How the engine works, from the morning news to a graded call

Most trading tools show you an output and ask you to trust it. This guide walks the whole path instead: how the engine goes from the morning's news to a single graded call on US index futures, every market day, with nothing hidden along the way.

The problem it solves

Before the open, the information that will move the day is scattered across dozens of places: a central bank speech here, an earnings surprise there, overnight futures, the economic calendar, how the market itself has been behaving. A person can read all of it, but not every morning, not consistently, and not without their own mood leaking into the read. The engine exists to do that reading the same way every single day and compress it into two things: a direction and a judgment on whether the day is worth trading.

Stage one: reading the news

Every market morning, pre-open, the engine reads across four broad areas: macro and central banks, earnings and guidance, global risk, and market internals. It is not scraping headlines for keywords. An AI agent reads the substance and asks the only question that matters: does this push US index futures up or down today, and how hard?

Stage two: weighing what matters

Not all news is equal, and how much each kind matters changes over time. Each factor in the read carries a learned weight: a measure of how much that factor has actually mattered to past sessions. When the engine gets a day wrong, the weights that led it there are adjusted. This is the part that makes the system a loop rather than a one-way pipe, and we will come back to it at the close.

The weights are not a black box. The factors behind each morning's call, and how hard each one pulled, are shown on the dashboard next to the call itself.

Stage three: the news gate

Some days are structurally unreadable before they start. When the calendar holds scheduled high-impact news, a Fed decision, a major inflation print, the honest position is that nobody knows which way the day breaks until the number is out. On those days the engine does not manufacture a confident opinion. It flags the day NEWS VETO and says so plainly.

Stage four: one call, before the open

By 9:30 AM ET the pipeline has printed the day's read, in plain English. It has three parts:

• The bias: bullish or bearish, for the day, on US index futures.
• The confidence: how strongly the morning's evidence agrees with itself. High confidence means the inputs lined up. It is a statement about the evidence, not a guarantee about the outcome.
• The conditions score: 0 to 100, answering a separate question: is today the kind of day where moves follow through, or the kind that chops both sides out? DIRECTIONAL at 65 and above, MIXED from 40 to 64, CHOPPY below 40. There is a full guide to reading it in reading trading conditions.

Stage five: the grade at the close

After the session ends, the engine checks its morning call against what the market actually did, and writes the result down where members can see it. The rules are deliberately simple:

• Called bullish and the session closed up, or bearish and it closed down: a hit.
• Called it the wrong way: a miss. Logged exactly like the hits.
• The session closed essentially flat, within about 0.15% of unchanged: neutral. No result is claimed either way, because claiming coin-flip days as wins is how records get quietly inflated.

The log is append-only. Yesterday's grade does not get edited when it ages badly. And every miss feeds stage two: the weights that produced the wrong call are moved, and the same pipeline runs again at the next open, slightly less wrong. Why we built the grading this way has its own guide.

What the engine deliberately does not do

• It does not give entries, exits, stops or targets. It reads the day; the trade is yours. See how to use the bias as confluence.
• It does not re-issue calls intraday to chase the tape. One read, before the open, graded after the close. A call you can move after the fact is not a call.
• It does not pretend markets are open when they are not. Weekends and holidays say so, plainly.
• It does not give financial advice. It describes conditions; it never tells anyone to buy or sell.

Why it is built this way

Every design choice above comes from one principle: a daily read is only worth something if it can be checked. Fixed timing, one call, simple grading rules, a public log, and a veto for the days that cannot honestly be called. Take any of those away and you have a tool that is easier to market and harder to trust. We chose the other one.