Page 1 of 1
Distribution analysis of offensive activity
Posted: Tue May 31, 2011 5:42 pm
by Crow
http://basketballanalytics.blogspot.com/
Have you compared the points per possession of teams with equal levels of the offensive stats you studied (or near equal bands of the overall distribution) but different standard deviations for their distribution? I would think that would be useful to see and might better isolate the significance of the distribution of the activity from the overall amount of the activity which can vary by the distribution (degree of standard deviation) and is not directly displayed in the blog charts.
Re: Distribution analysis of offensive activity
Posted: Tue May 31, 2011 5:55 pm
by Nima
I have not yet. I certainly want to and plan to take a look at specific cases and compare; most likely after the finals.
I'm not sure what similar levels of offensive stats I would compare, though I assume you mean compare teams with similar per possession fga, 3pa, fta, oreb, ast and to. I realized the flaw in my earlier analyses, where my standard deviation variables were being skewed by the total in the per possession statistics, so in my most recent analyses I reduced each player's stats to a percentage of the whole rather than taking the 5 different per possession stats across various categories (fga, 3pa, fta, oreb, ast, to). I think it would be a good idea, to look at lineups whose per possession numbers in those 6 categories are most similar to see what difference the distribution in those 6 categories makes.
Thank you for your feedback. It is always appreciated.
Re: Distribution analysis of offensive activity
Posted: Tue May 31, 2011 6:13 pm
by Crow
Yes. my quick comment was based on a quick look first at your original presentation materials and I now see your shift to percentages in more recent posts and it is helpful. Your commentary on FTAs and assists addresses part of what I was drawing attention to. Your later finding with regard %3ptFGA was interesting to see.
If you do further analysis of bands of near equal totals of certain offensive activity to look at the efficiency variation associated with different distributions of that activity, let us know.
Re: Distribution analysis of offensive activity
Posted: Tue May 31, 2011 6:49 pm
by Nima
Will do. Thanks, again.
The result with respect to %3PtFGA distribution in particular is why I like Dallas in this series. They have played better team basketball than anyone and seem to get good, open shots at will from shooters who hit open shots with the best of them, especially from beyond the arc. If Dirk remains unguardable inside without a double team, with Dallas' discipline and ability to move the ball to get uncontested three point shots, Miami will be hard pressed to stop them. It might require the refs letting Miami get away with aggressive D on Dirk that has found him at the line in the playoffs so far and it will be interesting to see if Dirk gets the same respect from the refs in the finals as he has in the rest of the playoffs. I also look forward to seeing how much Lebron will guard Dirk and whether he can slow him down without help. The same goes for Dallas creating open shots through the pick and roll/pop thanks to Barea's uncanny ability to finish in the paint. If Miami has to help, Dallas will find those shooters. Dallas' offense will be a great test for Miami's D.
Granted, my analyses so far are for the offensive side of the ball only, but Dallas has played near flawless team basketball offensively from what I have seen in the last couple series and casual observation tells me they've played some pretty good team D too.