Peter Attia· MD
the way that I've heard is that you want a bracket to contain at 95% of the time which means after you've done all 20 you should have exactly one incorrect and 19 correct
Loading…
The headline is broadly defensible, but the qualifications matter. Effect sizes vary by population, the strongest claims rest on shorter trials, and credible voices push back on how it's typically framed.
the way that I've heard is that you want a bracket to contain at 95% of the time which means after you've done all 20 you should have exactly one incorrect and 19 correct
Every Sunday: the week’s new conflicts and verdict changes — and nothing else.
Native comments, Twitter mentions, and Reddit threads about this claim — surfaced together so the conversation isn't fragmented across platforms.
Bookmarking — the dossier-vs-overview split is the right call. Most of the time I want overview; sometimes I want receipts.
Would love a "what would change this verdict" RSS feed. Sign me up if it exists.