Page 1 of 1

The notion of fit and value

Posted: Mon Nov 17, 2014 10:13 pm
by Sam Garriock
The idea of "skills" instead of positions is something that I really love.

In my opinion, basketball is more or less played by a primary creator, secondary creator, wing shooter/defender, stretch big and pick and roll dive man/rim protector. This explains the diminishing returns of throwing a ton of LeBron James' together - one of them is going to be "relegated" into the wing shooter/defender role and thus his value is severely diminished, all the while being the same player.

It is something that I have argued heavily with Cavs fans and it is why the value of having Dion Waiters out there when LeBron James and Kyrie Irving are already on the floor is severely diminished - in this role, Matthew Dellavedova is probably more valuable than Dion Waiters, which is just an insanely perplexing notion when watching them play; Delly can't dribble a ball without the world imploding.

This is something that needs to be taken into account with respect to RPM and those that use it - a player's RPM only really has any value if they are compared to another player who is playing that same role, not someone playing the same position.

Obviously this is nothing groundbreaking, but after listening to Seth Partnow and Ian Levy on the r-squared podcast (http://nyloncalculus.com/2014/11/14/the ... h-partnow/) I thought this may be a useful place for discussion on such a matter.

Re: The notion of fit and value

Posted: Mon Nov 17, 2014 11:37 pm
by Crow
I'd suggest reading the threads of David Sparks, listed in list of threads worth reading, player typology section. I'd like to redo with advanced boxscore stats and / or rpm but haven't made the time yet.

Re: The notion of fit and value

Posted: Mon Nov 17, 2014 11:51 pm
by Crow
Player quality vs quality of situational fit is a topic that has been touched on recently on twitter. I was tempted to start looking at bpm / rpm ratios at player, lineup and team levels but DSMok1 has cautioned about using bpm above player level and noted that it is not regressed to mean like rpm. Still that would be one crude way to start comparing player quality to situational fit quality.

Re: The notion of fit and value

Posted: Tue Nov 18, 2014 2:28 am
by Statman
Crow wrote:Player quality vs quality of situational fit is a topic that has been touched on recently on twitter. I was tempted to start looking at bpm / rpm ratios at player, lineup and team levels but DSMok1 has cautioned about using bpm above player level and noted that it is not regressed to mean like rpm. Still that would be one crude way to start comparing player quality to situational fit quality.
My post in Crow's lineup thread pretty much relates to this, and how I'd approach it if I could.

http://www.apbr.org/metrics/viewtopic.p ... 49b#p21462

Re: The notion of fit and value

Posted: Tue Nov 18, 2014 6:22 pm
by knarsu3
I looked at this on the defensive side of the ball: http://www.vantagesports.com/#story/VEB ... dard-terms

And I did find effects of optimizing the lineups based on these positions.

Re: The notion of fit and value

Posted: Tue Nov 18, 2014 6:44 pm
by Crow
is there a part 2 coming focused on types in lineups?

Re: The notion of fit and value

Posted: Tue Nov 18, 2014 9:17 pm
by knarsu3
Crow wrote:is there a part 2 coming focused on types in lineups?
Yeah there is. Don't have a date yet though.