Wednesday, August 24, 2011

Coach Writes in "Coaching Volleyball" on Using Stats

Oakland University (Michigan) women's volleyball coach Rob Beam has an article in the August/September issue of the AVCA journal Coaching Volleyball, on compiling statistics and using them to help one's team improve its performance.

The article, aimed mainly at high school coaches, provides a gentle introduction to volleyball statistics, with definitions of many key terms. Further, rather than presenting his ideas in technical terms (e.g., correlation coefficient), Beam uses a more intuitive form of exposition.

For each type of volleyball statistic, he presents a chart with the terms "absolute correlation" and "strong correlation" to winning or losing. The chart tells the reader how, for the midwestern high school competitions analyzed, performing above or below certain levels on eight different statistical indicators within a given match was associated with winning or losing. For example, teams recording a hitting percentage above .281 always won (an absolute correlation), those keeping their hitting-error rates below 14.2% usually won (a strong correlation), those having a ratio of service aces to reception errors below .33 always lost, and so forth.

With these benchmarks in hard, coaches can then work with their teams to raise (or lower) their numbers on the various statistical indicators. Beam also suggests a number of drills coaches can use in practice to enhance teams' competencies or minimize their deficiencies.

I e-mailed Coach Beam to share some thoughts and get his feedback. The hitting-percentage correlations to winning (.281 absolute, .172 strong) seemed fairly low to me, especially the .172. He attributed these findings to the closely matched competition in the late rounds of the high school tournaments he analyzed, in which teams' blocking and digging would keep down the hitting percentages.

I noticed one minor error in the article, in Chart A, where it appeared to define hitting percentage (efficiency) as ((KE)/TA). The numerator should, of course, be K minus E.

No comments: