Battle Cats: Cougars & Tigers - Round 2
Battle Cats: Cougars & Tigers - Round 2
BYU swept Princeton in round one, but the felines are right back at it tonight. We also have some less cat-centric matches as well.
In a rare Monday match, BYU was able to keep Princeton to .152 hitting while themselves hitting .328. Although this was a sweep, Princeton was challenging them the entire time. I say challenging because Princeton used a lot of challenges. In the first set alone, head coach Sam Shweisky used three challenges! Even though he had lost two earlier in the set, Shweisky went for it late in set 1 and won on his third challenge. This is the type of bold, calculated risk I like to see.
There are a handful of other matches around the league today as well. Most of them appear to be fairly one-sided, but looks can be a little deceiving. For the matches with a D-III and NAIA team, the VBelo model doesn't know much about them. So those projections are not as precise.1 We also have Thomas More, a new team, playing today. Since they are new, the model is still learning a lot about the strength of this team. One of the things that I love about predictive models is that they are always learning. The more data, the more it knows, and the better it can project outcomes.
A New #1
In the latest version of the AVCA Coaches Poll, Ohio State topped the rankings. They garnered a majority of the top-place votes with 12. After their impressive 5-set wins in Texas, the coaches saw it fit to reward them with the coveted #1 ranking.
Here is my problem with the polls: they have short-term memories. Early in the season, it is still difficult (impossible?) to know who the best team in the country is. The sample size is just too small. The fluctuation in rankings at this point in the season is based on one or two "impressive" matches.
I don't think we should get rid of the polls.2 I think they serve a very important purpose. Some of the best volleyball minds in the country are assessing what they are seeing. That is valuable. In addition to that, something like the VBelo model is helpful. VBelo has a longer memory and doesn't have a recency bias. If anything, it has a past bias. It also doesn't care about the name of a team's jersey. The model cares about results over time.
We need both short-term and long-term memory analysis to get a holistic picture of the volleyball landscape. One without the other ultimately falls short. So, as much as I dislike the one-match-sample-size of some of these rankings, I know we need it. It helps us ask better questions about these teams and the sport.
Match Recaps
Non-conference
BYU (67%) def. Princeton (33%)
3-0 (25-23, 25-19, 25-21)
Ben Harrington Ace Watch: 3 yesterday, 24 total, 1.09 aces/set. Those last two are nation leading.
Match Projections
Non-conference
North Greenville (77%) vs Tusculum (23%)
Princeton (29%) vs BYU (71%)
Mount St. Joseph3 (22%) vs Thomas More (78%)
Park4 (8%) vs Quincy (92%)
I’ve looked at the difference between these matches and matches that feature only D-I/II teams. The difference in their predictability was surprisingly low, but there is a difference. Since the model uses fixed ratings for D-III and NAIA, the projections can only be so accurate. Maybe one day I will have data for all the volleyball.
Full disclosure, I am a voter in the Off The Block media poll.
D-III
NAIA