Tuesday, November 18, 2008

So I am laying her in my bedroom, listening to the Blazer game. You see I don't have Comcast, nor is there a Trailblazer radio affiliate in the Astoria area. Luckily, from one spot in my bedroom I can pickup the broadcast on 95.5 out of Portland. Anyway, none of that has a whole lot to do with this post. This post is actually another in my sporadic series looking at modern statistical analysis and the NBA.

It is still pretty early in the season, so any statistics we could look at would be significantly limited by small sample size limitations. However, I came across a couple of articles at Basketball Prospectus (main page: here) that I think bear some comment, and with a little luck, some discussion. Kevin Pelton is the main writer for Basketball Prospectus, and he has worked to develop a projection system called SCHOENE that is (at least in some ways) similar to PECOTA, a baseball projection system. See here for a nice in depth explanation. In essence, SCHOENE uses last years stats to build a projection for every player that played at least 250 minutes in the previous season. Once the player projections are finished, those can then be used to build team projections. There are obviously several important adjustments made, if you are curious about those follow the link above.

So, that is a long explanation that you probably didn't need just to link to this. That is the SCHOENE projection for the Northwest Division this year. A little surprising to me was the projection for Portland: 51-31 and the division winner. Now don't get me wrong, I love the Blazers. I am just not sure they are 50+ win team this year. Compare this with Hollinger, who looks at Portland as more or less a .500 team. In some ways I tend to think Hollinger's projection is closer to reality, but honestly I am not qualified to discuss which system is stronger. I do think it is important to point out that Hollinger uses 3 years of weighted statistics rather then just last years. Anyway, I thought it was pretty interesting.

PS. Kind of funny, looking at Lamarcus Aldridge's closest comparison (that is what player of similar size, position, etc. did his last season match closest): Channing Frye. 

2 comments:

Rip Tatermen said...

Wow, TMI, buddy. I mean, the Blazers have been pretty sexy (exceptin' tonight's game, or at least the three quarters I've watched), but seriously, we don't need to know about your sports-radio fetish.

Beau said...

Ah crap. You know I try to proof-read before I hit publish, but clearly I missed that one. Now I have to resist the urge to edit and correct it...