Per-minute statistics in basketball. Minutes played affect performance?

We are pleased to announce a new publication. This is the provisional reference:

Martínez, J. A. y Martínez, L. (2010). El uso de indicadores de desempeño normalizados para la valoración de jugadores: el caso de las estadísticas por minuto en baloncesto. Motricidad. European Journal of Human Movement. Accepted for publication

This research empirically tests the hypothesis underlying the use of performance indicators in basketball, when they are normalized using the “per-minute” criterion: There is no association among normalized indicators and the minutes played by each player. This assumption is a necessary requirement in order to compare players using normalized indicators, because if this assumption is not met, then there would be a possible bias in the rankings of each statistical category. We apply several under-used methods in this type of research, such as the analysis of the intra-player variability and the structural equation modelling. Results show that there is no lineal association between the minutes played and the different per-minute normalized categories as: points, rebounds, assists, steals, and blocks. We discuss implications for sport performance and marketing. Therefore, this study contributes to the advance in one of the growing areas of sports research; such as basketball metrics.

Download paper: http://www.upct.es/~beside/Textos/EXMIN_JAMG.pdf

Please, be free to give us any feedback or comment.

A review of basketball player evaluation metrics

I would like to share a new publication. This is a series of three papers about player evaluation metrics. These are the abstracts of the three articles, and the links to a version of the papers. These papers have been accepted for publication in the "Revista Internacional de Derecho y Gestión del Deporte", and will be published shortly.

Papers are written in Spanish, but you can try to translate them to English (Thanks Crow) using: http://translate.google.com/# 

Please, be free to give any feedback.

Paper 1: A review of the basketball player evaluation metrics (I). A description of the existing methods

This research achieves a complete revision of the basketball player evaluation metrics, which are currently used by analysts. The aim is to depict the variety of systems proposed regarding the productivity of players, which can be handled by team managers in order to make more rational decisions. This is the first comprehensive review published in academic journals, and it provides a framework to analyse the “statistical revolution” started in United States. Likewise, systems applied by official leagues and fantasy games are depicted. This research is the first of a series of three papers that pretend to facilitate the understanding of a plethora of formulas and mathematical methods.

Download paper: http://www.upct.es/~beside/Textos/SVI__JAM.pdf  Updated! (5 April 2010)

Paper 2: A review of the basketball player evaluation metrics (II). Official competitions and fantasy games


In this second article of the trilogy on basketball player evaluation metrics, the systems used by the main competitions around the world and fantasy games are depicted. In both contexts, these systems play a fundamental role from an economic viewpoint, because these systems are used to grant players with individual awards, and to provide diverse prizes for success. The current state of development of information and communication technology is very asymmetric among leagues. However, the main competitions have the possibility to work with play by play data and, consequently, to use statistics in the most powerful way. The vast majority of leagues and fantasy games work with very simple systems, based on linear weights and box-score data, whilst only a few competitions use more advanced metrics, such as Net Plus Minus.

Download paper: http://www.upct.es/~beside/Textos/SVII__JAM.pdf  Updated! (5 April 2010)

Paper 3: A review of the basketball player evaluation metrics (III).  Overall discussion.

In this third article of the trilogy on basketball evaluation metrics, strengths and shortcomings of the different evaluation systems are discussed. In addition, diverse statistical analyses are achieved, in order to study the relationship among different systems, and discussing some assumptions. Moreover, this paper provides a framework to discuss about the basis of these metrics, proposing several ideas related to validity criteria, formative measurement, and the use of qualitative information. Finally, implications for further academic and applied research are discussed. 

Download paper: http://www.upct.es/~beside/Textos/SVIII__JAM.pdf  Updated! (5 April 2010)

Statistical winners of ACB League 2008/2009

Comparison of statistical winners of ACB League 2008/2009, after applying the probabilistic method depicted in Martínez and Martinez (2010). Changes are highlighted in red.

ACB criteria:
Evaluation/Ranking: F. Reyes
Points: I. Rakocevic
Assists: R. Rubio
Rebounds: C. Bordchart
Blocks: F. Vázquez
Steals: R. Rubio

Probabilistic criteria:
Evaluation/Ranking: F. Reyes
Points: I. Rakocevic
Assists: P. Prigioni / C. Thomas / K. Valters
Rebounds: F. Reyes
Blocks: F. Vázquez  / L. Barnes
Steals: P. Prigioni

.

New publication: A probabilistic method to statistically classify players in basketball

We introduce a method to re-elaborate the rankings of individual stats in the ACB League. The method is based on a probabilistic approach to interpret the individual performance achieved by each basketball player at the end of the regular season. Therefore, each individual record is an estimate of the real value of the parameter, with the corresponding associated error. The size of this error influences the final elaborated ranking. Under a parametric approach, we quantify the size of a maximum admissible error. Because of the growing interest of statistics in basketball, our proposal is a valuable contribution to analyse the performance of players, which is highly related to their market value.

This is a new research that has been accepted for publication in the "Revista Internacional de Ciencias del Deporte".

Martínez, J. A. y Martínez, L. (2010). Un método probabilístico para las clasificaciones estadísticas de jugadores en baloncesto. Revista Internacional de Ciencias del Deporte, 18 (6), 13-36.

You can freely download an expanded version of this article: a working paper entlited:

Martínez, J. A. y Martínez, L. (2009). Un método para reescribir las estadísticas individuales en la Liga ACB de baloncesto. Working paper. Universidad Politécnica de Cartagena.

Please, follow this link to download the article:

http://www.upct.es/~beside/Textos/Martinez&Martinez_Working%20Paper_Basket.pdf

Please, be free to give us any feedback or comment.

Macro-Panel of Experts

We are conducting a research about the importance of statistics in basketball. In this first stage of this research, we would be very grateful if you could participate in this survey, as a member of a panel of experts.

We created several panel of experts representing stakeholders in basketball (players, coaches, analysts, researchers, fans, bloggers, journalists...). If you are one of these stakeholders and you want to participate, please email us: josean.martinez@upct.es

We will send you a very brief questionnaire that can be fulfilled in 8-10 minutes. Your collaboration will be very important to this research. Experts of the main basketball leagues of the world are participating in this macro-research. All this information will be treated in strict confidence and you anonymity is guaranteed.

In addition, although your anonymity is guaranteed, if you desire we will promote the webs of all partners that are collaborating with this research. Moreover, we will reference these webs in tha academic publications derived from this study. Finally, all participants in this research will be informed in detail about the results.

Please, do not hesitate to contact me if you require further details. Thank you very much.