PalanerFree on the App Store
Get
PalanerFinds what's on in Los Angeles
Sign in

Probability and Statistics Seminar: Balancing Posets and the 1/3-2/3 Conjecture
Talk

Probability and Statistics Seminar: Balancing Posets and the 1/3-2/3 Conjecture

Wednesday 4 November 2026, 15:30

Venue time (Los Angeles)

Kaprielian Hall (KAP) · Directions

Wednesday 4 November 2026, 15:30Venue time (Los Angeles)
Kaprielian Hall (KAP)Directions
Price on the ticket pageTickets sold by calendar.usc.edu
Bookvia calendar.usc.edu
Add to calendar

Downloads an .ics file · Times are in America/Los_Angeles · Google Calendar

Hey, how's it going? Probability and Statistics Seminar looks like it should be pretty interesting. Want me to help you organise around it?

Event details

Tickets sold by calendar.usc.edu

Price on the ticket page

Add to calendar · Downloads an .ics file · Times are in America/Los_Angeles · Google Calendar

Max Aires, USC

Title: Balancing Posets and the 1/3-2/3 Conjecture

Abstract: A linear extension of P is a total ordering compatible with the poset relations. Let p(x<y) be the probability that x precedes y in a uniformly random linear extension, and let δ(x,y)=min(p(x<y),p(y<x)) and δ(P) be the maximum value of δ(x,y) over all x,y in P. The famous "1/3-2/3" Conjecture, first made over 50 years ago, posits that δ(P)≥1/3 for all posets with at least 2 linear extensions. For three decades, the best progress on this conjecture was a lower bound of (5-√5)/10, which is tight in the case of infinite posets. We discuss a recent improvement beyond this barrier by an (extremely tiny) positive absolute constant. We also discuss the successful resolution of the related "Kahn-Saks Conjecture" which concerns δ(P) in posets of large width. Both these results combine essentially all existing techniques for finding balanced pairs in posets, and were found by a mix of human labor and AI.

Ask Palaner

Going to Probability and Statistics Seminar? Ask me anything about it.

I read the organiser's pages and answer in a few seconds.

Answers are AI-generated · Privacy

Answers are AI-generated · Privacy