Fall 2023 Mailbag Pt3

Men's serving, serving metrics, and attacking a rotation defense.

Two more questions for the mailbag that didn’t fit into the first part. I’ll keep these both outside the paywall. Of course, if you’re enjoying these posts, please become a Premium Subscriber and get access to more mailbags, Q&As, and the ability to post-comment.

Fall Mailbag Pt1
Fall Mailbag Pt2

Email from a reader:

I was just looking at the email today from men’s WC and the question popped into my head. I wonder if there will be a team out there that goes with an all-float lineup, with the exception of one big bomber or something. Swing the pendulum back to the “develop our good pass block/D and make them play” kind of thinking.

A few thoughts:

Tactics can either be dominant or equilibrium in nature. A dominant strategy means it is superior to another full-stop. In a purely mathematical game like poker, this can be expressed very clearly. If the flop is 7/8/9 and I’m holding 10/J and you’re holding 5/6, I have you dominated. There’s no 2 cards that can improve your position and win the hand for you¹. And then you have situations like A/Q v A/J v K/10. A/J in a vacuum is a better hand than K/10, but if we both go all-in preflop and I turn over A/Q, you’d rather be holding K/10 than A/J.

And on the far end of this, you have rock-paper-scissors. All 3 are (theoretically) perfectly evenly matched in terms of expected value, but they do so by either 100% chance of win or loss depending on what your opponent throws. If you visited some strange land where nobody ever threw scissors on the first turn, you would win a lot of games by leading with paper. This is an equilibrium strategy because your success rate would be changed by what your opponent does.

I promise I’ll get to the volleyball aspect in a second, but I try to imagine strategies as either dominant, dominated, or equilibrium-based strategies. Always folding pocket pairs and only playing unpaired/offsuit low cards in poker is a mostly-dominated strategy. While the luck could bounce your way from time to time, you’re playing objectively worse cards. On the flip side, consistently folding weak hands and consistently playing strong hands is a mostly-dominant strategy. Sure, you don’t win every time you play AKs but you’re improving the odds.

From there, how you play the hands is where the equilibrium comes in. The old cliche of, “be aggressive at a conservative table and be conservative at an aggressive table,” is a cliche for a reason: it’s describing equilibrium dynamics. You can’t mathematically prove that aggressive raises or slow-play is the correct strategy² because the effectiveness depends on what your opponent does. Equilibrium.

Alright, alright. Volleyball. What does this have to do with float serving v spin serving.

So first, let’s put aside the question of, “how good is your float/spin serving,” and go with the axioms of:

Float serving will produce fewer KOs and aces, while also making fewer errors.
Spike serving will produce more KOs and aces, while also making more errors.

Is one of these strategies dominant over the other? Yes but… it does depend on you and your opponent. So there’s some equilibrium here, but not in the serving itself. We have to factor in Earned Sideout. Let’s look at 2 extreme situations.

Situation 1: The other team has an Earned Sideout of 10%. That is, if you serve the ball in the court, the other team will almost never win the rally. Or, as some would call this, 12-2s volleyball.

Situation 2: The other team has an Earned Sideout of 90%. That is, if you serve the ball in the court, the other team will almost always win the rally.

In Situation 1, a serving strategy of, “just get it in,” is clearly the dominant strategy. Any ace only gets you an “extra” 0.10 of a point, because the other team usually just makes an error as long as you serve it in. And any service error costs you a full 0.90 of a point. And in Situation 2, it’s flipped. Be insanely aggressive because the cost of an error is only 0.10 point while an ace is worth 0.90.

Okay, so real life isn’t 10% or 90%, what is it? Let’s look at 3 examples: 50% eSO, 60% eSO, and 70% eSO.

50% eSO is decent-level high school volleyball. Given that you’ve served the ball in the court, you and your opponent are relatively equally likely to win the rally.

60% is high-level women’s volleyball. Some of our matches here in Korea are at or above 60% eSO and a bit more than that are under.

