing proposed. The nearest unfired trigger closed a full point since the open - tranche 3 at $360 is 3.70% above GOOGL, against 4.51% at 06:57. Today's calendar row, the wash-sale window for the 6/05 GOOG loss lots, is live for the second time today, and its precondition still fails: re-checked against a fresh 30-day trade pull, the 8 GOOGL Sep-27 300 calls bought 2026-08-31 sit inside the window, so the answer stays do not sell. The forecasting pipeline's blocker is still clear - the tree was clean at 12:38 - but the 12:45 slot fires seven minutes after this check, so today's run remains unobservable from here, and the structured health check still reads a file nothing writes.

What to do

13 long puts sit above the $260 long-leg ceiling: 280P x5 and 300P x8.
BUY TO CLOSE 5 GOOGL Dec15'28 360 PUT / SELL TO CLOSE 5 GOOGL Dec15'28 280 PUT
one combo x5, limit near mid, day · ~$17,936.77 debit ($35.8735 net per spread at this pull's marks)
Your playbook: “"long leg never above $260 (must sit at/above the $230-250 floor zone, never below $230 where protection stops paying before history bottoms)"”
The same standing order, repriced at 12:38 PDT marks. $116.16 cheaper than this morning's $18,052.93. No live orders on the book and nothing has been traded on this rail. It repairs 5 of the 13 out-of-band contracts; the 8x Dec-28 300P still has no written remedy and stays under gaps.

Still over, unchanged

These fire every run until you act on them. No order attached.

rail	now	limit
long_leg_band	280P x5, 300P x8	$230-$260 band	13 contracts outside
gross_leverage	1.6673x	amber 1.3, red 1.5	red
top2_name_delta_ratio	0.62x	amber 0.70x, red 0.85x	clear
total_delta_ratio	0.87x	amber 1.30x, red 1.50x	clear

Why each one reads the way it does

long_leg_band
Unchanged from this morning. No leg has been traded on it.
gross_leverage
Third consecutive rise: 1.6546x at yesterday's midday, 1.6614x this morning, 1.6673x now. NLV FELL $166.84 to $1,468,352.66 while gross position value rose $8,324.14 to $2,448,112.01. Nothing was traded between the two reads - all 35 marks moved and no quantity did. Reaching the red line needs gross down $245,583 or NLV up $163,722.

delta by name

Delta notional per name as of 2026-09-17: stock at delta one, each option leg at the Black-Scholes delta implied by its own mark. Top two: GOOG, MU.

name	delta notional	of NLV
GOOG	$763,636	52.0%	1543 sh, 18 legs
MU	$150,985	10.3%	525 sh, 2 legs
META	$97,834	6.7%	144 sh, 0 legs
VOO	$86,273	5.9%	123 sh, 0 legs
RDDT	$78,145	5.3%	513 sh, 0 legs
NVDA	$50,375	3.4%	285 sh, 2 legs
AMZN	$25,143	1.7%	100 sh, 0 legs
EQT	$10,058	0.7%	200 sh, 0 legs
INTC	$8,520	0.6%	300 sh, 2 legs
TSLA	$367	0.0%	1 sh, 0 legs
DIS	$106	0.0%	1 sh, 0 legs

Everything else clear

Nearest trigger is GOOGL - tranche 3 confirmation (nearest written price level) at $360, 4% away.
INTC +8.3%, now $109.49. No written rule for it.
MU +5.6%, now $978.15. No written rule for it.

All 31 triggers swept, with today's number

price trigger	now	fires at	distance
GOOGL - tranche 3 confirmation (nearest written price level)	$347.17	≥ $360	3.7% up
GOOGL - tranche 2 weakness	$347.17	≤ $317	8.7% down
GOOGL - thesis break	$347.17	≤ $280	19.4% down
GOOGL - Rule 6, close every short put	$347.17	≥ $434	25.0% up
GOOGL - washout / Rule 8 crash deploy	$347.17	≤ $260	25.1% down
GOOG - Rule 6 on the GOOG basis	$343.44	≥ $434	26.4% up
MU - roll the collar up	$978.15	≥ $1,250	27.8% up
GOOGL - long spreads near max	$347.17	≥ $500	44.0% up
GOOGL - shares capped by the 600Cs	$347.17	≥ $600	72.8% up

