Setup, not story.
Conditions, not calls.
We do not tell you what to buy. We tell you where the conditions are present — and where they aren't. Read like a reference book, not a feed. Investing is research; the system is the library.
Five words do most of the work.
Each signal state is a defined term. We use them the same way every time, and we mean nothing more than what's written. No green lights, no buy signals, no hand-waving.
The setup we look for is in place. Worth a closer look, but not a green light.
NVDA matched on 1 May 2026 after Q1 print and four consecutive quarters of operating-margin expansion.
Setup is forming. We're watching for confirmation. May turn into Matched, may not.
CRWD has been stalking for 14 days. One more clean quarter and the window opens.
Setup is being reassessed after new data. Stand by, direction unclear.
RDDT moved to reviewing after Q4. The cohort line shifted; the window is being recut.
Last window has finished. Waiting for the next setup to form.
SHOP closed at $612 on 18 Dec 2025, a 13.4% advance from base. No active window today.
No setup, no movement worth watching. Skip it.
4,758 of the 4,820 listed equities are in Ignore right now. That's the point.
Built to surface where conditions are changing, not to tell you what to buy.
Three primary states, one per stock
Every name on the list reads as Setup detected, Worth watching, or Nothing here yet. The state is what fundamentals and valuation are doing, together, against the company's own history. No score, no rating, no ranking.
Conditions: opmargin↑, FCF↑, P/S compressed vs trailing 24m.
Years of cohort data behind each signal
Every state is grounded in thousands of historical windows with similar conditions. You see the outcome distribution, not a forecast. Past patterns do not guarantee future results. We say it because we mean it.
n = 218 windows · matched 67% · median +11.4%
Historical price levels behind every Matched
Helmbeam Pro shows the specific levels: where the window opened, where pullbacks confirmed, where current price sits relative to all of them. Levels derived from historical institutional activity, surfaced for your own research.
Current $948.21 · +15.3% from base
Sentinel alerts, only when something changes
Add any stock to your watchlist. We notify you the moment the state changes, and only then. No daily noise, no marketing emails, no streak counters. Watching the market is a feature; we just do it for you.
18 May, 14:32 GMT · view in app
Discover feed, editorial not algorithmic
Browse stocks where a signal has fired, filtered by quality and sorted by recency. Each entry has a short note from the system on what fundamentals and valuation are doing. Starting points for your research.
"Op margin inflected; valuation 18m low."
Full breakdown when you want it
Snapshot · Numbers · Timeline. Three tabs per ticker. Posture, fundamentals, and the chronology of every event that moved the system's read. Nothing hidden, nothing automated, nothing recommended.
18 Apr · 8-K · TSMC capacity extended
2 May · Pullback gate confirmed
The anatomy of a Matched window.
When a stock enters Matched, four objects move together. We surface all four. You decide what they mean.
Current price
Where the market is today, and how far it sits above every gate. Real-time-ish, on a 5-minute delay.
Base gate · amber
The price level where the window opened. Claimed the day fundamentals first inflected against the company's own history.
Pullback gates · sage
Levels where prior retracements stopped and reversed. Floor candidates for the active window, not predictions.
Cohort line
Where this setup sits historically. The system has seen 218 similar windows; 67% reached Matched, median outcome +11.4%.
What Helmbeam is. What it isn't.
We'd rather be useful for a narrow thing than vague about a wide one. Here is the line, drawn.
What it is
What it isn't
See which stocks have a detected setup. iOS and Android.
Search any stock, check its signal, build a focused watchlist. Upgrade to Helmbeam Pro for the price-level breakdown, the Discover feed, and the full historical cohort.
