LunaticWolf
Well-Known Member
Like I said, I understand what I'm looking at.
What you see though is striking in terms of both 1st unit vs. 2nd unit (since that's the real grouping trend) and in terms of the near mirror of the Williams vs. Watson defensive margins.
The unadjusted numbers speak to a couple of issues:
1) One of the reasons you adjust is to prevent overstating the value of guys who have bad back-ups or punishing back-ups who happen to have superstars in front of them. For that reason we would expect an adjustment to decrease William's defensive numbers and increase Watson's defensive numbers. What we get is the opposite: the gap narrows because we pull Watson down and push Deron up. That's the reverse of our expectations about these players based on reputation.
2) Basketballvalue breaks it out by defensive and offensive performance, as opposed to the 82games adjusted ratings which group the two components into a net effect. That obscures what's going on with Williams: that he is having a horrendous defensive season while simultaneously being the key to the offense. That actually puts him much closer to what we think of Monta Ellis (although he's obviously more efficient than Monta, I'm talking about it in terms of his "one-dimensionality".)
3) Another reason we adjust is to make sure that a player doesn't have his stats buoyed just because he plays with good teammates. I feel confident in stating that Earl's not getting that much extra benefit from playing with our bench guys.
I recognize that we're seeing the Earl stats with a large sample size against back-up PGs but I think 1 and 3 swamp those some as well as the use of Earl and Deron together in 2 PG backcourts as Sloan has done at various points this season. Obviously, there's no significant statistical difference between their performance in the 1 year adjusted ratings either and that's almost entirely because of Deron's horrendous defensive line. In fact, in one year adjusted +/- Earl Watson is the top player in the entire league (although the standard error issue raises its ugly head, as it always does with +/-)
https://basketballvalue.com/topplayers.php?year=2010-2011&mode=summary&sortnumber=94&sortorder=DESC
So the "adjustment" issue isn't affecting my primary reason for making the thread.
Why don't you go take a look at Carlos Boozer's +/- numbers? Amazing they look MUCH better then they ever were with the Jazz. Is he playing better defense? A better team game? Or are they higher because he doesn't have a starter backing him up?
+/- has everything to do with the players you play with, rotations, who replaces who, etc. You cannot extrapolate it to a single player without adjustments, as there are always 4 other players on the court. Works great for talking about lineups, but that is about it.
In 2008-09, Deron was a -9 and Knight was a +13. Was Knight a stellar defender?
Deron plays against lineups consisting of 2-3 stars for nearly the whole game. Watson and Price play against lineups consisting of 2-3 D-league scrubs for their minutes on the court. Which lineup do you think is going to have a better offensive output?
Additionally, you are comparing 1700 minutes by Deron to 700 minutes by Watson. You can also probably take away 200 minutes or more for the time they have been on the court together. Watson's sample size is simply too small to compare to Deron accurately.
In 09-10, Watson was a -1.5 in Indiana, which he totaled similar minutes to Deron. All you are seeing here is a contained sample. If Deron remains out for a while, watch Watsons defensive numbers plummet as he plays more minutes and against better lineups.
Production > On/off any day of the week.
Last edited: