Close Menu
    Daten

    [OC] Historische Wahrscheinlichkeit, eine perfekte NCAA-Klasse auszuwählen, 1985–2025


    [OC] Historische Wahrscheinlichkeit, eine perfekte NCAA-Klasse auszuwählen, 1985–2025

    Von possiblywrong

    View 3 Comments

    3 Kommentare

    1. possiblywrong on

      The figure shows, for each year, the prior probability of the exact outcome of all 63 games of that year’s NCAA men’s basketball tournament, using a couple of different models of game outcomes as a function of seeding. Interestingly, last year’s 2025 tournament was, by both models, the second most likely outcome in the history of the current format (2007 being the most likely).

      Source data is on [GitHub](https://github.com/possibly-wrong/ncaa-tournament), with links to multiple articles over the years describing methodology in more detail.

      I used Mathematica to aggregate the data, compute the probabilities, and create the figure.

    2. prediction_interval on

      It’ll be interesting to see if this trend towards chalk continues this year.

      There’s definitely an argument to be made that, with the NIL rules becoming increasingly influential in the last couple years, there’s been more movement of talent from smaller school to the big programs, and that’s [resulted in a bigger gulf between the haves and have-nots](https://www.theringer.com/2026/03/18/college-basketball/march-madness-2026-ncaa-tournament-upsets-cinderella). If this is true, we’ll continue to see a decline in upsets until the system is reigned in.

    Leave A Reply