I need to vent about statistics. Actually, it’s probably more about the systems they use to take stats. I’m referring primarily to the college game, but this could apply anywhere.
Let me explain.
In the NCAA they take stats during matches that end up in the box score. They’re all the standard ones you’ve seen for years. Unfortunately, they don’t cover things like serving and passing ratings or rotational breakdowns. If you want those, you have to use a secondary process, which I do.
Now, it’s not really a surprise that two different runs at stats on the same match come up with different numbers. But usually it’s a minor thing. It’s just small differences in how they count certain things. That’s not necessarily the case when it comes to some of the stuff I’ve seen coaching at Medaille as it relates to attacks, however.
Here’s an example. In one tournament we played three matches. The box score indicated we had 122, 101, and 132 attacks in those matches. My own secondary stats put it at 112, 88, and 111 respectively. That’s about a 12% difference.
A definition
To provide the baseline standard, here’s the exact verbiage from the 2021 NCAA Women’s Volleyball Statistician’s Manual.
SECTION 1—ATTACKS
Article 1. An attack attempt (ATT) is recorded any time a player attacks to attack (hit strategically) the ball into the opponent’s court. The ball may be spiked, set, tipped or hit as an overhead contact. There are three possible outcomes of an attack attempt:
(1) There can be a kill.
(2) There can be an attack error.
(3) The ball can stay in play. This is referred to as a “0 attack” (zero attack).
Philosophy. Any ball that is played over the net in an attempt to score a point should be considered an attack. Any ball played over the net simply to keep the ball alive should not be considered an attack attempt. The exceptions to an attack attempt are:
(1) An attempt is not charged on a ball played over the net on serve reception that is kept in play by the opposing team. This is called an overpass.
(2) An attempt is not charged on a free ball played over the net when, in the opinion of the statistician, the free ball is passed only to keep the ball in play.
(3) An attempt is not charged to a player if, in the opinion of the statistician, the set is bad and the player plays the ball over the net only to keep the ball in play.
(4) An attempt is not charged to a player if, in the opinion of the statistician, the player passes the ball over the net only to keep it in play.
However, if in any of the four above-mentioned instances the action results directly in a point for the team playing the ball, a kill (see Article 2), and therefore an attack attempt, must be awarded.
Too many attacks
In doing some of the matches for the 2021 Medaille women’s season I noticed a considerable difference between the total attacks I counted and what showed up in the box scores. My conclusion was that they were counting non-attacks as attacks, which shouldn’t be the case. The NCAA manual is very clear on that, as you can see above.
So how do we end up with too many attacks in the box score? The system people use likely has a lot to do with it.
I had the opportunity to actually witness in-match stat-taking once during a match hosted at Medaille. Basically, one person just called out the number of the player who just touched the ball, while the other entered it into the system. The software essentially fills in what those contacts represent. In other words, they aren’t actually entering what the touch was, just that a touch happened. All works great if teams play clean pass/dig-set-hit sequences.
Guess what, though. That’s not always the case. When you’re at the lower levels of play you get more free balls and other non-attack balls going over the net. They shouldn’t count as attacks, but this system does just that. And because it over-counts attacks, it also over-counts digs.
Uncounted aces and kills (and blocks)
And because the system doesn’t know whether contacts are under control or not, it actually creates problems in other places as well. I’m talking aces and kills particularly.
Consider the situation of a shanked dig or reception where one or two other players touch it, but can’t control it enough to keep it in play. In that case, you go back and credit an ace or kill. At least that’s the instruction. If you’re just typing in the number of a player touching the ball, however, the system doesn’t know that the ball never came under control. Thus, the no credit for the ace/kill, and some kind of error shows up in the wrong place.
This holds true for blocks as well.
An actual example
In a 5-set match during the 2021 season the box score showed Medaille with 55 kills and 32 errors on 192 attacks. That’s a .120 hitting efficiency. My stats for the match came up with 60 kills and 32 errors on 183 attempts. The hitting efficiency based on those figures is .153. Pretty big difference, eh?
And for the digs. The box score put Medaille at 97. By my count it was 82. You can quibble on a handful, but 15 is a pretty sizeable difference.
I counted 7 blocks for us. The official number was 5.
On the plus side, my ace counts matched the box score figures.
Implications
Statistical analysis can only be as good as the data. If we collect the stats incorrectly, then our data is incorrect. If our data is incorrect, our analysis is of dubious value. We have a situation here where we count too many attacks and not enough kills. That means our Kill %, Error %, and Hitting Efficiency numbers are all off. If the errors are consistent, and you know how big they are, you can factor that in. If not, there’s not much you can do.
6 Steps to Better Practices - Free Guide
Subscribe to my weekly newsletter today and get this free guide to making your practices the best, along with loads more coaching tips and information.
