Page 1 of 1

[2020 Draft] Prospects' OBPM vs. top 200, 125, and top 50 competition...

Posted: Thu Sep 17, 2020 5:09 am
by getrichordie
Read on Twitter


Feedback is much appreciated.

Re: [2020 Draft] Prospects' OBPM vs. top 200, 125, and top 50 competition...

Posted: Thu Sep 17, 2020 12:06 pm
by jezzerinho
All info is positive, but I'm interested to know how well college OBPM Vs better college opponents predicts NBA offensive success. Clearly it cant hurt to be good Vs good players, but how much does it tell us? Any data to support it?

Re: [2020 Draft] Prospects' OBPM vs. top 200, 125, and top 50 competition...

Posted: Thu Sep 17, 2020 6:40 pm
by getrichordie
jezzerinho wrote:All info is positive, but I'm interested to know how well college OBPM Vs better college opponents predicts NBA offensive success. Clearly it cant hurt to be good Vs good players, but how much does it tell us? Any data to support it?


So, I cannot answer that question for you in an absolute way.

I took on this little project just because I wanted to see what players had a significant drop off in OBPM relative to different levels of competition. The reason I wanted to see this is because my theory is that less-athletic and/or smaller players tend to struggle against the better level of competitions on average and I wanted to try to capture that using OBPM because it generally encapsulates offensive production fairly well. Also, I tend to think that the more skilled and more confident a player is in his offensive game, there's going to be less of a drop off or even an increase in OBPM versus higher levels of comp. Some players just figure out how to get there's no matter what, IMO while others get stifled which I think is important to understand when it comes to prospect analysis.

Re: [2020 Draft] Prospects' OBPM vs. top 200, 125, and top 50 competition...

Posted: Fri Sep 18, 2020 5:29 pm
by pad300
A question, how many actual data points are we looking at here, especially in the high ( top 75 competition) category, and also for the freshmen? I suspect that you're into really small sample sizes here, where 1 bad game could throw things off a lot.