2026-09-14 · midday

Gabriel traded for the first time in ten days, and the new position is one no rail can see. A 3-lot GOOGL September-2027 collar filled at 12:25 ET - three 305 puts bought, three 445 calls written, $103.70 net credit - and his own same-day straddle note records the part that matters: it needs 300 shares and only 243 are free, so 57 shares of it stand behind nothing. The book says the same thing from the other direction - 16 net short calls now cover 1,600 shares against 1,543 held, where at this morning's pull 13 covered 1,300. Two approved docs are now out of date because of it, and this run edits neither. Everything else eased: the afternoon took the morning's semiconductor selloff back by roughly half, GOOGL extended to $348.19, and net cash crossed back over its amber floor - by $21.33, on the cash the Berkshire sale brought in. Two rails fire instead of three, every runbook row is clear, and the nearest price level is tranche 3 at $360, 3.4% up and still being approached.

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 · ~$17,934.84 debit ($35.8697 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)"”
Seventeenth read of the same standing order, priced on a live tape 20 minutes before the close. The debit fell $553.47 since this morning as GOOGL rose $3.78 intraday, continuing the direction the structure says it should move. Reprice at the moment it is placed - these are 12:38 PDT marks, not a fill. 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	1.686x	amber 1.3x, red 1.5x	red
top2_name_delta_ratio	0.62x	amber 0.70x, red 0.85x	clear
total_delta_ratio	0.88x	amber 1.30x, red 1.50x	clear

Why each one reads the way it does

long_leg_band
The rail the action item addresses. No leg has been traded on it.
gross_leverage
1.57x when the policy was signed. Gross position value rose $11,214 to $2,435,754 as GOOGL extended, but net liquidation rose faster, so the ratio eased from 1.703x this morning and is still red. This rail counts both legs of every spread.

delta by name

Delta notional per name as of 2026-09-14: 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	$756,119	52.3%	1543 sh, 18 legs
MU	$137,976	9.6%	525 sh, 2 legs
META	$96,013	6.6%	144 sh, 0 legs
VOO	$86,120	6.0%	123 sh, 0 legs
RDDT	$84,168	5.8%	513 sh, 0 legs
NVDA	$48,482	3.4%	285 sh, 2 legs
INTC	$29,268	2.0%	300 sh, 0 legs
AMZN	$25,394	1.8%	100 sh, 0 legs
EQT	$10,618	0.7%	200 sh, 0 legs
TSLA	$360	0.0%	1 sh, 0 legs
DIS	$109	0.0%	1 sh, 0 legs
BRK B	$0	0.0%	0 sh, 0 legs

Everything else clear

Nearest trigger is tranche 3 confirmation (nearest written price level, and approaching) at $360, 3% away.
No position moved more than 5% today. Biggest was INTC -5.23%.
INTC -5.2%, now $97.56. No written rule for it.

All 35 triggers swept, with today's number

price trigger	now	fires at	distance
GOOGL · tranche 3 confirmation (nearest written price level, and approaching)	$348.19	≥ $360	3.4% up
GOOG · tranche 3 confirmation	$344.71	≥ $360	4.4% up
GOOG · tranche 2 weakness	$344.71	≤ $317	8.0% down
GOOGL · tranche 2 weakness	$348.19	≤ $317	9.0% down
GOOG · thesis break, close regardless of price	$344.71	≤ $280	18.8% down
GOOGL · thesis break	$348.19	≤ $280	19.6% down
GOOG · washout, stop selling spreads and buy stock	$344.71	≤ $260	24.6% down
GOOGL · close every short put	$348.19	≥ $434	24.6% up
GOOGL · crash powder, deploy the $100k reserve	$348.19	≤ $260	25.3% down
MU · collar roll-up	$926.90	≥ $1,250	34.9% up
GOOGL · long spreads near max, apply the buy-it-fresh test	$348.19	≥ $500	43.6% up
GOOGL · shares capped by the 600Cs, roll the collar up	$348.19	≥ $600	72.3% up

