Palantir fell 17% on a deleted tweet last week. I built something to find out if that reaction was even warranted.
Last week a deleted tweet from Michael Burry triggered a 17% correction in Palantir. It raised a question I kept coming back to: why does a single comment move a stock that much, and how often do calls like that actually hold up?
The Burry/PLTR episode is a good example of a recurring pattern. A pundit makes a bold public call, markets react immediately, and by the time anyone checks the underlying thesis the damage is done. Whether PLTR is overvalued is a separate debate, but a tweet about Anthropic eating its lunch is a thin basis for a 17% move in either direction.
I've been digging into this for a while now, tracking public predictions from 10 well-known commentators (Burry, Tom Lee, Cathie Wood, Ackman, and others) against what actually happened. A few things that stood out:
\- Several experts' funds significantly outperformed their own public guidance, suggesting the TV persona and the trading desk operate very differently
\- Macro crash calls (SPX -40%+) have a near-0% hit rate across almost every expert in the past 15 years
\- Single-stock accuracy is materially higher than macro timing accuracy, across the board
\- Strict accuracy ranges from roughly 0% to 60% depending on the expert
One thing that makes the Burry case interesting specifically is that his 13F portfolio has repeatedly contradicted his public commentary. The "Sell" tweet in January 2023 came while he was buying Chinese tech. The public narrative and the actual capital allocation are often two different stories.
I put the full dataset, methodology, and per-expert breakdowns at [alphaaudit.net](http://alphaaudit.net) if anyone wants to dig in. Happy to discuss whether a 12-month scoring window is the right standard, or whether others have a better framework for measuring this kind of thing.
What do you think drove the PLTR reaction last week, and do you think the market is right to move this much on a single deleted tweet?