Council · Round 01 · Horizon 30 September 2026
Outlook & risk
What the calendar has scheduled, what each scenario requires to happen, and the risks the council flags with a level attached rather than an adjective.
Critical windows
every date verified by calculation| Date | Event | Agents affected | Importance |
|---|---|---|---|
| Fri 4 Sep | Nonfarm payrolls (August) — July printed −23k | Macro · Quant | High |
| 8–15 Sep | Q3 buyback blackout window opens | Earnings · Sentiment | High |
| Thu 10 Sep | PPI (August) | Macro | Medium |
| Fri 11 Sep | CPI (August) — first print after the oil shock | Macro · Earnings · Quant | Critical |
| 15–16 Sep | FOMC — decision Wed 16 Sep, 2:00pm ET. Market prices a 33% chance of a hike | Macro · Sentiment · Quant | Critical |
| Wed 16 Sep | Retail sales (August), same day as the FOMC | Macro | Medium |
| Fri 18 Sep | Quadruple witching + quarterly S&P rebalance | Gann · Quant · Sentiment | High |
| 22–23 Sep | September equinox | Gann | Low |
| Wed 30 Sep | PCE (August) + GDP third estimate + quarter end | Macro · Earnings | High |
| Oct 2026 | Seasonal turn of presidential cycle year 2 | Gann | High |
| Wed 4 Nov | Midterm elections | Gann · Macro | Medium |
The clustering matters more than any single date: CPI on 11 Sep, FOMC on 16 Sep and witching on 18 Sep fall inside six trading sessions, and the buyback blackout opens right on top of that block, removing one of the US market's largest structural buyers on a scheduled date.
Probability matrix
horizon to 30 Sep 2026 · levels in SPYStanding alerts
Bad asymmetry into the 16 September FOMC. The market prices roughly a 33% chance of a hike — three governors already dissented in July asking for tightening — while options positioning sits in the 15th percentile of defence, with total put/call at 0.84 and a 9-day average of 0.83. A binary event meeting an unhedged book.
An oil supply shock in progress. WTI rose 5.2% to $90.22 and Brent to $92.30 after Iran announced a military operation in response to US strikes. Core PCE is already at 3.3% and headline at 3.7%. The 11 September CPI is the first print to absorb this.
Breadth deteriorating for two and a half weeks. New 52-week highs at 1–4% of members against 10–12% in a healthy market, McClellan negative since mid-August, share above the 50-day falling from 70% to 54%. Breadth divergence leads tops by months — treat it as a loss of quality, not an exit trigger.
Semiconductors already turned. SOXX is down −23.6% from its 22 June high and lags SPY by −12.9% over three months while the index went sideways. It is the classic leading indicator of the industrial cycle, and it broke before the index did.
The buyback blackout opens between 8 and 15 September, derived from the 30 September quarter end and the big banks opening the season around 13 October. It removes a demand source on a scheduled date, right on top of the CPI–FOMC–witching cluster.
Credit does not confirm the stress. HY OAS at 2.63% on 31 August, compressing from 3.03% in March and 2.72% in June. Historically lenders turn before shareholders; here they have not turned. It is the strongest reason to read this as a correction rather than a cycle end.
The calendar improves in about 30 days. We are in month 8 of year 2 of the presidential cycle, and October of a midterm year is historically the best entry point of the four. September's vulnerability window and October's opportunity window are neighbours.
Data protocol — round 01
Route taken: mixed. The Alpha Vantage key was confirmed as a free tier, capped at 25 calls a day, and was spent entirely on daily series for 19 tickers. Breadth, credit, sentiment, earnings, the calendar and the quotes for indices, VIX and commodities came from web sources.
What was missing, and is declared rather than filled in:
- IWF/IWD (growth vs value): not measured — Alpha Vantage quota exhausted. Cycle coherence was assessed across 8 of the 9 ratios.
- Copper/gold: directional reading only, from news sources (copper near highs, gold down a third day). No series computed.
- Six-month window: the free series returns 100 sessions, reaching back to 10 April 2026. The "RS 5m" columns cover ~4.7 months, not 6.
- Prices are not dividend-adjusted. That understates the higher-yielding sectors (XLU, XLP) by roughly 1 pp a year over the longer windows.
- Not obtained: DXY, IG OAS, the Fed balance sheet, reverse repo, NAAIM, CFTC COT, dealer gamma, insiders, short interest, implied correlation, the Kitchin cycle low.
Checks run in code: weights sum to 1.00; weighted sum = −2.15; Council score = 50 + (−2.15 × 5) = 39.25; probabilities sum to 100%; third Friday of September 2026 = 18 Sep; second Friday of December 2026 = 11 Dec; midterms = 4 Nov 2026; presidential cycle year and month derived from the 20 January 2025 inauguration; blackout window derived from the 30 September quarter end and the ~13 October season opening.