level trigger	now	fires at	distance
GOOGL - tranche 3 at $360 is the nearest unfired trigger on the board	347.17	close above $360 on the 50-day, 1.2x volume	3.70% above spot, in from 4.51% at the open as GOOGL rose $2.72 intraday
Gross position value / NLV	1.6673	> 1.5 red	fired; would need gross down $245,583 or NLV up $163,722 to reach the red line
Net cash / NLV	-0.0367	< -0.05 amber	$19,526.61 of headroom; cash unchanged since the open, NLV down $167
Excess liquidity / NLV	0.7601	< 0.40 amber	36.0 points of headroom
Excess liquidity floor	1116118.75	< $650,000	$466,119 above the floor
Top-2 name delta / NLV	0.6229	> 0.70 amber	7.7 points of headroom; up from 0.6206 at the open as MU and GOOGL marked up
Total delta notional / NLV	0.8659	> 1.30 amber	43.4 points of headroom
GOOGL Dec-28 30-delta put IV	0.3546	> 0.43 tranche 2, > 0.50 crash	7.5 points below the tranche-2 level; up 0.4 points from the open's 0.3510
Dec-28 340P extrinsic - assignment watch	53.52	< $5.00	zero intrinsic at a $347.17 spot, so the whole $53.52 mark is extrinsic
Longest call spread vs the 90%-of-width sell row	0.432	>= 0.90 of width	Dec-28 300/400 at 43.20%, Sep-27 300/440 at 40.84%, Dec-28 350/500 at 30.55% - all recomputed from this pull's marks, all far from the exit
NLV on the week vs the 72-hour freeze	4.21	-8% in one week	up 4.21% ($59,303) against the 2026-09-10 midday read of $1,409,049.45

calendar	date	away
Wash-sale window opens for the 6/05 GOOG loss lots - TODAY, and the prior-30-day precondition still fails on a fresh trade pull	2026-09-17	0d
MU fiscal Q4 print, after the close	2026-09-30	13d
Earliest clean sale date for the 6/05 loss lots if only bought calls count	2026-10-01	14d
Earliest clean sale date if a WRITTEN call is also replacement property	2026-10-15	28d
GOOG print, first scheduled thesis adjudicator	2026-10-28	41d
Deploy the $100k reserve at market if no crash has come	2026-10-29	42d
GOOG capex guide, second scheduled adjudicator	2027-02-03	139d
GOOG lots vest long-term, add collar put leg 9 (+66 sh)	2027-03-13	177d
GOOG lots vest long-term, add collar put legs 10-11 (+200 sh)	2027-04-02	197d
GOOG lots vest long-term, legs 12-14 (+350 sh) - 193 of them clock-stopped by the 09-14 collar	2027-06-05	261d
GOOG lots vest long-term, leg 15 (+50 sh) - all 50 clock-stopped by the 09-14 collar	2027-06-25	281d

Spot prices

	last	day
GOOG	$343.44	+1.2%
GOOGL	$347.17	+1.3%
MU	$978.15	+5.6%
AMZN	$251.43	+2.2%
DIS	$105.67	-1.2%
EQT	$50.29	-0.2%
INTC	$109.49	+8.4%
META	$679.40	+0.9%
NVDA	$219.70	+2.7%
RDDT	$152.33	-3.3%
TSLA	$366.74	+2.4%
VOO	$701.41	+1.2%

The check itself

will happen again The pipeline's pre-flight blocker is still clear, but the 12:45 slot fires seven minutes after this check, so today's run cannot be observed from here.
git status --porcelain -- data stocks_dashboard/dist returns empty at 12:38 PDT, the same answer this morning got, so nothing has re-dirtied the tree during the session. git log now shows a commit dated today, 138a65d, Test the growth-multiple hypothesis on 56 firms, and refuse the arm it implies. launchctl print gui/501/com.gabriel.stock-dashboard.daily reports the job loaded, state not running, runs = 2, last exit code = 1, calendar interval Hour 12 Minute 45. The first slot that can produce a gate count therefore fires after this run ends; tomorrow's 06:35 run is the first that can read it.
never worked ops/status.py reads a file that nothing writes, and it returned its non-answer again today.
backend/.venv/bin/python ops/status.py --check exits 10 with: no daily record in /Users/gabedottl/code/financials/ops/status.jsonl: the job has not run, or the log was removed. ops/status.jsonl exists but holds zero lines, across two runs that did happen. grep -rn 'status.jsonl' ops/ finds it only in ops/status.py, ops/tests/test_status.py and ops/README.md - never in the 511 lines of ops/run_daily.sh. The reader half is built and the writer half is not, so the check's verdict is identical whether the job ran, refused or never fired. Today that verdict happens to be true by accident, because the 12:45 slot has not run yet. Wiring the writer changes his repo, so it is named and not done.
will happen again Two consecutive days produced nothing, and the logs confirming it are the only evidence there is.
ops/logs holds 2026-09-15.log and 2026-09-16.log, each ending SUMMARY: <date> pre-flight refused, 0 step(s) run (data/ or dist/ dirty at start), with COMMIT_REFUSED-2026-09-15.md and COMMIT_REFUSED-2026-09-16.md naming the five modified data/tickers/{AMZN,GOOGL,META,MU,NVDA}/config.json files. No check_no_data_loss: line appears in either, because no step ran, so gate_examined has still never been a number. Read-only: nothing in ~/code/financials was edited, committed, reverted, rerun or reloaded by this run.
structural Neither market-check slot can ever observe a completed same-day pipeline run.
The job fires at 12:45 Mon-Fri; this check runs at 06:35 and 12:00-12:45. Question 1 answers no by construction and question 4 always measures yesterday's outputs. Unchanged, and still a scheduling decision for Gabriel rather than something this run takes.
will happen again The markets prompt's own 4a fallback points at a directory that no longer exists.
Section 4a's documented fallback globs ~/code/financials/stocks-dashboard/cron/logs/*.log. That directory is gone - the shell returns no matches - and ops/logs/ is the live path, which is what was read. The block needs repointing or removing.

