13D vs 13F: Timing, Thresholds, and What AI Can Infer from Each
Both filings tell you who owns what. But they answer different questions on different clocks — and conflating them is one of the most common mistakes in ownership analysis, human or AI. Here's the clean mental model, and what each filing type lets an AI assistant actually infer.
The two clocks
Schedule 13D is event-driven. Cross 5% of a company's shares with activist intent and you must file within a matter of days — disclosing who you are, how you accumulated, and, in Item 4, your purpose: board seats, a sale process, a proxy fight. (Passive crossers can file the lighter 13G instead.) The clock starts at the event.
Form 13F is calendar-driven. Institutional managers over $100M in 13F securities disclose their long positions quarterly, with up to 45 days of lag. It's a portfolio snapshot, filed on a schedule, with no intent attached. The clock is the quarter.
The practical consequence: a 13D can tell you something is happening now; a 13F tells you what a manager's book looked like up to 45 days ago. One is a flare; the other is a ledger.
What each filing lets you infer
From 13D/13G: intent (stated in the filing itself), urgency (accumulation trail from first crossing to latest amendment), and posture — the language of Item 4 reads very differently in a friendly letter versus a hostile one. This is the signal quarterly data structurally cannot capture: by the time an activist position shows up in a 13F, the campaign is old news.
From 13F: breadth and persistence. Single filings are noisy, but stacked quarter over quarter they yield the durable signals — position trails (build / hold / trim arcs), consensus moves (multiple notable funds entering the same name in the same quarter), quiet accumulation patterns, and manager-level track records. The value isn't the snapshot; it's the panel.
Where analysis goes wrong
- Treating 13F positions as current. They can be 45 days stale on filing day — and the manager may have exited since. Share-based quarter-over-quarter changes are meaningful; "current holdings" claims are not.
- Reading value changes as trades. A position's dollar value can double on price alone. Real flow analysis uses share counts, not market values.
- Missing the 13D/13F join. The richest reads come from crossing the two: an activist files a 13D, and the following quarters of 13F data show whether other notable funds followed them in.
Asking in plain English
With the Kaleidoscope SEC server connected, both clocks are one question away — and every answer is cited to the filing:
- "Show active hostile activist campaigns demanding board seats" — event-driven 13D data with stated intent, accumulation trail, and classified tone.
- "Which stocks saw three or more notable funds enter last quarter?" — 13F cluster-entry screening across hundreds of notable managers.
- "Reconstruct how this fund built and trimmed its position, quarter by quarter" — the multi-year, share-based position trail.
- "This company just got a 13D — which notable funds were already holders?" — the join, in one sentence.
The refresh cadence matters as much as the coverage: 13F data updates every ~4 hours from EDGAR, and 13D/13G campaigns land as they're filed — so the flare and the ledger are both current to the source. (For how event-driven desks put this together, see MCP for hedge funds.)