Every trade that executives, members of Congress, and billion-dollar funds are required by law to make public — sorted by the one thing that decides whether it is worth anything: how old it is by the time you can read it.
All three sources tell you what someone bought. They differ enormously in how long they're allowed to wait before telling you. That delay is the whole story — by the time slow filings arrive, the trade is ancient history.
Bar length = days between the trade happening and you being able to read about it.
The day the person actually bought or sold. This is their decision date.
The day it became public and you could have known. Nothing before this date was knowable to you.
The gap between them is the delay. On this page, FILED is the date that matters — it is the first moment the information existed for anyone outside the company.
When an executive or board member trades their own company's stock, they must file within two business days. This is the only source here that's fast enough to act on, and the only one with real research behind it.
One row = one company. It appears here only when two or more different insiders bought shares on the open market with their own money in the same week. One person buying could be anything; four people buying is a conversation.
Dollar value of open-market trades. Selling almost always dwarfs buying — that's normal, executives are paid in stock and have to diversify.
A CEO buying carries more information than a board member buying. They know more.
A 10b5-1 plan is scheduled months ahead — "sell 5,000 shares every quarter." It says nothing about what the seller thinks today. Only the other slice is a live decision.
One row = one person selling one stock on one day. Only sales that were not on a pre-set schedule, so each was a decision made that week. "% out" is how much of everything they held in that company went out the door.
The STOCK Act requires representatives to disclose trades, but allows 45 days and only requires a dollar range, not an amount. Many file at the deadline. The penalty for filing late is $200.
One bar = one stock. Bar length is how many separate trades members of Congress made in it. Stocks touched by more than one member sort to the top.
Days between making the trade and disclosing it. Anything past 45 days is legally late.
Most disclosed volume isn't stock-picking at all — it's bonds, CDs, and funds, often in accounts an advisor runs without the member's input.
By disclosed dollar range. Ranges are wide because the law only asks for a bracket.
Any manager over $100M files a 13F — but only once a quarter, 45 days after the quarter ends. It lists only long US stock positions. No short bets, no cash, no bonds, no hedges. A fund can buy and completely exit inside a quarter and you will never see it happened.
The same 13F filings, for a wider set of well-known investors. These are summaries rather than full quarter-over-quarter diffs — enough to see size, concentration, and what each one's biggest bet is.
One row = one investment manager. "Top 10" is how much of the portfolio sits in its ten biggest positions — the higher that number, the more the filing expresses a real opinion rather than tracking the market.
JPMorgan, BlackRock, Vanguard and the rest file the same 13F form — but theirs are the least informative filings on this page, and the chart below shows why.
Each bar is one filer's top-10 concentration — how much of the portfolio sits in its ten biggest positions. A concentrated manager is making a bet you can read. A bank holding thousands of positions is holding the market, mostly on behalf of clients: index funds, pensions, and trading inventory. Its "top holdings" are just the biggest companies in America, in order.
Bubble area = portfolio size. Left = spread across everything. Right = a real opinion.
Same data as a table. Note how position count and concentration move in opposite directions.
A 13F can never show you this. Form 13F covers long US stock positions only — shorts are excluded by law, so no 13F tracker anywhere has them. The UK is the exception: any net short reaching 0.5% of a company must be disclosed publicly, by name. That is what this section reads.
One row = one fund shorting one company. The percentage is how much of the company's total shares that fund has sold short.
One row = one company. How many separate funds are betting against it, and how much of the company is sold short in total. Crowded shorts can squeeze violently.
These managers appear in the long 13F holdings above and in the UK short register. It is the closest this data gets to seeing both halves of what a fund is actually doing — though the long side is US and the short side is UK, so they are not two halves of one book.
Anyone who is an officer, director, or 10%+ owner of a listed company has their own SEC file. That makes individuals followable — including the Trump family, the most-asked-about case. What the record actually shows is below.
SEC forms only cover holdings in public companies where the person is an insider. A sitting President's broader finances are disclosed annually to the Office of Government Ethics on Form 278e — a separate system, filed once a year, in wide value ranges, as a PDF. It is public but it is not machine-readable the way EDGAR is, so this tool does not pretend to track it. Search oge.gov for the current filing.
When something material happens, a company must file an 8-K — usually within four business days. This is where financial journalists get the story. Each filing declares a numbered item saying what kind of event it was; those codes are translated to plain English here.
The events that mean something went materially wrong: restatements, bankruptcy, auditor changes, accelerated debt, cyber incidents.
Breakdown of every notable 8-K in the window.
The cross-reference worth having: a company whose insiders were buying, or that Congress was trading, and which also announced something material. Empty is a normal result — most weeks these sets don't overlap.
A page of signals is worthless if nobody checked whether the signals predict anything. So we checked, and the honest answer is published here rather than buried.
Every signal entered at the next session's open after the filing date — never the trade date, which would be using information nobody had yet. Take-profit 10%, stop-loss 6%, 40-day maximum hold, costs charged both ways.
The published research finds insider buying works in diversified baskets held for months, not in stop-loss trades over two weeks. This result agrees with that literature rather than contradicting it.