level trigger	now	fires at	distance
Net cash / NLV (nearest trigger on the board)	-0.049985	< -0.05 amber	$21.33 of cash
Dec-28 30-delta put IV	35.65%	> 43%	+7.3 pt
GOOGL IV, crash trigger	35.65%	> 50%	+14.3 pt
Dec-28 300/400 call spread, mark vs width	43.46%	>= 90%	+46.5 pt
8x Sep-27 300/440 call spread, mark vs width	41.23%	>= 90%	+48.8 pt
10x Dec-28 350/500 call spread, mark vs width	30.76%	>= 90%	+59.2 pt
GOOGL Dec-28 340P extrinsic	$54.05	< $5	$49.05
Excess liquidity / NLV	76.39%	< 40%	+36.4 pt
NLV change on the week	+1.13%	-8% in one week	+9.13 pt
Defined max loss vs program cap	$175,059	$200,000	$24,941 headroom
Short GOOG/GOOGL calls vs shares held	1,600 shares called against 1,543 held	no written rule exists	57 shares uncovered
Biggest single-name move	INTC -5.23%	>= 5%	BREACHED, reported under moves

calendar	date	away
Wash-sale window opens for the 6/05 GOOG loss lots - the prior-30-day precondition still fails, see gaps	2026-09-17	3d
Micron fiscal Q4 print (not a trigger row)	2026-09-30	16d
Clean sale date for the loss lots IF only the 08-31 300C buys count as replacement property	2026-10-01	17d
Clean sale date IF today's written 445C also counts - the doc does not say which	2026-10-15	31d
Alphabet Q3 print, scheduled thesis adjudicator - re-read the house view within 72h	2026-10-28	44d
No crash by this date: deploy the $100k reserve at market anyway	2026-10-29	45d
Earliest date the GPU glut tripwire can evaluate a 90-day baseline	2026-11-16	63d
Alphabet capex guide, second scheduled adjudicator	2027-02-03	142d
GOOG lots vest long-term (+66 sh): add collar put leg 9 - unaffected by today's collar	2027-03-13	180d
Runbook says +350 sh vest (legs 12-14). 193 of those 350 are now clock-stopped by today's collar; the row no longer holds as written	2027-06-05	264d
Runbook says +50 sh vest (leg 15). All 50 are clock-stopped by today's collar; the row no longer holds as written	2027-06-25	284d

Spot prices

	last	day
GOOG	$344.71	+2.8%
GOOGL	$348.19	+2.9%
MU	$926.90	-5.0%
AMZN	$253.94	-1.1%
BRK B	$515.20	+0.9%
DIS	$108.87	+2.2%
EQT	$53.09	-1.8%
INTC	$97.56	-5.2%
META	$666.76	+2.9%
NVDA	$211.81	-3.0%
RDDT	$164.07	+4.0%
TSLA	$360.17	-1.4%
VOO	$700.16	-0.3%

