I discovered a great site over the week. It’s called dougstats . The greatness of the site is in that the owner does the public service of collecting real-time stats for the nba and mlb for use and recreation.
What attracted me was that he was collecting pre-season stats for the NBA and as I’m model building to predict wins and project player performance I was curious to see if the pre-season had any kind of statistical significance to the following season. To answer the question I decided to look at 2009 pre-season stats and compare them to 2009-2010 stats and look for correlation.
After I did that I went ahead and worked out the correlation for ADJP48. That’s here:
We see 11 guys who were out of the top 50 for the year. We also see Blair and Lawson (the two draft steals). The preseason can show you rookies who are arriving early but for returning players the larger sample size of previous seasons is more appropriate.
Given that I will work out pre-season WP numbers for all rookies in a future post.