70% is right at the top of elite-level men’s volleyball. For example, Italy won the 2022 World Championships with a 71.6% overall Sideout and 69.6% eSO.

From a purely aces and errors strategy, in the 50% eSO meta-game, you’d want aces to be equal to errors, because their magnitude is equal. Missing a serve costs you the 0.50 of a point that you’d expect to get by staying in the point and playing defense. Whereas serving an ace gains you the 0.50 of a point you’d expect the other team to get by winning half the rallies when you serve the ball in the court.

Alright, so how about 60% eSO? You guessed it, a 4:6 (or 2:3) Ace:Error ratio is probably about the equilibrium. (2 Aces x 0.60 per ace) = (3 Errors x 0.40 per error).

And how about the 70% eSO elite men’s game? The math would suggest 3:7. (3 Aces x 0.70 per ace) = (7 errors x 0.30 per error).

How did Italy’s World Championship winning team match up with that? They served 22 aces and missed 101 serves. By that metric, they were (22 x 0.7) - (101 * 0.30) = -14.9 points in the whole. They played 25 sets, so, in theory, they were costing themselves about 6/10 of a point by serving “too aggressively.” And their opponents were almost exactly the same, serving 23 aces and missing 105. Although worth noting that Italy had quite a few more “zero” serves than their opponents. The raw totals are similar, but because Italy was winning most of their sets, on a % basis, Italy was at 20.1% overall Terminal Serves (Aces + Errors) while their opponents were 25.4%.

Now, people are not computers where you can cleanly dial up and down some Aggressiveness Meter. Perhaps their opponents were being more aggressive against Italy because they knew Italy was a great team and they were okay with missing some serves. Perhaps they weren’t and they just missed serves because they screwed up. Impossible to know for sure.

But the initial question is essentially asking: is high-risk, high-reward serving a dominant strategy in the men’s game, or might we see the pendulum swing back?

From a mathematical sense, the answer seems like it has to be, yes, it’s possible than a men’s team could win by employing all float servers because it’s possible that most teams are a bit too aggressive.

And also, from a pure point-scoring perspective, there are dozens of float-serving men who have a Break Point % at or over 40%. In FIVB Men’s Volleyball, if you keep the other team at or under 60% Sideout, you win. So there’s no theoretical reason why you couldn’t assemble a team with 6 float-serving (or, given modern men’s game, they would probably be hybrid/off-speed guys more than pure floaters) guys who could BP at 40%.

But also, because more of the elite servers (at or above 40% BP) are spin servers, I also can’t say that you should try to assemble this team.

And if we’re going to talk equilibrium strategies, we’re also starting to see fewer pure float or pure spin guys. We see a lot more guys who have multiple options with their serves: dropping floaters, short rollers, cross-court cuts, etc.

And it’s inevitable that we’re going to see that in the women’s game as well.

---

An email from a college coach:

I hope this email finds you well. My name is [X] at [Y]. I had a question about calculating serving percentage. I was trying to figure out how to calculate one. I first thought about doing one based on what the opponents passed based on each server. However, I thought that this way doesn't calculate any serving errors. So I thought about doing it as a 5 point scale as follows 4=ace, 3=bad bass by opp, 2=medium pass by opp, 1=good pass by opp, and 0=serving error. Am I on the right track? Do you have a scoring system for getting a percentage for servers?

For me, the 3 main serving metrics I like are:

1. KO%

2. Expected Sideout %

3. Opponent Sideout

If you don't mind me linking those articles directly as explanations, they explain those pretty clearly. And then Opponent Sideout is what they ACTUALLY sided out when that server is serving. Which, of course, is the ultimate measure of a server, but can be thrown off by rotation-specific factors like your best blockers being in the front-row, etc.

KO% is the simplest and you can create drills based off it because most servers are going to be between 40-60% KO so it sort of slots in there.

Expected Sideout is complex and doesn't necessarily make sense to players when you explain it to them, but it's the most "mathematically pure" formula of the 3.

Hope this helps!

1. 10/J would result in a draw.

2. Maybe they have in the past few years, I haven’t been following the research lately…