How to compare 13F filings without inventing a trade
A quarterly holdings table is a snapshot, not a trade blotter. This revision withdraws the prior claims about Berkshire's $17 billion Alphabet purchases, its ranking among holdings, and the motives of named fund managers. The article did not provide filing accession numbers and share-count reconciliations sufficient to support those statements. The method below shows what a defensible comparison requires.
What the filing covers
The SEC's Form 13F FAQ explains the reporting rules. Reports generally arrive up to 45 days after quarter end. They cover reportable securities, not every economic exposure, and do not report short stock positions. Option entries and confidential treatment can complicate interpretation. The manager may have changed positions after the reporting date.
Compare quantities, not just dollars
Imagine a filing shows 100 shares at a quarter-end value of $10,000, then 100 shares valued at $12,000 next quarter. The reported value increased by $2,000, but the share count did not change. That is not evidence of $2,000 in purchases. If the later filing shows 120 shares, the net increase is 20 shares after checking for splits and amendments. It still does not reveal exact trade dates, prices, or gross buys and sells inside the quarter.
A reconciliation worksheet
For each comparison record the manager's identifier, both accession numbers, reporting dates, amendment status, security identifier, share class, share count, reported value, and any put/call designation. Read the form's unit instructions rather than assuming every downloaded table uses the same scaling. Keep share classes separate and check corporate actions before treating a change as a transaction.
If estimating a portfolio weight, state the denominator. A percentage of disclosed 13F holdings is not necessarily a percentage of the manager's complete assets. Cash, shorts, nonreportable securities, and other assets can be absent. A large disclosed weight therefore does not fully specify the manager's risk.
Do not infer the investment thesis
The filing does not say whether a purchase expresses optimism, hedges another position, or belongs to a strategy managed by a particular person. Attribute a motive only to a separate, dated statement from the manager. “Smart money is rotating” is not a conclusion a share-count table can establish on its own.
Link back to your own experiment
A simulator replay can describe what holding a named security over a selected period would have done. It cannot reproduce an undisclosed manager's portfolio or demonstrate that copying a delayed report is profitable. The rolling-window study is a useful reminder that choosing dates after seeing an outcome changes the story.