HYPE TESTEDReel #1 · Copy what billionaires buy
Copying what billionaires buy is a waste of time.
No edgeThe claim
"Copy what billionaires buy." Famous investors have to publish their US stock holdings every quarter (a 13F filing). Buy what they newly bought and you'll ride their skill.
How we tested it
Every 13F filing since 2016 from seven funds. A "new position" is a stock in this quarter's filing that wasn't in the last one. We buy at the close of the first trading day after the filing becomes public, hold 63 trading days (about 3 months), pay 0.10% per round trip, and compare with the S&P 500 (SPY) over exactly the same days.
1601 new positions found; 1163 tested. 375 had no ticker we could match, 10 had no price on the buy day, and 53 are less than 3 months old. We judge each filing's basket of new positions together (164 baskets), because positions from the same filing aren't independent.
By fund
| Fund | Baskets | Beat S&P 500 | Avg difference | Verdict |
|---|---|---|---|---|
| Berkshire (Buffett) | 28 | 53.6% | -0.17% ±3% | Too few |
| Appaloosa (Tepper) | 36 | 47.2% | +0.55% ±3.38% | No edge |
| Duquesne (Druckenmiller) | 42 | 52.4% | +0.24% ±2.59% | No edge |
| Pershing Square (Ackman) | 19 | 47.4% | -1.26% ±5.28% | Too few |
| Scion (Burry) | 29 | 41.4% | -0.73% ±4.11% | Too few |
| Nvidia 13F (Huang) | 6 | 50% | +1.40% ±42.88% | Too few |
| Situational Awareness (Aschenbrenner) | 4 | 75% | +30.04% ±23.06% | Too few |
A fund with fewer than 30 baskets is marked "too few": a handful of good quarters can easily be luck.
What this doesn't tell you
- This result is probably a bit too kind. Stocks that went bankrupt or were delisted often have no price history and drop out of the test, which flatters the strategy.
- Past results don't predict future results. Taxes are ignored.
- This tests a strategy, not the people. It says nothing about how good these investors are at running their own funds.
Data: SEC EDGAR 13F filings (public domain); stock prices from yfinance (Yahoo Finance), auto-adjusted closes. Summary figures only; per-trade data isn't republished because of the price source's terms.