The check itself

confirmed The run label came from the log, and it mattered more than usual today.
The rails log already carried a 2026-09-14 markets-open entry stamped 06:42:30, so this is the midday run: rails only, no GPU pull, no forecast pass. Had the clock been used instead it would also have said midday, but the log is what was read.
found this run The report JSON's chips block is dead - the generator has never read it.
Found this run by grepping the generator for the key: build_markets_report.py contains no reference to 'chips' anywhere, and all ten of this run's chip lines were checked against the rendered HTML and none appears. Every report in this series since the schema was set has written roughly ten one-line summaries that go nowhere on the page. They are not wasted entirely - the same lines are copied into the check-in doc, which is read - but anyone assuming the page shows them is wrong. Not fixed here: the generator owns the layout and adding a section to it mid-run is a styling decision for Gabriel, not a correctness fix.
confirmed First live-session midday read with actual fills in the window, and the book diff caught all three.
positions.json was copied to positions.prev.json before the new pull and diffed contract_id by contract_id. 32 rows became 34: BRK B went 40 to 0, and two new rows appeared, GOOGL Sep17'27 305 PUT at +3 and GOOGL Sep17'27 445 CALL at -3. Both new rows carry daily_pnl exactly equal to unrealized_pnl (-$179.792915 and -$51.42227), which is the signature of a position opened today. All fourteen instruments returned top_status REALTIME; the Dec-28 340P option quote came back DELAYED and is flagged under gaps.
confirmed Prior closes were derived, not reported, and every one matched the open run.
The connector's prior-close field returned empty on ten of the thirteen equities again; BRK B, EQT and RDDT populated it. Each empty one was derived as last minus the same response's change field. All thirteen derived values reproduce the 06:42 run's prior closes to the cent, which is the cross-check. No price was estimated.
confirmed The campaign IV was read, not solved.
GOOGL Dec-28 340P returned option-midpoint-iv annualIv 0.3565322832631698 with isValid true, cross-checked against the separate implied_vol field at 0.35803345443017315. The rails read the midpoint figure, as written.
standing A midday deploy strips the cost-of-compute section off the morning's page.
Measured on 2026-09-13 and unchanged. Both runs of a day write to the same <out>/<date>/index.html, the midday report legitimately carries no gpu block because the spot pull is an open-run job, and the generator renders that section only when the block is present. The data is not lost - it is in report-2026-09-14.open.json and in gpu-spot-daily.csv. The generator was not edited: which run should own the dated page is a design question for Gabriel, not a styling change to make mid-run.

Rails that could not be evaluated

No rail, in any approved doc, covers short-call coverage against the share count. Today's fill pushed 16 net short GOOG/GOOGL calls against 1,543 shares - 1,600 shares called, 57 uncovered. Before today it was 13 net short, 1,300 shares, comfortably inside. rails.json carries exactly one program, the Dec-2028 put campaign, whose every_short_capped rail tests put legs only; it reports all 30 short legs capped and is correct on its own terms. Nothing anywhere tests calls against stock. This run does not add the rule - a rule that is not in a doc Gabriel approved does not go in the file.
The runbook's June-2027 vest calendar no longer holds and the runbook still says it does. Part 2's calendar row reads: 'Calendar: 2027-03-13, 04-02, 06-05, 06-25 | GOOG lots vest LT on these dates -> ADD the corresponding collar put legs ... 06-05 +350 -> legs 12-14 - 06-25 +50 -> leg 15.' Today's straddle identification states its own consequence: the identified lots' holding-period clock stops while the collar is open, per Reg 1.1092(b)-2T. The identified lots are 193 shares from 06-05 (73+56+14+50) and all 50 from 06-25. So legs 12-14 and leg 15 are scheduled against conversion dates that have stopped advancing. The runbook is a doc Gabriel approved, so this run does not edit it; the divergence is the report. The 2027-03-13 and 04-02 rows are unaffected.
tax-rules does not say whether a WRITTEN call is replacement property, and the answer moves a date three days away. The doc reads 'For STOCK losses, ANY call on the stock is replacement property regardless of strike/expiry; bought puts and index options are not.' Today's two fills were a bought put, which the sentence explicitly excludes, and a written 445 call, which the sentence neither includes nor excludes. Under S1091 the test is acquisition and writing a call is not acquiring, but the approved doc does not say that, and this run does not decide it. Operationally it changes nothing this week: the window opens Thursday 09-17, the runbook's precondition already fails on the 8x Sep-27 300C bought 08-31, and nothing is proposed either way. It changes the clean date: 2026-10-01 if only the 300C buys count, 2026-10-15 if today's written call counts too.
stressed_excess_liquidity is still unevaluated. It needs IBKR Risk Navigator at -30% price and +50% implied vol with the TWS gateway up, and has never once been measured. It is the rail that would actually price the Part 3 crash scenarios, and it is the one the new 57-share gap would show up in.
The 8x Dec-28 300P leg is outside the long-leg band and has no written repair order. Every page in this series proposes the 280P repair and says nothing actionable about the 300Ps, because nothing actionable is written. 8 of the 13 out-of-band contracts have no remedy on paper.
short_dec28_340p_extrinsic is listed unevaluated in rails.json but was computable this run. Swept by hand against the runbook: $54.05 against a $5 trigger, GOOGL $348.19 leaving the 340 put fully out of the money. The file's stated reason - that it needs a live Dec-2028 chain pull - is wrong; the position row's own mark is enough. The file was not edited, because this check does not edit the rails to match the book.
The Dec-28 340P quote came back DELAYED while all thirteen equities were REALTIME. Its last trade stamps at 16:38:22Z against equity stamps at 19:38:10-19:38:51Z. The IV reading, the 340P extrinsic and the repair debit all rest on option marks from the position rows rather than that quote, and the position marks are the broker's own, so the practical exposure is the IV figure. It moved 0.17 vol points from this morning's 35.82%, which is consistent with a stale-but-close read rather than a wrong one. Named rather than worked around.
The glut tripwire still cannot fire. It is defined as H100 spot 20% or more below its level 90 days back, and the daily series starts 2026-08-18, so no 90-day reference exists. Earliest evaluable date is around 2026-11-16. The SemiAnalysis composite it is written against is still the 2026-04 print at $2.82, five months stale.

