Investigating the multivariate nature of NHL player performance with structural equation modeling

Abstract

Hockey is a complex and multifaceted game, yet many of the statistical tools used to evaluate performance are univariate. To garner a better understanding of hockey’s multifaceted nature, two structural equation models (SEMs) assessing the interrelations between offense, defense, and possession were built from three seasons of NHL data. Overall, it was found that the concepts of offense, defense, and possession are best understood via a small constellation of measured variables, and that offense mediates the relationship between possession and defense such that higher levels of offense leads to poorer defensive performance. These findings are discussed within the context of ranking player performance.

Silencing of LIMD1 promotes proliferation and reverses cell adhesion‑mediated drug resistance in non‑Hodgkin's lymphoma

“…Non-Hodgkin's lymphoma (NHL) is a group of malignant tumors of lymphoid cells, which are classified as indolent or aggressive, according to histopathological and biological characteristics of carcinomatous cells (1)(2)(3). The mortality rate of NHL was ranked in the top ten for cancer-associated mortalities worldwide (4).…”

Oliver’s four-factor model: Validation through causality

“…Mardia’s multivariate normality test reveals that the data do not follow a multivariate normal distribution: the chi-square statistic for skewness is 148.5 with an associated p-value of less than 0.0001, while the kurtosis Z-score is 2.025 with an associated p-value equal to 0.043. However, prior research on SEM suggests that violating multivariate normality does not undermine the findings ((6), (11)): there is compelling evidence that maximum likelihood estimation is robust to normality variations, especially when sample sizes are large (e.g., N > 600). On page 175 in (6), a table shows the simulated results of Mardia’s multivariate estimates of skew and centred kurtosis on datasets with different distributions and with a different number of observations.…”

“…There has been very little work involving the analysis of teams or players using structural equation modeling. Two examples are the article by Riley on N H L players performance analysis ((11)) and the work by Baghal on N B A teams winning percentage ((12)). Baghal presented a study of Oliver’s four-factor model that includes three latent variables into the model specifications: the offensive quality of the team, the defensive quality of the team and the salary of the team.…”

Analysis of goal scoring opportunities in elite male ice hockey in relation to tactical and contextual variables

“…In competitive ice hockey, studies focusing on technical and tactical aspects are sparse. In recent years, technical/tactical studies in ice hockey have been investigated with focus on individual players contribution (Macdonald, 2012;Riley, 2017), as well as differences between winning and losing teams . Due to the complexity of ice hockey, the sport requires several different approaches of match analysis to cover the whole performance dimension spectrum.…”