Behind the Laptop Week 12
Metrics don’t always measure what we want
If you’re interested in watching any of our matches, all of them are on ESPN+
Saturday 11/16
A friend of mine sent me an email this week. They asked my opinion about some concepts around serving and blocking. My reply had several visualizations, explanations, and further thoughts. Below is a bit of a summary of and reflection on my response to my friend.
Their first question was around blocking, particularly stuff blocks and what I call positive block touches (block touches that make the ball go up, go slower, or generally make it easier for the defense to dig the attack). Their question prompted me to explore the relationship between stuff blocks and other kinds of block touches. Along with positive block touches, I also use negative block touches (touches that make the attack go down or sideways, generally making it harder for the defense to dig the attack) and block backs (touches where the ball goes immediately back to the attacking side and the attacking team covers).
The plot above shows the relationship between stuff blocks and the three non-terminal block touches I listed above. Each mark represents a Pac-12 (R.I.P.) team. The counts come from matches I coded in the seven seasons I was at CU and in the P12, so each team’s marks represent 12-14 matches of block touches. The lines are regression lines for each of the three non-terminal block touches. The green marks and lines, showing block backs, provide the most robust conclusion of the three. If teams blocked more balls back at us, they also stuffed more of our attacks. The increase is almost one-to-one. Positive and negative block touches also show positive correlations but the relationships are not statistically strong. In sum, this plot has me thinking that using stuff blocks to describe the overall blocking quality of a team isn’t very helpful. Stuff blocks do tell me something about how well a team scores directly with their block but the non-terminal block touches should be accounted for somehow. From other research I have done, I know that positive block touches and block backs greatly increase a team’s indirect scoring. My friend’s question was helpful in pointing out that we don’t often pay attention to the indirect scoring improvements that blocking can bring.
Their next question was about serving, particularly what they perceive as large numbers of service errors in the name of serving aggressively. I shared another plot with them. This plot uses data from Evollve, a new site for college volleyball data.
The plot shows service error percentage as a function of service ace percentage. Each mark is a single team’s average service ace and error percentage for a single season. The data is saying that serving “more aggressively” does lead to more errors but it also leads to even more aces. There is a positive correlation, but it isn’t particularly strong. (You can tell it isn’t strong by looking at the distance between any mark and the regression line. While the dots trend upward, many of them are far from the line.) So my friend may be right about seeing more service errors but they may not be paying attention to the commensurate increase in aces. Again, these are measures of direct scoring. There’s plenty left to be explored in terms of indirect scoring that serving can create.
The last question my friend asked tied the two previous questions together. They pointed out that the best blocking teams in Division I earn about three blocks per set while the best serving teams earn closer to two and half aces per set. They were wondering if it wouldn’t be better to invest time in blocking rather than serving since there’s more scoring to be had. I responded with a few plots. Here’s one of them.
This plot shows the relationship between blocks per set and aces per set. Like the last plot, each mark is a single team’s average aces per set and blocks per set for a single season. Also like the last plot, there is a very weak positive relationship between them. The relationship is so weak that it would be more accurate to say that the two are not related. I created this plot to address an underlying assumption in my friend’s question, namely that coaches are choosing to serve more aggressively at the expense of blocking more aggressively. If that assumption were true, I think this plot would look very different. Teams that score more aces would have fewer blocks per set but they don’t. This suggests that coaches aren’t forsaking blocking for serving.
That relationship got me thinking about if some teams are good at both. With the exceptions of UTEP and Pitt, both in 2024 (they’re the two marks nearest to the top right), teams tend to be good at one or the other. Teams that get a decent amount of aces tend to be average at blocking. This doesn’t address my friend’s question about coaching serving instead of blocking. I can’t tell from the data how coaches are spending their time in training, all I can tell for sure is that the data doesn’t show a plethora of teams that are good at serving and bad at blocking (according to the metrics my friend brought up).
My friend asked a question which suggested that coaches are emphasizing serving while spending relatively little time on blocking. In the end, I don’t have data that can prove or disprove that relationship. But I think that the difference between points scored serving and points scored blocking is small enough that coaches can keep coaching the way they currently do. My suspicion is that it is easier to make average servers better than it is to make average blockers good. I think that it is hard to measure the differences between “average” and “good” in either of these skills. It is certain that blunt instruments like blocks per set and aces per set aren’t designed to measure those differences.
I’ll be presenting at the AVCA convention next month and I hope you can attend both the convention as a whole and my session in particular. My session, scheduled for Saturday, 12/21 at 11:30am, is called “Risk _____ to Win _____: Identifying and Managing Competitive Opportunities”.
