Route to understanding and implementing statistical timing models

Question

The research paper Forecasting NBA Player Performance by Douglas Hwang describes a method for predicting player performance by utilizing a statistical timing model and a 'Weibull-Gamma' distribution.

I'm approaching this from a rudimentary understanding of descriptive statistics and Bayesian inference, and would like to get some background to better appreciate, and potentially implement a similar approach to predicting individual performance (outside of the sports arena).

• What branch of statistics encompasses this sort of analysis?
• What is the usual route students/practitioners follow in learning how to design and implement these analysis?
• Any suggestions on books or resources (as applied as possible)? Or open-source libraries that provide implementations of this sort of thing?

Answer 1

The ASA has a section on statistics in sports. It cosponsors conferences like the one where you found Hwang's paper. There are sessions and a luncheon talk every year at the JSM. There is now a journal on quantitative analysis in sports. Serious articles for the layperson can be found in Chance Magazine and Significance Magazine. I was chair of the sections program at the JSM in 2002. Michael Schell has written two very serious books on baseball both published by Princeton University Press. One is about greatest hitters of all time (adjusting for era and ballpark effects) and the other is about home run hitting (where the ballpark effect is taken int account). Wizardry by Humphreys is an analysis of fielding skill in baseball. Michael Lewis' Moneyball that was very popular and made into a movie is about how usign sabermetrics allowed a low market team like the Oakland As get talent to form a winning team without paying out the big bucks.

Baseball is rich with data especially in the major leagues and now data is being collected and analyzed about every pitch and every park of the ballpark where a ball is hit.

You will also find another statistician Jim Albert has written popular books on baseball including one he coauthored called Curve Ball. There are now jobs in major league baseball for people like Bill James as well as for professionally trained statisticians. Bud Goode was one of the pioneers to consult with the Dallas Cowboys and other NFL teams years ago on football strategy based on simple regression methods.

There are hired statisticians in the NBA, NFL and MLB and the use of statistics in sports is growing. I have seen papers on tennis, golf, soccer and even cricket. I have spanned the gamut of my knowledge on this. I hope this has addressed your question.