Cleverium: Which Executives Really Get It Right When They Buy Their Own Shares

By Kike Gandia · Research · 9 min read

Illustration of loose filings flowing like streaks of light into a circle of four nodes linked by arrows, with a bar-chart ranking inside and the top bar highlighted in orange

Every trade an executive makes in their own company's stock is public. Whenever a CEO, a CFO or a board member buys or sells shares in their company, they have to report it to the SEC within two business days. Europe and Asia have similar rules, with somewhat longer deadlines. The information is open and free, and practically nobody reads it — because “public” is not the same as “usable”.

The forms arrive as XML, spread across thousands of filings a day, with no ranking, no context and without the one question that actually matters: did they get it right?

2 days: the business-day deadline to report each purchase or sale to the SEC
11: pages on Cleverium (arranged in a circle, not a loose menu)
4: big questions (Method, Rank, Signals and Reality-check)
€0: to see what is happening right now (the history and alerts are paid)

Public what the regulator publishes

Forms in XML
Thousands of filings a day
No ranking
No context
Missing the question that matters

Usable what Cleverium does with it

Executives and funds ranked by realized performance
Fund overlaps and clustered purchases
A cross-check against prediction markets
The method explained before the ranking
The answer to did they get it right?

The gap between “available” and “reviewed” is the gap Cleverium fills.

That is why we built Cleverium.

What makes Cleverium different

The tools that already exist for this — Dataroma is the best known — solve half the problem: they aggregate funds' 13F portfolios and present them in order. Very useful for seeing what someone holds. But they don't tell you whether holding it was a good idea.

Cleverium ranks executives and funds by realized performance: what the stock did after they bought it, not just how much they bought.

A data point aggregated 13F portfolios

“Warren Buffett holds this”
What each one holds
How much they bought

An answer realized performance

“This is what happened to every stock Warren Buffett has bought in the last three months”
Whether holding it was a good idea
What the stock did afterward

The first sentence is a data point. The second is an answer.

If you're interested in the underlying question — whether these purchases, taken together, say anything — we measured it in When an executive buys their own shares: 17,746 reported purchases, a control group and a net edge of 2.28 points at 20 trading sessions. That analysis also flags its own limit: within each category, between 40% and 46% of trades underperform the market. The aggregate says the signal exists; the individual case is noisy. That is why, on Cleverium, the method comes before the ranking.

A circle, not a menu

The whole site is organized as a circle: Method explains why to trust Rank, Rank is what you watch through Signals, Signals are tested in Reality-check, and Reality-check is exactly what Method exists to justify. Eleven pages, four big questions.

The Cleverium map: eleven pages in four phases that link up in a circle. Method, why to trust the number: How the score works (cleverium.es/methodology). Rank, who really gets it right: Scoreboard (cleverium.es/scoreboard) and Funds (cleverium.es/funds). Signals, what to watch: Cluster buys (cleverium.es/clusters), Who invests like whom (cleverium.es/patterns), Explore (cleverium.es/explore) and Activity (cleverium.es/activity). Reality-check, testing it against reality: Insiders vs. the crowd (cleverium.es/crowd), Is the crowd right? (cleverium.es/calibration), Smart money (cleverium.es/smart-money) and The Short Book (cleverium.es/shorts). Method justifies Rank, Rank is watched through Signals, Signals are tested in Reality-check, and Reality-check is what Method exists to justify.

Method: why to trust the number

Method How the score works: Before showing a single ranking, the page that explains what the score is, how it is calculated and, above all, what it CANNOT tell you.

A small sample is not a track record. The site says it in so many words before you have to find out the hard way.

Rank: the scoreboard

Rank Scoreboard: Which insiders have the best and worst realized track record after buying. Not “who paid the most”, but “who has actually come out ahead”, rank by rank, with the trade behind every name.
Rank Funds: What the tracked funds hold, where they overlap and who is adding to a position right now. This is where “what each one holds” does live — but as a starting point, not a conclusion.

