Why news matters for short-horizon options trades
Charts are downstream of news
Every meaningful intraday move in an Indian F&O stock has a cause, and most causes leave a trail in the news flow before they leave a trail on the chart:
- A corporate filing on the exchange bulletin board that traders haven't priced yet
- A regulatory headline overnight that re-rates an entire sector at the open
- A guidance change buried in an investor-presentation deck released the prior evening
- A commodity-price shock in another time zone that hasn't propagated to the local sector yet
Charts catch up. By the time the price chart "tells you" something, the wire has usually known for a while. The hypothesis behind Project H's news layer is that some of that lead is capturable — that the market digests news imperfectly over short horizons, and a system that reads broadly and weights carefully can lean into the digestion window.
That's a hypothesis. It sounds convincing — most trading theses do. Whether it survives contact with real forward returns is precisely what the paper scoreboard exists to measure, and the honest answer is "the jury stays out until the evidence accrues."
What the news layer actually has to do
Three jobs, in principle terms:
1. Aggregate broadly, not deeply
A single news source is a single failure mode. The system pulls from a wide mix — exchange filings, wire services, broker actions, sentiment streams, macro calendars — and treats redundancy as a feature. When independent sources agree, the signal strengthens. When they disagree, abstaining beats guessing.
2. Score for materiality, not just sentiment
Not every headline matters equally. "Company reports earnings" deserves different weight from "company acquires a small startup," even when both read as positive. Sentiment-only models miss this. Materiality scoring asks a harder question: would this news, seen before the market fully priced it, have changed your position? Modern language models are unusually good at that classification — which is what makes a serious news layer feasible on a single-developer budget at all.
One more filter matters: what the market has already priced. A headline arriving after options have re-priced for the move it describes is late news wearing early news' clothes. The system discounts news accordingly — a mechanism you can watch working on the live dashboard, where each headline is shown alongside how much weight it was actually given.
3. Cross-validate against price — continuously
Every news source eventually has to be tested against realised forward returns. A source that scores headlines beautifully but doesn't predict anything is decoration. The audit runs continuously; sources that don't earn their place get cut, and no source is grandfathered in because it feels informative.
Why options make timing the whole game
Options add two constraints that stocks don't have:
- Theta decay — premium leaks every minute a position is held. You can't be "right but slow."
- Convexity — when you're right with timing, the payoff is non-linear.
Together they mean the news layer isn't trying to answer "is this company good?" It's trying to answer "is there a fresh, material, un-priced catalyst right now?" — and a name with no fresh catalyst gets treated more conservatively than one with a filing an hour old, no matter what the chart says.
The honest caveat, again
It would be easy to end a post like this with "and that's why the news layer prints money." It doesn't get to claim that. News is one input into a multi-signal blend, under continuous measurement like everything else in the system — and the measurement, not the story, decides whether it keeps its seat. The scoreboard is public; watch it decide.