Rails that could not be evaluated

NO RAIL COVERS SHORT CALLS AGAINST SHARES HELD. Counted straight off this pull's position rows: 45 short GOOG/GOOGL calls against 29 long is 16 net short, 1,600 shares against 1,543 held, so 57 shares of the Sep-2027 collar stand behind nothing. Unchanged since 09-14. Separately, INTC is now $109.49 against a $140 short strike on a 3x/300-share collar that is fully covered - but no rail anywhere would say anything if it were not.
THREE Sep-2027 legs remain invisible to every rail: the 8x 300/440 spread opened 08-31, the 3x 305/445 collar opened 09-14, and the INTC 87.5/140 collar opened 09-16. rails.json carries exactly one program, the Dec-2028 GOOGL put campaign, and nothing in it can see a Sep-2027 leg or an INTC leg.
MU rose 5.57% to $978.15 and has NO written price level anywhere below spot. The runbook's only MU rows are the $1,250 collar roll-up, now 27.79% above, and a Part 5 thesis-break test on HBM contract pricing that is a qualitative judgement rather than a level. The second-largest delta in the book moved 5.57% and produced no rail reading of any kind.
INTC rose 8.35% to $109.49 and is in assets_without_written_rules, so a 300-share position at a $34.42 average cost, now collared, has no written rule for any outcome. Named, nothing proposed.
THE 8x Dec-28 300P STILL HAS NO WRITTEN REPAIR ORDER. The long_leg_band fire covers 13 contracts; the written remedy addresses 5 of them. Open every run.
stressed_excess_liquidity has still never been measured. It needs IBKR Risk Navigator at -30% price and +50% IV with the TWS gateway up.
short_dec28_340p_extrinsic is still listed in rails.json as needing a live Dec-2028 chain pull, when the position row's own mark answers it at $53.52 of pure extrinsic. rails.json NOT edited - the file holds only rules from an approved doc, and changing how a rail is measured is Gabriel's call.
tax-rules is still silent on whether a WRITTEN call is replacement property. Today's answer does not depend on it - the 8x GOOGL Sep-27 300C bought to open 2026-08-31 settles it either way - but the clean-sale date does: 2026-10-01 if only bought calls count, 2026-10-15 if written ones do.
The runbook's Jun-2027 vest-calendar rows still list dates for 243 shares whose holding-period clock is stopped by the 09-14 collar under Reg 1.1092(b)-2T, per Gabriel's own note in investing/straddle-identifications.md: 193 of the 350 on 2027-06-05 and all 50 on 2027-06-25. The runbook has not been edited and will not be by this run.
ops/status.py reads a file nothing writes, so section 4a's structured route still cannot answer any of its four questions.
The markets prompt's section 4a log fallback globs a directory that no longer exists.

Where the numbers came from

Run label decided from the rails log, not the clock: the 2026-09-17 entry in investing/checks/rails-log.jsonl already reads markets-open, so this is the midday run. Sections 4b (GPU) and 5 (forecasts) were skipped by that rule, not by failure.
Book pulled 19:38Z / 12:38 PDT, 82 minutes before the close. 35 rows, the same 35 as this morning.
Zero quantity changes and zero rows added or removed against the 06:57 pull; all 35 marks moved. get_account_orders returns an empty list, so nothing is working.
positions.json, summary.json and quotes.json copied to .prev.json as the first writes of the run.
The connector's prior_close field returned empty on 11 of 12 equities, so every prior close is derived as last minus the broker's own change field. All twelve match this morning's independently-derived set to the cent, and RDDT's derived 157.60 matches its own populated prior_close field.
IV from option-midpoint-iv on the Dec-28 340P, the campaign's written IV contract, cross-checked against implied_vol on the same contract (0.35458 vs 0.35507).
The wash-sale precondition was re-verified in this run against a fresh get_account_trades DAYS_30 pull rather than carried forward from the morning: 8 GOOGL calls bought 2026-08-31 at 75.40-75.66 and 2 more bought 2026-08-18 at 58.65 both sit inside the 30-day window.
Weekly NLV comparison struck against the 2026-09-10 midday read of $1,409,049.45, the midday run one week back in this log.
Every call-spread percentage recomputed from this pull's position marks, not carried forward.
Section 4a was read-only throughout: nothing in ~/code/financials was edited, committed, reverted, rerun or reloaded, and the launchd job was not touched.