Dudley has been looking positively statuesque on the court. Little guards rebounding over his head. And his 3pt% has declined for the 4 straight seasons. Now at .351.
If a "lockerroom statesman" type is needed, I think there are better options than Dudley. Guys who can play that role while also being a solid 10th man or something.
Dudley can be a solid 10th man. For the season the Nets were better with him on the court, and in the playoffs that was amplified: the Nets actually outscored the Sixers when Dudley was in the game, and he had an unreal +21.2 point differential per 100 possessions when in the game versus on the bench.
He can contribute both on the court and in the locker room.
On/off numbers can be useful with very large samples of 10 seasons, ideally with players playing a lot of different line-ups. For a single season, they have no statistical value due to the multicollinearity problem - those numbers depend on the actions of Dudley and other 9 guys, not just Dudley. That's why the list of players by net rating as a bunch of guys from Milwaukee, Golden State and Toronto in there: 14 out of the first 15 among players with GP>50 belong to those 3 teams.
IOW, the Nets point differential was higher with Dudley on the court is completely different than the Nets point differential was higher because Dudley was on the court.
There are some attempts to sort out those multicollinearity problems, using boxscore numbers and other data as bayesian priors and ridge regressions, but none of them is any good IMO.
Dudley is painfully slow. He's the kind of guy that could still be useful maybe 4 or 5 years ago, but in today's switch heavy NBA, with lots of guys who can shoot off the dribble, will just be mercilessly hunted. Nowadays defenses are defined by their worse players more than their best players - not having a sieve is more important than having a Bruce Bowen.