Can you copy a famous investor’s 13F? Start with the clock.

The Buffett Effect, a working paper submitted to arXiv on October 12, 2026 by Pawel Robert Smolinski and Siddharth Saridevi, compares new positions disclosed by 24 recognizable funds with those of 24 matched funds that are not widely recognized, from 2000 to 2025. It reports a positive disclosure premium that rises with recognition. It also reports that portfolios copying the most recognized funds’ newly disclosed purchases, entering on the first trading day after each filing, earned a statistically significant four-factor alpha: a return beyond what the market, company size, value and momentum explain. That alpha was strongest in the first month and faded over the following quarter without reversing.

Primary source: The Buffett Effect · working paper, October 2026 (opens in a new tab).

That is a reason to study timing, not a promise of an executable return. The authors read the evidence as attention and information working together rather than as two separate effects. They also note that the main premium is driven by five managers, three of whom appear in this lab: Berkshire, Scion and Pershing Square. Without Buffett or without Burry, the 30-day premium stays positive but is no longer statistically significant. Our replays neither replicate their study nor estimate the effect of fame. They answer a narrower question: how did the observed price comparison change when the assumed entry moved?

Berkshire and Alphabet: the November 14, 2025 filing arrived on a Friday after the close. The first replay entry is Monday, November 17. A chart that starts the follower at Friday’s close would quietly give that follower a price from before the disclosure. The lab keeps the earlier price movement separate from the follower’s window.

Berkshire and UnitedHealth: the August 14 filing provides another after-close example. Compare the first-session entry with the delayed entries while keeping the exit fixed. Waiting can change the price in either direction. The direction of that change is an observation, not evidence that delay always helps or hurts.

Pershing Square and Uber: the May 15 filing was not the first public news of the stake. Reuters had reported Ackman’s announcement on February 7. This case exposes a different mistake: treating every new row in a quarterly table as a new revelation. Our May replay cannot measure the February announcement reaction.

Berkshire and TSMC: the November 2022 disclosure produces a small loss over the first 20-session window and a gain over 60 sessions. Choosing an endpoint can change the headline without changing the underlying filing.

Scion and Alibaba: Burry’s firm filed during market hours on February 14, 2023. The next-session entry is a consistent convention, not the earliest possible trade. The stock lost value and trailed SPY over both observation windows.

Soros Fund Management and Rivian: the February 2022 filing followed Rivian’s IPO. The replay includes a steep decline. A large stake associated with a famous name did not give the hypothetical follower protection from losses.

The lab offers exits 20 or 60 trading sessions after the first eligible replay entry. For each exit, every entry scenario ends on the same date. A five-session wait therefore produces a shorter holding period, not a fresh 20- or 60-session window. This makes the entry-price comparison explicit.

Stock and SPY price changes use the same entry and exit dates. Their difference is expressed in percentage points. It is not risk-adjusted alpha. The chart starts both series at 100 on the selected entry date, and a daily price table and CSV make the calculation inspectable.

The calculations exclude dividends, fees, slippage and taxes. They do not reconstruct a manager’s portfolio, options exposure or original purchase price. Earnings and other market news can explain part or all of a move. A positive number alone does not establish that the filing caused it.

The SEC lists November 16, 2026 as the deadline for Q3 2026 Form 13F reports. When a new disclosure attracts attention, first establish the portfolio date, the public timestamp, whether the position was already discussed publicly, and which price your comparison assumes.

A useful research habit is to write down those choices before checking the outcome. A genuine strategy test would also need a pre-specified universe, all eligible events rather than memorable examples, complete historical security coverage, and a period kept separate for validation. This release does not make that claim.