24 NCAA Triangles Part 2
More on Terminal Serving
2024 NCAAW Season Analysis
Every year on SmarterVolley I have at least one content cycle that’s based around extensive statistical analysis of a large data set. American NCAA Women’s volleyball is pretty good at providing that, so I tend to pull heavily from data from the fall NCAA seasons, but some of the past analysis includes FIVB stuff, men’s stuff, etc.
2023 Defensive Analysis
2023 Offensive Analysis
2022 Major Competition Analysis
2021 Triangle Profiles
All of these have a similar flavor to them and I use many of the same tools, but I try to come at things a little differently each year so you won’t get bored and cancel your subscription can see some of the different methods of analyzing teams. Let’s dive into this year’s.
Quick notes:
First of all, if you are unfamiliar with the Triangle analysis framework or some of the other terminology I use here, check our the Triangle Primer that I updated for this year.
Second, some details on the data-set: I used the top-100 teams in the NCAA and pulled all of their dvw files from all matches. I didn’t exclude any matches, so everything regular-season and playoffs is in here. Also, Drake is not in here, even though they were top-100 in NCAA. Every now and then a match won’t important properly into Volleystation but for whatever reason more than half of Drake’s matches didn’t important properly. So… nothing from Drake in here. Sorry about that folks.
2024 Terminal Serving Details
Let’s go a little more into the sub-components of Terminal Serving, as well as look at the correlations between them and winning.
Terminal Serving and Win %
The X-axis is the % of Terminal Serving points won and the Y-axis is Win %. You have Pitt and Nebraska the two highest Terminal Serving teams and Lipscomb and OU as the lowest.
The correlation here is 0.35 to Win %, which is reasonably high. We can see there’s quite a bit of variation among top teams. Nebraska was way out on one edge of the graph (and so was Pitt, to an extent), while PSU was highly successful despite a fairly ordinary Terminal Serving rating. We’ve seen this before where Terminal Serving clearly influences winning, but not as strongly as the other two points of The Triangle.
Other Correlations
0.09 Terminal Serving Share
0.20 Ace %
0.351 Opponent Ace %
0.14 Serve-In %
Terminal Serving Share was uncorrelated to winning. So there are winning teams that are high-ace, high-error and there’s also winning teams that are more conservative. Both strategies seem to have potential to work just fine.
Ace % had some correlation to winning. Slightly higher than it did last year, but in the same ballpark: 0.20 vs 0.15. Serve-In had a similar, but slightly-lower correlation to winning than it did last year: 0.14 to 0.21.
Opponent Ace % had the highest correlation of any sub-factor at 0.35. Basically the same as the whole Terminal Serving aspect itself. I showed the leaderboard for Opponent Ace % in Part 1, so let’s look at the scatter plot.
There you can see some of the outliers in terms of giving up more or fewer aces and, in particular, a few of the more winning teams. If you look at a trend-line like this, teams that are above it won more than “expected” based on how many aces they gave up. Teams below it won less than “expected.” UT-Arlington is the outlier in terms of winning more than expected, given that they gave up a lot of aces. Indiana is the outlier in terms of winning less than expected, given that they gave up relatively few aces.
Takeaways
It’s interesting that we see passing metrics consistently correlate more with winning than serving metrics. My guess is that the dominance of float serving in NCAA Women’s volleyball means that serving is essentially uncorrelated with attacking, whereas passing is correlated to attacking. I’d be curious if this is the case in FIVB men’s volleyball, where serving ability (in the form of spike serve velocity) seems to be more correlated to winning.
I hesitate to make broad training recommendations based off these correlations, because I think the amount of subs and importance of recruiting always skews a few of these things when looking at NCAA data. However, I would say:
NCAA teams should probably feel pretty confident that they can train their way into good serving, rather than need to recruit it. I’ve been on the NCAA men’s side as well, and, on that side, I think there’s a recruiting element in terms of getting guys that are going to be able to hit at the velocity needed at that level. But in women’s volleyball, all the evidence seems to suggest that serving is trainable.
Quality passing is more rare than quality serving, and thus, relatively more valuable in women’s NCAA volleyball. I don’t know if that’s the case in elite men’s volleyball. Quality serving might be more rare. But I think in women’s/girl’s volleyball, it’s easier to train up your serving than your passing. This isn’t to say that you shouldn’t train passing… the difficulty of getting good at passing means you probably need to dedicate more time and effort to training your passers since there appears to be a greater difference at the margins.
Any training you do for passing probably also trains your servers, so its a bit of a moot point.
Next up in this series, we’ll start to look at First Ball performance and see how those factors influence winning!
The correlation is -0.35, because the more aces allowed the less you win and/or the fewer aces allowed the more you win, but I just made everything positive for ease of reading for less statistically inclined readers.