Why is P/BB so variable?

This forum is for discussing tournament formats, question styles, strategy, and such.
Post Reply
User avatar
DavidB256
Lulu
Posts: 15
Joined: Sun Nov 04, 2018 7:37 pm

Why is P/BB so variable?

Post by DavidB256 »

Looking at the stats for the 2019 PACE NSC, I noticed that the P/BB (points per bounceback) statistic correlates very poorly with both overall placement and P/B (points per bonus). I understand that P/BB is a function of both the skill of the team and the skill of the teams they play against (playing against better teams would decrease your P/BB because they are more likely to get all of the easier bonus parts), but I don't understand how that can lead to some of the existing results. For example, the Championship bracket had a range of 4.49 P/BB, the 9-16 bracket had a range of 4.74 P/BB, and my team, which got 59th, had more P/BB than DCC A.

Does anyone know why this is?
David Bass
Jamestown High School '19
University of Virginia '23
"General P. P. Peckem was, as he always remarked when he was about to criticize the work of some close associate publicly, a realist." -Joseph Heller

User avatar
Santa Claus
Wakka
Posts: 181
Joined: Fri Aug 23, 2013 10:58 pm

Re: Why is P/BB so variable?

Post by Santa Claus »

DavidB256 wrote:
Mon Jun 10, 2019 8:04 pm
I understand that P/BB is a function of both the skill of the team and the skill of the teams they play against (playing against better teams would decrease your P/BB because they are more likely to get all of the easier bonus parts)
Kevin Wang
Arcadia High School 2015
Amherst College 2019

2018 PACE NSC Champion
2019 PACE NSC Champion

User avatar
Cody
2008-09 Male Athlete of the Year
Posts: 2385
Joined: Sun Nov 15, 2009 12:57 am
Location: Richmond

Re: Why is P/BB so variable?

Post by Cody »

As noted, bouncebacks are extremely opponent-dependent and won't tend to average out over a large sample size like other statistics such as PPG because there isn't a large sample size as compared to the variability in getting the bounceback correct.

Another consideration is that (iirc) SQBS reports bouncebacks as a function of fractions of a bonus. You have to divide P/BB by 3 to get points per bounceback. This greatly reduces the range and the other usual statistics (std dev).
Cody Voight, VCU ‘14. I wrote lots of science and am an electrical engineer.
VCU Tournament Director ‘13-‘17. HSAPQ President ‘15-16. ACF Treasurer ‘19-20.
Hero of Socialist Quizbowl Labor (NSC ‘14). “esteemed colleague” of Snap Wexley, ca. 2016. Stats Hero (Nats ‘16).
Quizbowl at VCU

Post Reply