Measuring Improvement in Volleyball (Critical Thinking)

Even the most data driven coaches will likely have a hard time measuring improvement over time in volleyball. This is due to the cyclical nature of volleyball and the interplay of different skills over the course of every rally.

A player could receive serve at 55% perfect & good in a week of practice in September and she receives 65% in a week of practice in January. Some would say that 10% is a lot better… But can you say that that player is actually better?

What about if you have a player receiving 55% in September and she receives 55% in January? As a coach, you failed to make that athlete improve her reception in the past four months?

In both cases I don’t think the answer is very obvious.

In the second example, the receiver might be a lot better but your servers in practice are a lot better too. So, statistically, not only is your receiver the same, but so are your servers. But! Both are actually better. The servers are creating more difficult serves and the receiver is better able to handle those serves. So it seems like neither are better.

In the first example, it seems more obvious, that yes, your receiver is much improved because if the servers are the same she has a 10% improvement and if the servers are better than she has a better than 10% improvement rate. But if the servers are, in September, forcing bad receptions 45% of the time then, in January, they are only forcing bad receptions 35% of the time it seems like the servers are worse!

Maybe that receiver was just having a good week in January and she’s going to return to the mean the next week. Maybe she has got better but she’s just got better at receiving her teammates serves in practice and hasn’t got better at receiving the opponent you’ll be playing on the weekend!

Because of this interplay of skills in volleyball it is extremely difficult to judge improvement at the higher LTAD stages. The quality of serve is judged by the ability to disrupt the reception and the quality of the reception is based on the ability to manage the serve.

At the lower stages of LTAD it might be a lot easier to measure because we might have a model of what good technique looks like and we are measuring how close we are to attaining that model. If we know there is a certain way we want to put our wrists and hands together on serve receive we can measure how often our athlete is putting his wrists & hands together and see if it happens more and more frequently.

In that sense we are measuring improvement and keeping the athlete focused on the process rather than the outcome. This is not possible at the higher stages of LTAD where an athlete might have already developed their technique and individual style and we are largely outcome driven.

So measuring improvement at the T2C and T2W stages might be more subjective than we care to admit. I would argue that when you’re measuring improvement you might just be measuring what the athlete did that day (or week) and it might not be possible to use that to predict actual improvement. I guess that this is why we need to test statistical and practical significance when testing a hypothesis (for instance, hypothesizing that athlete X has improved from time point A to time point B). This might be what needs to happen in high performance gyms to really measuring the improvement of our athletes more scientifically.

No two rallies in volleyball are alike. What happened once will never happen again exactly as it did before. This means that for those of us who want to be data driven and scientific we might have to be honest with ourselves and admit that we’re going to have to rely on subjectivity when making decisions.

One thought on “Measuring Improvement in Volleyball (Critical Thinking)

  1. Another interesting blog, which is a bit more aligned to your research. While it is interesting to look at the absolute or relative change in performance, one must also consider the variability (Eg. Standard Deviation) in the data. Another way to look at this is through this example. In game A X player made average performance of 50 units with a standard deviation of 3.5 units. While in game B, X player made an average performance of 50 units with a standard deviation of 7 units. Which game did the player perform better. Overall the average performance was similar, but in game A, one could say that they were more consistent given a smaller deviation around the mean. This may not work well with individual (within subject) data but would certainly be a way of looking at team data. Anyway, it is an attempt to look at consistency as opposed to overall effect.

Leave a Reply

Your email address will not be published. Required fields are marked *