Everyone says it: just copy what the great investors buy. Their 13F filings are public, so the strategy sounds free. We rebuilt five years of those filings from SEC data and tested the simplest version of the idea.
The test. We tracked 83 well-known funds. Whenever three or more of them increased their stake in the same stock in the same quarter, that counted as a signal. Then we "bought" on the filing deadline — the first day every filing is actually public — and sold one quarter later. No cherry-picking, no hindsight entry prices. 2,724 signals, 2021 Q3 to 2026 Q1.
The result.
- Win rate: 56.8%
- Average return: +2.74% per quarter
- S&P 500 over the same windows: +3.28%
- Excess return: −0.54 percentage points
Read that again. The copied trades made money almost 57% of the time — and still underperformed simply holding SPY. A high win rate feels like an edge. It isn't one, because the losses, when they come, are bigger than the wins. Sea Limited fell 62% in a quarter after a crowded buy signal. Carvana, the best case, rose 255%. The distribution is a lottery ticket stapled to an index fund — minus a little.
We also tested the more aggressive version: stocks where a fund at least tripled its position (598 signals, positions of $5M+). Win rate 55.5%, average +1.65%, excess −1.55 points. Bigger conviction did not mean better outcomes.
Why we publish this. Most "smart money" content shows you the buys and never the scoreboard. The filings are real; the edge, at least in this naive form, is not. That doesn't make 13F data useless — it makes it a starting point for research rather than a buy list.
Full methodology and the quarter-by-quarter scoreboard are on our Track Record page.
Data: SEC Form 13F filings, 83 funds, 2021Q3–2026Q1. Returns are dividend-adjusted, measured from the filing deadline (quarter-end + 45 days) to the next filing deadline. Past patterns are not a promise of future results. Not investment advice.