TradingHQ

What it is

TradingHQ is a personal trading pipeline that runs on every trading day. It scans the market in the pre-market window, classifies the regime, surfaces candidates that match several specific setup patterns, and ships a written report before the open. A second pass runs in the last half-hour of the session and produces the Close report — the day's late-tape story stocks while the market is still live.

The reports you read on this site are the public output of that pipeline. The morning Daily and Sunday Weekly come out of the main run. The Close comes out of the evening scan. Everything visible here was generated by the same system that informs the trades.

What it helps with

Pre-market prep used to mean ninety minutes of clicking through screeners, checking breadth indicators, eyeballing leader charts, and writing down what to watch — every day, before the open, while still half-asleep. The pipeline replaces that with ten minutes of reading a structured report.

The four things it does that matter:

• Regime read in one number. The composite score (CMRF) reduces the broader market's character to a single classification: are we in a regime that rewards pressing, or one that rewards sitting out? That decision is the most consequential one a swing trader makes each day, and it's the one most often skipped under pressure.

• Candidate universe with conviction-ranking. Several independent setup scanners run in parallel. Names that show up in more than one are ranked higher — convergent signals across different methodologies are the strongest read. The morning report shows the top of that list with a grade and reasoning.

• Late-tape story stocks. The Close report surfaces names from the High-Volume Edge universe that came alive in the second half of the session. Each is classified — clean leader (CAT), broken setup (DOG), or volatile two-sided action (LIQUID_LAVA) — so the overnight watchlist is written before the bell, not at midnight.

• Outcome tracking that feeds back into scoring. Every signal the system produces gets tracked at fixed horizons (5 / 10 / 20 sessions). The weights inside the scoring model are updated from the actual hit rates, not from intuition. The system learns what's working in the current regime.

How it works

The pipeline runs in a fixed sequence: data, regime, candidates, synthesis, deliberation, narration, report.

Three filters run before any chart pattern matching: the underlying name has to be in a tradeable structural state — either an established uptrend or a confirmed late-stage bottoming pattern — it has to be a leader within a leading industry group, and the market regime has to be at least neutral. If any of the three fails, the name is out of the playable universe regardless of how the chart looks. These filters eliminate roughly 70% of any given universe and are the single most useful piece of the system.

Inside that filtered universe, several setup scanners run in parallel. Each one looks for a different structural moment — the first reclaim of key moving averages after a downtrend, tight consolidations near a pivot, late-stage continuation breakouts, structural reversals out of late Stage-4 bottoming patterns, and breakout-from-IPO patterns. The convergence between scanners is the signal. A name that's flagged by three scanners is a much stronger read than one flagged by one.

A language model layer handles narration and classification. Once the convergent candidates are ranked, the pipeline uses an LLM to write a short rationale per name and (for the Close report) to classify each story stock into one of three behavior tags. The LLM never makes the entry decision — that stays with the trader during live execution.

The output is multi-channel. An HTML report goes to email. A structured archive goes to this site via an authenticated ingest endpoint. The raw JSON stays in cloud storage for the outcome tracker to pick up later.

What it doesn't do

It stops at the bell. The pipeline's contract is a written pre-market plan — regime, candidates, rationale — delivered before 9:30 ET. Everything from the bell onward is the trader's job at the keyboard: live-volume confirmation, entry timing, stop placement, position management, exits. None of that lives inside TradingHQ.

It doesn't predict price. It identifies names in a known structural state — typically a Contraction phase or a late-stage bottoming pattern — and surfaces them as candidates. Whether the move that follows is real or fails is decided by the live tape, not the screener.

It doesn't override discretion. The trader makes every bet. The pipeline frames the morning by removing names that shouldn't be considered and ranking the ones that should — but it never pushes the button, and it doesn't tell the trader how to manage what happens after the entry fills.

Why it exists

Trading is mostly about not making bad decisions under pressure. The bad decisions tend to happen in two places: choosing what to trade (selection error) and deciding how aggressive to be (sizing error). Both compound when made under time pressure with incomplete information, which is the default state at 9:25 AM ET.

The pipeline removes the time pressure from selection and writes the sizing input (the regime) at a timestamp before the open. The trader inherits a defined universe and a defined regime instead of building both from scratch every morning. The system isn't trying to be smart — it's trying to make the smart decision the easy one.

The reports archive serves a second purpose: a regime read with a timestamp is auditable, and an auditable regime read makes the trader honest about what was actually thought before the bell. That's the discipline the market pays for.