ElGee wrote:90sAllDecade wrote:That Elgee post was interesting but I have one concern about the main underpinning of the whole post.
Would'nt using a Drtg number (which is useful in comparing a team within a season, not cross seasons imo among other things) which fluctuates with league average defensive ratings make faster paced league Drtg 80's players look worse than slower paced mid 2000's lower league average Drtg players?
It would naturally skew the sample for mid 2000's players having more games and therefore better balanced averages against lower Drtg numbers (like below 103).
A comparison using relative Drtg competition or something like that might be stronger perhaps.
A second study in how each player succeeds defensively against higher relative Ortg competition might be interesting as well, to balance out comparing combined two way impact.
(I may attempt it if I have that much time available)
You can see the relative DRtg in the tables. I started the run with relative DRtg's but I think they introduce more issues. You have a defensive environment based on how hard points are to come by -- that seems like a more normalized environment to me than relative DRtg. How would it be otherwise?
If you use relative DRtg, you've got David Robinson playing against dams who give up 54% TS on average compared to a guy whose opponent gives up 50% on average...
I'm not 100% as a better choice as I'm not an expert and I appreciate the work involved, I guess it just seems uneven when comparing Drtg across seasons imo.
The defensive environment isn't the same decade to decade or season to season and some players could only play so many sub 103 Drtg teams in their season versus others who could have better balanced averages playing more.
The sample size against sub 103 defenses increases during the mid 2000's versus players like Jordan, Olajuwon and Malone who played in the 80's while the others didn't.
Rule and pace changes also affect league average Ortg and Drtgs.
If a player in the 80's played a #1 ranked defense in a league that had rules/pace conducive to a higher league Drtg it would look like they played a worse defense more times than a mid 00's player who played against a #15 ranked defense with rules/pace that encouraged a very low league Drtg.
Perhaps it was a tougher environment but it also doesn't factor that those same 80's players would be older by the time those lower league Drtg averages rolled around in the mid 90's or mid 2000's in the comparisons altering the sample.
If those players were in their prime during that era it would boost their stats against a larger higher Drtg sample and hurt them even more in the comparison as they age finally playing sub 103 teams more often and declined.
I wonder if some type of per100 comparison would help too? Again, I'm not sure and nothing alternative would be perfect, it just doesn't seem even imo.