APBRmetrics

drebelx · Posted: Thu Mar 10, 2005 2:03 pm Post subject:

FrontRange,

Assists aren't an accurately measured. There are numerous threads in the internet discussing the scoring official's inaccuracies on when to score an assist. I'm sure many here know about this.

Knight this year is doing 0.291 assists per minute on 28.5 minutes per game. For his career before this season he was doing 0.247 assists per minute on 25.2 minutes per game. Run the percentage and that is a 15.3% increase in the production of assists for Knight over his career numbers. I'm thinking 15.3% is better than slightly better.

Haha. Mt Mutombo doesn't assists because he doesn't have the skills. No argument there. I dunno Mutombo at all since I just jumped into basketball, sounds like he would be a burden on a team from your description.

Dan Rosenbaum · Posted: Thu Mar 10, 2005 2:10 pm Post subject:

drebelx · Posted: Thu Mar 10, 2005 2:25 pm Post subject:

Dan,

Yeah, when I think about it, I am in a very odd situation. I'm not sure of which forums I read that on and it may have been on this one. This was probably a few weeks ago that I had a discussion on assists with my pals on the fantasy league, so I did some quick google search reading on Assists. I have heard of Mutumbo, but I don't really know of his situation, skill set, etc. I've only jumped into basketball since around August when I had to develope a draft sheet for my fantasy league. That is how I ran into your "Measuring How NBA Players Help Their Teams Win" site. As the season wore on, I developed some ideas about basketball that I wonder if anyone else has concidered which has ultimately led me here to the premiere locaition on the internet of basketball thinking.

I just hope my ideas don't get tossed out so quickly because of my inexperience with basketball.

Dan Rosenbaum · Posted: Thu Mar 10, 2005 2:57 pm Post subject:

S.K. · Joined: 18 Feb 2005 Posts: 52 Location: Toronto

S.K. · Joined: 18 Feb 2005 Posts: 52 Location: Toronto

FrontRange · Joined: 27 Jan 2005 Posts: 124

Just to be clear . .assists and turnovers are a poor proxy for passing and creating skills. They don't tell the whole story but we don't have alot more detailed information. In a similiar vein ERA tells something about a pitcher even if it captures alot of noise (defense, ball parks, skills of oposing teams, etc.)

If baseball didn't track OPS and slugging, RBI could be a poor man's proxy for hitting. If your point is we need better passing stats, I think the entire forum would agree (although I personally believe a greater understanding of how to measure defensive ability is larger need), but I think you can see why RBI are not that good a comparison to assists. Without assits and turnovers, what would you look at to determine statistically who are good ball handlers / passers?

replayhoops · Joined: 18 Feb 2005 Posts: 10 Location: Connecticut

Some of the earlier posts have alluded to the different degree of passes that can lead to baskets. It would be great if there were available stats breaking out assists into "incidental" and "exceptional". Then we could better quantify the truly great passers than from those who are just lucky enough to be on the same team as great offensive players.

drebelx · Posted: Thu Mar 10, 2005 5:29 pm Post subject:

Dan,

I know what you are saying about needing to know more about the game. I think it is good that I'm forming my own opinions and thoughts before really getting into it. This way I can for realize for myself if my thoughts match up with better with the basketball reality as opposed to what someone telling what it is supposed to be. Lots of new and innovating things have been realized in this manner.

I'm not so sure about the deep understanding of the game part. Anybody can develop a superficial knowledge about any sport and understand what the stats are saying. If that was not true the idea of sports as a passtime would fail.

My understanding of the two spheres of stat people and baseball people are that they are merging. Example, Boston Red Sox. The two "spheres" work very close with one another under the guide of Theo Epstien.

Not enough numbers being collected in basketball is annoying! When are new things going to be counted, or at least issued for public consumption?

FrontRange:

drebelx · Posted: Fri Mar 11, 2005 8:10 am Post subject:

that last silly little equation won't work. It should be more like (passes completed) / (passes attempted) and it's not really equivalent to OBP, but more like Batting Average. OPS would be more like something that explains Ball Retention, the ability of keeping the ball on your team and not making turn overs. So that would end up being something like (possession - TO) / (possessions). Something like that.

kjb · Joined: 03 Jan 2005 Posts: 671 Location: Washington, DC

Passing turnovers gives us information on failed passes. The data exists in the play-by-plays, and is summarized on the player pages at 82games. I'm dubious about whether a completion percentage would be of much value for the NBA.

Something like (ast - passing TO) / passes attemtped might be useful. That might start getting at whether players are using their passes to set up teammates or whether they're making "safe" passes that might not be doing much to help the team score.
_________________
My blog

drebelx · Posted: Sat Mar 12, 2005 11:43 pm Post subject:

hmm. So I'm guessing the consensus here is that Assists aren't like RBI's and that Assists are not concidered glorified passes. Also, there there is not much basketball can learn from baseball since they are different sports. Interesting...

Dan Rosenbaum · Posted: Sun Mar 13, 2005 12:13 am Post subject:

drebelx · Posted: Fri Mar 18, 2005 9:20 am Post subject:

It still feels like there hasn't been a real crazy breakout idea. Maybe I missed one or just don't really understand. In baseball, sabermetrics brought the ideas that bunting and "small ball" was generally a waste and it is still a controversial topic. To find a good pitcher, DIPS was created that removed the defense behind him because it was realized some numbers that were credited to pitchers were partially out of their control, like ERA and even Hits. So yeah, new solutions are needed. I like the thread that is currently going on about Assist% and creating your own shot. Interesting stuff.

HoopStudies · Posted: Fri Mar 18, 2005 12:17 pm Post subject: Things to Study