What a 13F filing does not show you
One manager's entire disclosed 13F book is three names. Another's ten largest positions do not even add up to his whole book. Both facts come from the same kind of filing, and both are only half the picture.
Michael Burry's most recently disclosed 13F long book, on Tenachine's Michael Burry page, holds exactly three stocks. LULU makes up 38.7 percent of it, MOH2 makes up 36.5 percent, and SLM makes up the remaining 24.8 percent. Three numbers, adding to the whole disclosed book.
Warren Buffett's page reads nothing like that. His ten largest disclosed positions add up to only 87.5 percent of the reported book. Apple alone is 28.4 percent, Chevron 16.8 percent, BAC2 16.2 percent, and OXY2 9.1 percent, four names covering 70.4 percent, with the rest spread across six more inside the top ten and a further 12.5 percent sitting outside it. Two managers, the same kind of filing, and two completely different pictures of conviction.
What a 13F actually requires
A Form 13F is a quarterly SEC filing required of any institutional manager with over 100 million dollars in qualifying US equity assets. It has to be filed within 45 days of the calendar quarter ending, and it discloses long positions in US-listed stocks, certain options, and convertible securities, along with share counts and market values. That is the entire scope of what the form legally requires. Everything else about a manager's book is optional information the form does not ask for.
What it leaves out
A 13F does not show short positions, most derivatives, foreign-listed shares, fixed income, cash, or anything held outside a public equity account. Burry is a case study in why that gap matters: he built his public reputation on short and put positions against subprime mortgage bonds and, more recently, specific sectors, none of which a 13F requires him to report. The three-stock long book above is real, but it is not an account of what Scion is actually positioned for. It is an account of the one slice of that positioning the SEC requires him to disclose.
Buffett's filing has the same kind of gap, shaped differently. Berkshire Hathaway owns BNSF Railway, Geico, and See's Candies outright, none of which show up in a 13F because they are not public equity positions traded on an exchange, they are operating subsidiaries. The 87.5 percent figure above describes only the portion of Berkshire that looks like a conventional stock portfolio, not the company as a whole.
The lag nobody puts on the chart
The 45 day filing window means a 13F is never current. A position reported for the quarter ending March 31 does not have to be filed until May 15, and it describes a snapshot from six weeks earlier that could already be stale. If a manager builds a position in April and exits it in early May, before the March 31 filing is even due, that trade never appears in any 13F at all. What you are reading is always a photograph of where a book stood on one day, developed and delivered weeks later.
The average and the typical trade, again
Tenachine's per-manager pages report two return figures side by side for exactly this reason. Burry's page shows 170 trades analysed, a 64.1 percent win rate, a weighted return of positive 4.5 percent, and a median IRR of positive 18.3 percent. Buffett's page shows 140 trades, a 68.6 percent win rate, a weighted return of positive 98.8 percent, and a median IRR of only positive 4.7 percent.
Those two managers pull apart in opposite directions. Buffett's weighted return towers over his median IRR, which means a small number of very large winners, Apple foremost among them, are doing most of the work in the aggregate number while a typical disclosed Berkshire trade returned far less. Burry's numbers run the other way: his median IRR beats his weighted return, which means his typical trade did better than the size-weighted total, and some larger, weaker performers are pulling the aggregate down. Tenachine's guide on reading backtest distributions makes the same point about a strategy's average return. The lesson holds across a completely different kind of data: an aggregate number and a typical outcome can disagree, and only one of them describes what usually happened.
How to actually read a 13F page
- Check how many positions are disclosed and how concentrated they are. Three names and ten names mean different things about conviction.
- Ask what kind of manager you are looking at. A hedge fund known for shorts and options, like Scion, is showing you a fraction of its real positioning. A holding company like Berkshire is showing you the public-market slice of a much larger business.
- Note the filing date against today's date. The gap is never less than 45 days and is often closer to three months.
- Read the weighted return and the median figure together, not just the headline one. A large gap between them means a handful of positions are carrying the average.
- Treat the whole page as a description of one manager's past disclosed activity, not a signal to act on today.
Why Tenachine publishes it this way
Tenachine's screener, per-manager analyzer, and rank page all run on the same public 13F data every other tool in this space uses. What differs is what gets put next to the headline number. Every manager page carries both a weighted return and a median figure, the position concentration among the manager's largest disclosed holdings, and a plain statement of what that particular filing type does and does not require its manager to report. A three-stock book and a ten-stock book that only covers 87.5 percent of the total are both real numbers worth knowing. Neither one, on its own, tells you what to do with them.
None of this is investment advice, and a 13F was never built to give real time signals. It is a quarterly, backward looking, partial disclosure, and the gaps in it are not a flaw to work around, they are the shape of the form itself.