When regression to the mean isn’t your friend

Bill Connelly, in looking at coaching underachievers and overachievers, notes this disquieting (if you’re a Georgia fan, anyway) information:

For preview purposes, I note when teams strayed pretty far from their win expectations, one way or the other, and in 2016 two teams strayed really far from expectations: Idaho overachieved by 2.3 wins, and Notre Dame underachieved by 3.2 wins, the fourth-highest (lowest?) in 12 years.

From year to year, this is a sign of randomness. The teams on the extreme ends are all but guaranteed to regress (or progress) toward the mean the next year. Notre Dame was the fourth team to underachieve its second-order win total by at least three games; the other Power 5 team on that list: 2013 TCU, which improved from 4-8 to 12-1 the next season.

12-1?  Dayum, Bill, I really didn’t need to see that.



Filed under Notre Dame's Faint Echoes, Stats Geek!

2 responses to “When regression to the mean isn’t your friend

  1. Hogbody Spradlin

    Breathe easy Blutarsky. Small sample size. Just an anecdote.


  2. KornDawg

    I don’t mind if ND goes 12-1, as long as we’re the 1.