I'm a statistician. It's what I do.
This post is short and sweet. I've covered this topic before, so I won't belabor it. I think there are players that rack up a ton of aces, but this is offset - and then some - by boat loads of errors. I think many people see the aces totals and believe this helps their teams. I disagree.
If a player has 47 aces and 77 service errors, then you have to demonstrate to me 30 non-ace definitive times that their serve has manifested itself in a point immediately following the serve without rally JUST TO BREAK EVEN. Once these 30 times have been demonstrated, this server has had the same impact on scoring as someone who has zero aces and zero service errors, or 20 service aces and 20 service errors. This assumes, of course, that these last two servers NEVER had situations where their serves resulted in non-ace, yet immediate scoring opportunities for their team (which probably is not true and thus makes the 47 ace, 77 error server still look less valuable than them).
Think about what has been written above. Really think about it. We glorify aces, but turn our back on a mountain of errors. Have we lost our minds?
Now, here is a graph of the Top 40 players in NCAA Division I volleyball in aces. The horizontal axis is aces and the vertical axis is service errors. Can you spot the outlier? Hint: 47 aces, 77 errors.
Look at where the bulk of the data lies and then consider this: If you remove the outlier, there are a total of 1100 aces represented and 1152 service errors- two values that are roughly the same when considering there are 40 players represented.
In fact, just to add a little statistical flavor to this, if you calculate the correlation between aces and errors WITHOUT the outlier, you get a correlation coefficient of r =.17. For the nerds among us, the p-value for that being significantly different than zero is .30. So, simply speaking, without the outlier, the above cluster of points is essentially random scatter.
If you re-run the correlation analysis WITH the outlier in the data set, then the correlation coefficient jumps to .54, which is significantly different than zero (p-value now .0005). Thus, the inclusion of the outlier changes the correlation from being considered statistically zero to statistically non-zero BY ITSELF. THAT, my friends, is an OUTLIER.
Four of the 40 data points in the above graph belong to Southland Conference players:
Heather Schnars (UCA): 47 aces, 77 errors
Nat Jaeger (NW St): 33 aces, 40 errors
Jessica Wooten (HBU): 29 aces, 21 errors
Kristyn Nicholson (TAMUCC): 25 aces, 41 errors
Nicholson's data point looks slightly out of line with the national trend among ace leaders, but Schnars is the outlier.
Heather Schnars is the reigning Southland Conference Player of the Year. Heather Schnars is one of the best players to ever play in this conference.
I picked Devaney Wells-Gibson as my player of the year last year in part... I emphasize IN PART... due to this type of analysis. And, I will reiterate one more time for the record, that I have ZERO qualms with Schnars winning last year.
Heather Schnars leads the nation in aces. Heather Schnars leads the nation in service errors.
By the way, Northwestern State leads the nation in service errors and UCA is second.
The Southland Conference leads all conferences in total aces. The Southland Conference leads all conferences in total service errors.
Aggressive serving to the point of being an outlier as mentioned above? I say it isn't worth it. And I'd coach the player to stop. You'll need to provide me with quantitative evidence to the contrary to get me to change my mind. Don't start telling me anecdotal stories of getting teams out of system here and there. Service errors give the other team a point 100% of the time they occur. I want data.
Has anyone else noticed all this?