Where the numbers came from

Run label decided from the rails log, not the clock: a 2026-09-14 markets-open entry stamped 06:42:30 already existed, so this is the midday run - rails only, no GPU pull, no forecast pass.
Book pulled 19:38Z / 12:38 PDT, twenty-two minutes before the close, and diffed contract_id by contract_id against the 06:42 pull. 32 rows became 34, with three quantity changes and 32 of 34 marks moved.
The three fills are read off the book diff and corroborated by investing/straddle-identifications.md, which Gabriel wrote the same day and which carries the order number, the exchange, the fill timestamp and the per-lot attribution. The 57-share figure appears in his note and falls out independently of the position rows; both are reported because they agree.
Short-call coverage was counted straight off the position rows: 45 short GOOG/GOOGL call contracts against 29 long, 16 net short, 1,600 shares against 1,543 held. The pre-trade figure of 13 net short comes from the same count run on positions.prev.json.
Prior closes: the connector's prior-close field returned empty on ten of thirteen equities, so each was derived as last minus the same response's change field. All thirteen reproduce the 06:42 run's values to the cent.
Campaign IV read, not solved: option-midpoint-iv annualIv 0.3565322832631698, isValid true, with implied_vol cross-checking at 0.35803345443017315.
Runbook Parts 2, 3 and 4 were swept row by row against this run's own quotes and position marks rather than copied from the morning. The three call-spread width percentages (43.46%, 41.23%, 30.76%), the 340P extrinsic ($54.05), the excess-liquidity ratio (76.39%) and the one-week NLV change (+1.13% against 2026-09-07's $1,428,489.17) were each recomputed.
Repair debit is the 360P mark minus the 280P mark on 5 spreads: (64.54338835 - 28.67370605) x 500 = $17,934.84, down $553.47 from the morning's $18,488.31 as GOOGL rose $3.78 intraday.
MU complex P&L is the sum of the three position rows' own daily_pnl fields: shares -$25,459.87, Dec-28 1740 call +$13,852.11, Dec-28 900 put +$4,676.35, net -$6,931.41.
net_cash_ratio cleared arithmetically, not structurally: -$72,209.99 on $1,444,626.31 is -0.049985 against a -0.05 amber. The cash arrived from the Berkshire sale and the collar credit. The rail is reported as the nearest trigger rather than as a rail that passed.
No forecast source was fetched and no Seeking Alpha page was opened this run. The forecast pass is an open-run job.