The week turned on prices, not commentary
The cleanest market story this week was not a speech or a platform release. It was the way inflation data moved the event book into the September FOMC.
This post is educational. It is not financial advice and it does not promise returns.
Inflation rebuilt the event window
On 10 September, BLS reported that August final-demand PPI rose 0.4% after a 0.1% July increase and a 0.1% June decline. The goods side mattered: final-demand goods rose 1.1%, with energy doing much of the work.
One day later, August CPI also printed 0.4% month over month. Gasoline rose 3.9%, energy rose 2.1%, and core CPI rose 0.3% after a 0.2% July increase.
For discretionary traders, that is a macro debate. For automated traders, it is an input shock. A system that was sized for a slow policy week suddenly had to trade around a live rate decision with inflation and energy back in the driver seat.
The important lesson is that the calendar did not change. The risk attached to the calendar changed.
Rate probabilities became execution risk
Cboe's 14 September macro-volatility note put the mechanism plainly: higher oil prices and hotter core inflation helped push the market-implied odds of a Fed hike at this week's FOMC from 58% to 90% in the OIS market. Weekly SPX options were pricing roughly a 1.1% implied move for Wednesday's Fed announcement.
That is the part automation should respect. Policy probability is not a research footnote. It is a traded price, and it can be pulled from futures or swaps into volatility, skew, index futures, FX, gold, and broker margin behavior.
CME's FedWatch API page makes the same plumbing point from the data side: FedWatch is derived from 30-day Fed Funds futures, supports real-time data, and is positioned for algorithmic execution, duration hedging, macro strategy, and portfolio monitoring.
In other words, the "Fed odds" headline is already part of the machine-readable market. If your model consumes it only after a news rewrite, you are downstream of the repricing.
Vol buyers were not waiting for the press conference
The volatility surface started adjusting before the decision. Cboe noted that the VIX Index gained 1.3 points last week to 15.8, with more than half of the move coming from steeper SPX skew and convexity. It also flagged heavier tail hedging through VIX options, including several of the largest VIX trades of the year in the prior two weeks.
That matters because many automated strategies read realized volatility, spread, and recent slippage as if they are current state. Around an event, those variables can be stale precisely when the book is most sensitive. The options market may already be charging for a distribution your execution layer still treats as normal.
The practical response is not "turn everything off." It is to decide, in advance, which systems are allowed to quote, add, copy, or rebalance while implied move and skew are doing the warning.
The 24-hour market question is the same control problem
The SEC's 17 September roundtable on 24-hour trading is formally about U.S. equity-market structure, but the agenda reads like an automation checklist: overnight surveillance, expected liquidity, operational resiliency, Regulation SCI, failover, market-data continuity, shorter maintenance windows, cybersecurity, and staffing models.
That belongs in the same conversation as FOMC week. More hours do not create more liquidity by themselves. They create more states in which liquidity can be thin, staffing can be partial, market data can be degraded, and an automated strategy can still be live.
If a system is allowed to trade during a thin overnight session, the control plane has to know that the session is thin. If it can copy trades across accounts, the downstream accounts need the same event-state constraints as the source. If it can recover after a disconnect, the reconnect path needs to know whether the book it is rejoining is the same book it left.
Durable lesson
Event risk is not only a directional view. It is a permissions problem.
Before the next CPI, FOMC, or 24-hour-session rollout, the useful questions are mechanical:
- Which strategies may open new risk inside the event window?
- Which may only reduce risk?
- Which symbols or accounts get wider spread and slippage limits?
- Which feeds are trusted enough for automated repricing?
- Which kill switch is independent of the model that wants to keep trading?
The worst automation failures often start with a reasonable signal running in an unreasonable market state. This week was a good reminder that policy, inflation, volatility, and session design are not separate topics for an operator. They are different names for the same live control surface.
Soft next step
If you are reviewing copy-path sizing after a macro-volatility reset, the free copy ratio and lot size calculators can help with the arithmetic. They do not replace broker specs, exchange rules, or a kill switch.