Your trading desk, prepared before the open.

01 · Gaps

07:00 ET. The overnight gaps, read first.

Before Morning Picks runs, the Gap Analyzer takes the 12 biggest large-cap moves of at least 0.75% and reads them together in one high-effort LLM call. Its default is to fade the move (long a gap down, short a gap up), and the model vetoes that where the news explains it. Calls at 60+ confidence, up to 10, are carried into the run.

The 12 biggest gaps, one LLM read. Calls at 60+ confidence are carried in.

• Panic: fade itLong / Short · mean reversion
• Justified by the newsAvoid
• UnclearWait
• Carried into the run≥ 60 confidence, up to 10

02 · Score

07:03 ET. 300 stocks, scored.

Then Morning Picks runs on its own. It scores the S&P 500 Top 300 on 25 factors across 7 categories, percentile-ranked into one score from 0 to 100; a carried-in mover outside the list is scored in beside them. Swing mode shown; day-trade mode runs its own momentum preset the same morning.

Weekdays before the open: 300 names, 25 factors, one score from 0 to 100.

• Value
• Quality
• Growth
• Financial Health
• Momentum
• Earnings Quality
• Analyst Sentiment

03 · Flows

Where money moved, sector by sector.

Each of the 11 sector ETFs gets a flow score from 5 legs (5-day return, 10-day return, today's move, volume pulse, breadth) and past ±15 it reads as inflow or outflow. An LLM turns each session's read into sectors to favour and to avoid, and a sector leans only when at least 2 of the last 3 to 4 sessions agree. Today's read runs after the picks.

11 sectors, 5 flow legs each. A sector leans only when the last sessions agree.

• Inflowflow score ≥ +15
• Outflowflow score ≤ −15
• Leans on agreement≥ 2 of the last 3 to 4 sessions
• One session is not a trendXLF +17 · 1 of 4 sessions

04 · Tilt

Tilted toward where money is moving.

The list is re-ranked by that vote: the tilt only reorders it (the score from 0 to 100 is never rewritten) and caps stop one hot sector from taking over. 70 survive.

Re-ranked by the vote, capped per sector. 70 survive.

• Sector with inflow×1.8
• …already RSI 70+×1.2
• Sector with outflow×0.6
• At most 40% from one sector28 of 70
• Outflow sectors keep their best2

05 · Read

Then every candidate gets read.

All 70 are analysed in parallel. A trade validator walks price, technicals and fundamentals to a confluence score and asks an LLM for a verdict and a confidence; another model summarises the news; each name gets its own fear/greed reading.

Validator verdict, news summary and fear/greed, for all 70 at once.

• Trade validatorBUY · HOLD · SELL + confidence
• NewsLLM headline summary
• Fear / greedper stock, 0 to 100
• Tapequotes · after-hours · ATR · targets

06 · Select

An LLM keeps up to 20.

A language model reads all 70 rows against the morning's regime and keeps up to 20 setups, ranked, favouring confluence: a strong score, a validator BUY, supportive news. It names its 15 closest misses with a reason each, so you can check what it passed on.

Up to 20 kept against today's regime; 15 closest misses explained.

• Reads the regimeSPY · QQQ · DIA · VIX · rotation
• Keeps up to 20, ranked≤ 7 per sector
• Passed on SYK“sector already has 7 picks”
• Passed on DGX“sector already has 7 picks”

07 · Vision

A vision model reads each finalist's chart.

Each finalist's chart is rendered twice. A classical-pattern analyst reads the line chart, a candlestick analyst reads the candles, and both reads feed the write-up, the same chart tool the Trade Validator runs.

Line chart and candles, read by a vision model, for each of the 20.

• Line chart: classical analystbull flag
• Candles: candlestick analystbullish engulfing
• Pattern readhanded to the synthesis

08 · Review

Checked as a book, not one name at a time.

Three passes run over the 20 finalists. A health check marks each one healthy, warning or critical. An opportunity ranker blends 4 sub-scores and adds a one-line LLM read on each. A portfolio-risk pass treats them as an equal-dollar book (correlation, beta, diversification, sector mix) and an LLM narrates it for the write-up.

Health, an opportunity rank and the book's risk, over all 20.

• Health checkhealthy · warning · critical
• Opportunity ranker4 sub-scores + an LLM line
• Portfolio riskcorrelation · beta · diversification
• Risk narrativeread by the synthesis

09 · Plan

Written up as trades.

A synthesis model, at high reasoning effort, reads the regime, the chart reads and the risk narrative, and writes up trade plans.