Spring 2023 Mailbag: Part 1

Men's reception, service error rate

Let’s kick this one off with Macabe Rappold’s comment in a recent post:

Where has beach week been?

Delayed, ironically enough, by the NCAA Beach Nationals. I was down there this past week doing some video and statistical analysis. I’ll have a lot of takeaways on the beach side next week. Until then, enjoy Part 1 of the latest Mailbag.

Every 3 months or so I put together questions that come in via Substack comment, email, or text message and post my responses here. I have a few other mailbag posts that you might want to check out as well.

---

Comment from Todd Elwell in the previous mailbag:

We've been having a discussion amongst our coaching staff about the best way to deal with managing seams on serve receive and two different approaches have come up. I wanted to run them by you.

Approach 1 is that the player moving to their right always takes the short seam, and the player moving to their left always takes the deep seam. This makes it so both players can make attempts to play the ball without colliding.

One of the really nice things about this approach is simplicity.

Approach 2 is that the seam management is dependent on where the server is serving from. If the server is serving from zone 1, the player moving to their right takes the short seam and the player moving to their left takes the deep seam - just like before. But if the server is serving from zone 5, the player moving to their right now has the deep seam and the player moving to their left has the short seam.

The nice thing about this approach is that when our players are standing and facing the server in zone 5, their right foot is already naturally behind their left foot - which makes it easier for them to take the deep seam to their right - and vice-versa.

I was wondering if you have any data on either of these approaches, or if you have any advice/experience on what is being done at the top levels. Any thoughts/ideas are appreciated.

(As a note here, Todd is coaching a pretty strong men’s team, and I’ll respond considering that level.)

Great question Todd and yeah, that's exactly the debate that you and it feels like every other team have.

To put it simply, I prefer Approach 1, but I think I'm in the minority there. I think a slightly higher % of teams use Approach 2. I see both though.

I like Approach 1 for two reasons. (1) As you said, the simplicity. (2) The most important thing that I know is that the zone 6 guy is better on his left than the zone 5 guy is on his right and that is regardless of where the serve is coming from. How can I set things up so that the 5/6 seam is handled more by the zone 6 guy than the zone 5 guy? That's why I favor Approach 1.

I get the arguments for Approach 2, but I favor both the simplicity and the handling of the zone 5 seam better in Approach 1. I think also if you slightly adjust the depths of your passers, this gets a little easier. If zone 5 is a touch shallower than zone 6 when the server is in 5, this crossing move gets easier.

---

In the post on NCAA Tournament Triangles, super-commenter Cal Pringle says:

I'd be curious to see some data from Stanford. Hambly had a quote that the broadcast was repeating over and over: "there is no correlation between missing serves and winning and losing games." I'd be curious to see how that philosophy played out in their serving game.

So first of all, let’s ask, is that statement by Hambly true true- is there really no correlation between missing serves and winning and losing games?

In the Club Month series, I shared some data from club nationals medal-round matches:

We can see here that, from a raw numbers perspective, that indeed there was no correlation between missing serves and winning and losing games. At U-15, the losing teams missed slightly more serves (as seen by the winning teams earning more points via opponent serve errors) but the winning teams at U-13 and U-17 missed slightly more.

Okay, but of course, the winning teams are serving more often, so let’s look at it from a % basis. Here’s a chart generated by Untan. The dataset is P5 NCAA Women.

Okay what the heck is this telling us? There’s a bar at different %s of service error and the corresponding # of sets won and lost. I put the slider bar at 0.084 (or: 8.4%) for a reason: if you win a set while missing 2 serves, you probably missed either 8% of your serves (2-of-25) or 8.3% (2-of-24) of your serves, depending on whether you earned the last point while serving or siding out.

You’ll also see that this is pretty common. About half of the time, a team misses <8.4% of their serves. You could do that while winning (for example, missing 1 out of 25 serves) or losing (missing 1 out of 20 serves, or missing 2 out of 25 serves and losing 25-27, etc).

You can also see that teams do win a bit more than they lose when they are under 8.4% error.¹ In general, missing 2 serves in a 25 point match is slightly predictive of winning. If you do a 2-variable analysis:

You can see there’s one result that’s impossible: Miss more than 2 serves AND have a service error rate under 8.7%.² You also can miss 2 serves or less with an error rate in the 8s or lower. That means either you missed 1 or 2 serves and won, or missed 1 serve and lost, or you missed 2 serves and lost a set that went into bonus points 26 or higher. You can also miss 3 or more serves and either win or lose.

And what this shows is that teams who miss less than 2 serves win more often than teams that miss 3 or more serves, when you also account for error rate.

However: this effect is pretty small. It changes your odds of winning the set by a few % points. If Kevin Hambly wins 56.7% of his matches for the next few years, he’s not getting his contract renews. So you can read his statement as directionally true and probably expand it a little more to say, “it’s not enough just to serve in, at a high level you also have to serve tough or you’re going to get clobbered by strong sideout offense.”

Now, let’s look at Stanford specifically. If your coach is saying that service errors don’t really hurt you, presumably you might serve a bit more aggressively, no?

Indeed, Stanford was on the aggressive side: they missed 12.1% of serves, which is above-average for an NCAA Woman’s team, especially a very good one. They earned 44.1% Break Points, a very strong number. And they held teams to just 45.3% Good Pass. So: Hambly is preaching an aggressive mindset, Stanford is playing that way, and having success.

Finally, let’s look at Stanford’s record when they serve in more or less often than their opponents:

10 Wins where they missed less often than opponents
16 Wins where they missed more often than opponents
1 Loss where they missed less often than opponents
4 Losses where they missed more often than opponents

Strangely enough, Stanford served in a lot more in the tournament than the regular season, missing only 7.6% of their serves. And in their epic battle against USD, Stanford served in at 96.2%, missing only 4 of 104 serves in their loss to USD, who also missed only 6 of 103 serves. (And it’s also worth noting that, while Stanford did lose that match, they outscored USD.)

You could look at this and say:

Stanford went 10-1 when missing less often than opponents.
Stanford went 16-4 when missing more often than opponents.

But I dunno… when you’re talking about a team that only loses 5 matches all year (and outscored their opponents in one of those 5), I wouldn’t parse that up. I would summarize by saying:

1. Yes, Stanford was more aggressive than most teams.

2. This aggressive serving helped (or at least, did not hinder) their Break Point ability.

3. Teams in general are served well by serving in, and 2 serve errors or fewer per set is a good goal for women’s volleyball. But this isn’t nearly as big of a factor as other aspects of the game.

Okay one more, behind the paywall: