OT: basketball sabermetrics?

by lincypoo i wuv u on Jul 28, 2009 2:15 AM PDT Comment 15 comments

For baseball I enjoy some of the statistical analysis on this site, fangraphs and so on. It goes well with my fantasy sports addiction. I realize that since basketball is largely a team sport. It's also completely totally offseason but still. I was wondering if there was any metrics community for basketball. I go to basketball-reference sometimes, which was what I used to make incredibly incorrect playoff predictions. Any suggestions? Also I'm going to assume that predicting football using math (except maybe turnover differential) is absolute madness but if you can point me to any wizardry about that I'm all for that as well.

This FanPost is reader-generated, and it does not necessarily reflect the views of McCovey Chronicles. If the author uses filler to achieve the minimum word requirement, a moderator may edit the FanPost for his or her own amusement.

0 recs | Comment 15 comments

Story-email Email Printer Print

Comments

I thought this was gonna be a fanshot. ok fanpost it is.

go rowand

by lincypoo i wuv u on Jul 28, 2009 2:17 AM PDT reply actions 0 recs

I know John Hollinger at ESPN has his PER metric system, but his formulas predict some crazy thing sometimes. Overall, you can use sabermetrics to evaluate player value, but there are so many intangibles in basketball that really aren’t there in basketball that they don’t hold nearly the same accuracy. For instance, team chemistry is infinitely more important in basketball then baseball.

Bruce Bochy would like you to look at the career numbers and stop complaining.

by cheno on Jul 28, 2009 3:52 AM PDT reply actions 0 recs

By the way, I’ve heard that most of the big-time basketball statisticians have already been picked up as consultants by the teams.

Bruce Bochy would like you to look at the career numbers and stop complaining.

by cheno on Jul 28, 2009 3:53 AM PDT up reply actions 0 recs

From what I’ve looked into, a lot of the stats for higher analysis aren’t easy to get a hold of, compared to baseball.
A lot of it is based on possession by possession numbers and production given a specific lineup.

by DesertFox on Jul 28, 2009 10:05 AM PDT reply actions 0 recs

Check out APBRmetrics

There’s a bunch of links on that Wiki article that you might want to check out.

Bay Area fan going to UCLA.

by Yoyo on Jul 28, 2009 10:23 AM PDT reply actions 0 recs

ah nice, rss subscribed

go rowand

by lincypoo i wuv u on Jul 28, 2009 3:22 PM PDT up reply actions 0 recs

A couple of months ago, Michael Lewis did an interesting article in the NY Times (elitist!) Magazine about the new metrics in basketball, focusing on Shane Battier and the Houston Rockets. It is a good read, though like Moneyball, isn’t really heavy in the stats, just giving an introduction to what people are doing. Here is the link.

by Uribe nee Gonzalez on Jul 28, 2009 10:44 AM PDT reply actions 0 recs

Warning: It’s like 20 pages long. Good, though.

53-46 - Giants Series Record: 17-10-4, 13-2 at home

Adopted Giant: Daryl "Dealio" Maday - 3.66 ERA, .275 BAA, crappy 1.66 K/BB ratio in Norwich. OK.

by rhys on Jul 28, 2009 11:38 AM PDT up reply actions 0 recs

Oh yeah I read that! In related news Shane Battier is now one of my favorite players, rooted long and hard for them against the lakers. IIRC dude has funny low amount of self esteem.

go rowand

by lincypoo i wuv u on Jul 28, 2009 3:24 PM PDT up reply actions 0 recs

Basketballvalue.com
Basketballprospectus.com
82games.com
popcornmachine.net
storytellerscontracts.info

They’re out there, and they’re good.

Rufus on Fire -- SB Nation's Charlotte Bobcats Blog

by David Arnott on Jul 28, 2009 11:12 AM PDT reply actions 0 recs

I would go here

http://dberri.wordpress.com/

Uses a system called win score that isn’t too different from Hollinger’s, but seems to be way more accurate. Based largely around scoring efficiency and gaining posessions for your team ala rebounds, steals, etc.

What’s impressive is that the number of wins derived from the stats for each player often add up pretty closely to the number of wins the teams get. Seems like there’s some subjective massaging of the numbers that helps him get so close, but there’s no doubt he’s basically onto the “winning” factors.

Where he differs from a lot of the other stat stuff is that his system kills one-dimensional scorers that are sometimes the biggest stars. Iverson has been a below-average player throughout his career according to Berri, Stephen Jackson makes almost no positive contributions. On the other hand Biedrens slaughters in win score and was even a top 15 or so player two years back.

by Stoned Slacker on Jul 28, 2009 11:20 AM PDT reply actions 0 recs

I’ll say one piece about Berri’s work, though, and then stop, because I could go on all night, and this isn’t the forum, but… it’s easily torn apart. I read the book, and they’ve concentrated on basketball on their blog seemingly because their football/baseball work was so laughably awful in light of work done by more prominent folks already established. The general principle that the person who scores the points gets too much credit in the MSM is a good one, but after that, it’s one suspect mathematical assumption after another, leading to retroactively “accurate” measures that don’t predict anything and are useless on an individual player level because they split defensive responsibility purely by minutes. So it breaks at the extremes. For example, Bruce Bowen, just past his prime, was worse than useless according to their calculations, which is simply wrong.

Rufus on Fire -- SB Nation's Charlotte Bobcats Blog

by David Arnott on Jul 28, 2009 2:04 PM PDT up reply actions 0 recs

I’ve seen that and I’m not a fan
http://www.amazon.com/Wages-Wins-Taking-Measure-Modern/product-reviews/0804752877/ref=cm_cr_dp_all_summary?ie=UTF8&showViewpoints=1&sortBy=bySubmissionDateDescending

Also according to his ‘underrated players of the year’ Troy Murphy is 3x more valuble than Battier. I realize its hard to account for Battier’s ability to force bad shots but damn, Troy Murphy is not better than Rondo. With that aside, I do think some of these win values are more intuitive than other systems, Lebron being > 20 wins for example, though the calculations to getting that may be ‘getting lucky’.

go rowand

by lincypoo i wuv u on Jul 28, 2009 3:38 PM PDT up reply actions 0 recs

Interesting

I always had a ton of problems with his system, mainly the lack of meaningful stats for defense or team dynamics and drawing too many conclusions from it. And the way he insists on slotting players into positions kills some tweeners, guys he calls power forwards that aren’t rebounders, etc, but I figure that really helps him manage his results.

I had to comment on his post a while back about Chauncey Billups, how he tried to explain a rapid 50% decrease in Chauncey Billups’ output without attributing anything to the change in team, role, and style of play.

But its always been very good with the end numbers, so I had to give him his proper credit…..if that’s all suspect too then he’s not looking good.

by Stoned Slacker on Jul 28, 2009 3:21 PM PDT reply actions 0 recs

As people have mentioned, the biggest problem with basketballs sabermetrics is that the advanced stat data is not available to the public. Most of the stuff you find around the web is based off box-score statistics and the play by play. The real in-depth analysis done by teams is rarely disclosed, and comes from their advanced data collection methods where they record information like how open a shooter is, etc.

by snafu on Jul 28, 2009 10:34 PM PDT reply actions 0 recs