Insider conviction45%
Only genuine open-market purchases count — a Form 4 with transaction code P, shares acquired, in the last twelve months. Grants, vesting and tax withholding are not buying.
SEC Form 4 · cluster bonus · 30-day recency decay
GemRadar runs a full pass over every sector on NASDAQ and NYSE between $300M and $5B, scores each one on three independent signals, and ranks the day by where they disagree. Nothing is hand-picked.
No card for the free plan.
Everything above is a claim until you can check it. These two pages are where the checking happens — dated, with the denominator, and with the calls that went nowhere left in.
Every sector, both exchanges, in full. Rules define the universe, so nothing is quietly left out.
The day's top three in your inbox, with the date each one was first flagged.
Entry, stop and target are already on the page — derived from rules, not opinion.
Every call keeps being measured. The winners and the losers, with the denominator next to them.
A single strong number is easy to find and usually already priced in. What is harder to find is disagreement — insiders buying while analysts stay silent, a valuation that looks cheap on a stock nobody is watching. Each pillar scores 0–100 on its own data, and only then are they combined.
Only genuine open-market purchases count — a Form 4 with transaction code P, shares acquired, in the last twelve months. Grants, vesting and tax withholding are not buying.
Published ratings against the forward P/E. When two or fewer analysts cover a stock, the page says so rather than presenting a thin average as a consensus.
How easily the stock actually trades: volume against market cap, and whether the price series is trustworthy enough to score at all.
Zero to a hundred, weighted from the three pillars. It says how many signals it had to work with — a stock scored on one pillar is not the same as a stock scored on three, and the page never pretends otherwise.
From the list the sweep produced to the record of what became of it — in the order you would actually meet them.
Everything the sweep found, ranked by Divergence Score. Free sees positions 4 and 5 of the top five; Pro sees all 429, across every sector and both exchanges.
The score, the three pillars that produced it, and how many of them actually had data. The ring sits in the page header and stays there on every tab, so the number you are judging never scrolls away.
Based on 3/3 pillars. Still a first, unvalidated weighting — not investment advice.
Entry, stop and target derived from rules rather than a hunch. The stop is ATR-based and you choose how much room it gets — the risk/reward is then a consequence of that choice, not a claim.
Open a simulated position at the close and keep an honest record of what your decisions would have done — corrected for splits and dividends, so a 3-for-1 does not read as a collapse.
Grouped by sweep day, so any call can be checked against the date it was actually flagged. A hit is +15% or more; the total sits next to it, and Pro sees the flat ones and the losers too.
Measured from the day each gem was first flagged to its most recent close — including after it drops out of the sweep. Four sweep days is not yet a track record — the denominator says so.
“The weights are a defensible argument, not a measured result.”
Every screener you have ever seen leads with its wins. This page is where GemRadar tells you where it is weak, because a tool you cannot calibrate is a tool you cannot use.
45/30/25 is a reasoned argument about what matters, not a result measured against outcomes. It has not been backtested into those numbers.
The record starts the day the sweep started, and it is still only a handful of sweep days long. Whatever the hit count happens to be, the denominator sits next to it — on every page, every time.
When two or fewer analysts cover a stock, the page says so instead of dressing a thin average up as a consensus.
A price in the wrong currency once produced a perfect 100. Rows that fail the scale checks are dropped rather than scored — Unknown, not a guess.
GemRadar surfaces candidates to look at. It does not tell you what to buy, and it never will.
Pro sees all of them, including the three at the top — on the day they are flagged, not after they have moved.
Follow the sweep and the proof, with the top three covered.
No card required.
Create an accountThe whole sweep, the reports, and the tools to act on them.
Cancel any time from Settings.
Start with ProFor people who bill clients: the sweep as data, on your own universe.
Built on request.
Start free firstPrices exclude any sales tax that applies where you are. Not investment advice — the Divergence Score is a first, unvalidated weighting of three signals, and GemRadar surfaces candidates to look at rather than telling you what to buy.
Free account, no card, and the top five in your inbox before the bell.
Start free