Comments on: Friday Fun with Numbers: Shoot More Threes?

By: Today’s Celtics Links 10/17 « Flceltsfan’s Weblog

Today’s Celtics Links 10/17 « Flceltsfan’s Weblog — Sat, 17 Oct 2009 13:55:59 +0000

[...] Making Sense NBA Changes Its Traveling Rule… Sort Of I Gotta Read This Book Celtics Hub Friday Fun with Numbers: Shoot More Threes? ESPN Report: Stephon Marbury to sit out this season, eyes 2010-11 return to NBA [...]

By: Jason

Jason — Fri, 16 Oct 2009 17:43:02 +0000

Sorry, but still not sure what “raw” means. Is that all attempts (including fouled attempts) or just official attempts. I think the “all” data is out there somewhere, so you could re-run your comparison the same as before, but just with better data. You don’t have to do the regression.

By: Jason

Jason — Fri, 16 Oct 2009 17:40:09 +0000

Anyway, speaking of Sheed’s 3s, they are a double-edged sword to me. He’s percentage is acceptable, but not great, so it’s not like I want him taking a lot. Four a game isn’t a lot and some games he’ll shoot 10. I’m not sure that’s a winning formula.

Yes he spreads the floor, which has benefits. But he still seems to shoot just a few too many. He shoots them early when it’s not necessary. He shoots them contested when it’s not necessary. With just a tad more selectivity and patience, (kind of like how Scal only shoots when wide-open, but not that extreme), he could probably avoid at least one ill-advised 3 a game. That’s 82 low percentage 3s off the ledger. If he did that, his efficiency could skyrocket.

By: Zach Lowe

Zach Lowe — Fri, 16 Oct 2009 17:38:08 +0000

@Jason–nope, it’s just raw attempts. As I say in there, no sophistication at all. I have never done a multi-variable regression, though I know what they are. I honestly lack the expertise/training to do that, but I’m confident I could learn.

By: Jason

Jason — Fri, 16 Oct 2009 17:30:55 +0000

When you say “took the most shots” does it include all attempts, even those that weren’t official FGA because of a foul? If a team takes 10 inside shots and gets fouled on 5, that’s great and leads to winning (I would guess), but if in your breakdown they are only credited with 5 inside shots, then I think your .545 vs .529 winning %age comparison is skewed. If you put back in the wiped out attempts, the list of teams with the least and most inside shots changes and so to will the winning $age spread between the top and bottom.

Obviously, you could make the same adjustments to the other two categories, but I’m guessing things won’t change much there since fouling on jump shots is so rare.

Honestly, the proper way to do this analysis is a multi-variable regression where you can then get the weight of each variable as it pertains to winning. If you get the proper data, you should be able to run one (and understand the results) in Excel with very little research (if you’ve never run one before).