average age for teams

Post #1 » by **C.lupus** » Thu Jan 6, 2011 4:51 pm

Does anybody know if there is an online database or website that lists the average player age for NBA teams for all years (or at least going back a while)? It seems that, with all the sports stats out there this would be easy enough to find but I'm stumped.

What I want to eventually do is a regression analysis to see the correlation between team age and winning percentage.

thanks.

Manuel Calavera · Post #2 » by **Manuel Calavera** » Fri Jan 7, 2011 1:45 pm

I'm pretty sure that's already been done (maybe by BBR?). I don't think they list average ages though but I could be wrong. It would be really easy to add though, maybe you can request it.

droponov · Post #3 » by **droponov** » Fri Jan 7, 2011 3:30 pm

I remember Kevin Pelton doing something like this but for the correlation between age and defensive efficiency.

No idea where you can get that data. But I wouldn't use simple average age. I'd have more interested in the correlation between wins and (1) average age weighted for minutes or, ideally, (2) average NBA experience weighted for minutes.

droponov · Post #4 » by **droponov** » Fri Jan 7, 2011 3:51 pm

I found Kevin Pelton's work I was thinking about:

http://www.blazersedge.com/2009/3/2/773 ... ws-minds-a

but it turns out he actually did it for average age too:

http://www.basketballprospectus.com/art ... icleid=571

I also got this from BlazersEdge - author PoliSam. He actually calculated the average age/experience weighted for floor minutes but also for FGA (a decent estimation of a player's importance in the team):

http://www.blazersedge.com/2008/12/9/68 ... in-the-nba (part I)
http://www.blazersedge.com/2008/12/12/6 ... and-win-in (part II)

Here's the relevant chart:

He also did for experience (in this chart measured in games played):

The relationship between success and age is pretty strong indeed.

Post #5 » by **C.lupus** » Fri Jan 7, 2011 3:54 pm

Sweet! Thanks droponov.

Chicago76 · Post #6 » by **Chicago76** » Sat Jan 8, 2011 5:42 am

Keep in mind that there is a survivor bias on that first graph that droponov presented.

He touches on it here, but he misses the larger point:

Now, all of this needs to be understood in context. The data shows a correlation between age and team success, but this does not mean the former leads to the latter. Going out and signing a 40-year-old free agent and putting him in the lineup won't improve your team. Much of the explanation for the relationship between age and success comes from self-selection--teams competing for a championship will fill out their bench with veterans looking for a ring, while younger teams are giving heavy minutes to youngsters in the name of player development. The strength of the team often determines its age, not the other way around.

The veteran point raised only shifts the needle a little bit. If your team avg age is 29, and you go and pick up a 36 year old reserve shooting specialist/rebounding specialist, team age will only jump about a half year. The bigger issue is that teams don't usually get old together unless they're good. In other words, if your 29 year old on average franchise went out and won 58 games last year, there is a good chance you'll ride them out another year. If a team of the same age finished .500 and barely made/missed the playoffs, there is a good chance you'll blow the team up and rebuild through youth. Bad teams usually get younger while good teams are afforded the time to grow older.

EDIT: that might be an interesting study. A team with an avg age of X in yr 0. Based upon floor time from the year 0, what percent of floor time the next year is just organic 1 yr aging from retained players, what is the net change from dismissed vs. FAs/traded players the next, and what is the net change from dismissed vs. drafted players, and what is the resulting delta in record from yr 0 to yr 1? It would be a ridiculous amount of work to get a decent sample size though.