Signals: what to watch

Signals Cluster buys: Which companies have seen several purchases by different insiders within the same month.
Signals Who invests like whom: Which funds run nearly identical portfolios, and which of the two moves first. If two managers buy the same thing weeks apart, one of them is copying — and it's worth knowing which one to follow.
Signals Explore: Every reported purchase, sliced by date, size, sector and filer. The “give me the raw data and let me filter it myself” view, for when the ranking isn't enough.
Signals Activity: What has been filed in the last few days, and how the prediction market is pricing it right now. The site's most up-to-date pulse.

A single purchase may mean nothing

It may be about liquidity
It may be a bonus

Three or more at once on the open market

Different insiders
Same company, same month
A pattern that can't be faked

Why Cluster buys counts purchases, not just dollar amounts.

Reality-check: testing it against reality

Reality-check Insiders vs. the crowd: Where a company's insiders and the prediction markets (Polymarket, Kalshi) point in different directions. When the people with inside information disagree with the people putting money in from outside, there is something worth looking at.
Reality-check Is the crowd right?: If a contract trades at 70 cents, does it really pay out 70% of the time? The question almost nobody asks about prediction markets, answered with data, not intuition.
Reality-check Smart money: Which open markets Polymarket's big money is in, and at what price it got in.
Reality-check The Short Book: Which funds are betting against a company, and whether the stock actually fell afterward. The same treatment as the Scoreboard, but for those betting on a decline.

An illustrative diagram, with no data, of how to read a calibration curve like the one on Cleverium's Is the crowd right? page. The horizontal axis shows the contract price in cents; the vertical axis, the percentage of times it pays out. The diagonal is perfect calibration: a contract at 70 cents should pay out 70 times out of 100. If it pays out more often, the crowd was being too pessimistic; if less often, it was overpaying for a certainty it didn't have. The real results are at cleverium.es/calibration.

Smart money is not a hit-rate ranking. Polymarket's data comes filtered toward winning wallets, and the site admits it instead of faking a rigged scoreboard. What it shows is where serious capital is positioned right now.

The present is free; the past and the future are paid

Free the present

The eleven pages above
What is happening right now

Paid the past and the future — cheap

The full history
Real-time alerts
A Telegram channel that notifies you as the forms come in

This is not investment advice. It is a transparency tool, not a signals service. That line is written on the site itself, not in the fine print — and it is the same one How the score works draws from its first paragraph.

Why QuantumSec built it

At QuantumSec we spend our days looking at attack surfaces that people think are under control and aren't:

Open ports: That nobody remembers opening.
Reused credentials: The same key for doors that shouldn't share one.
Third parties with too much access: More than they should have.

The pattern repeats outside cybersecurity: there is data that is public by legal mandate and that, in practice, nobody audits, and that gap between “available” and “reviewed” is exactly where problems hide — whether it's a port or an insider trading pattern that nobody has bothered to cross-reference with the share price.

Cleverium applies that same obsession to SEC filings. And if you visit the site, you will see a few of our ads — we don't hide it: that is how we pay for the infrastructure and keep the subscription cheap for everyone.

An SEC Form 4 what Cleverium reviews

Public by legal mandate
Nobody cross-references it with the price
The pattern sits in plain sight, unread

Your perimeter what we review at QuantumSec

Public repositories
Forgotten subdomains
Configurations nobody has looked at again since they were deployed

It's the same question: is it as clean as you think?

If your company has never asked itself whether its own “public information” is as clean as it thinks, let's talk. That is what we do in an external ethical hacking engagement and a vulnerability assessment: look at your perimeter the way an outsider would, before one does.

cleverium.es · quantumsec.es

Topics

#Cleverium #SEC #Open-sources #OSINT #Data-analysis #Prediction-markets #Insider-trading

Keep reading

Review your public footprint