Suns News: Week 15 2/2-2/8

[carey]

TJ's 40:

Seems to me we have a lot more offensive than defensive talent on the roster. I don't see how this guy fits with the Morris twins.

If we could get a Montrezl Harrell to pair with Warren we wouldn't need the MoBros

He should be there at 14.

[INFORMER]

TJ's 40:

Seems to me we have a lot more offensive than defensive talent on the roster. I don't see how this guy fits with the Morris twins.

If we could get a Montrezl Harrell to pair with Warren we wouldn't need the MoBros

We've never needed the Morris Twins.

[SDC]

question: why isnt gerald green get to be part of the group that gets to take the last second shot in a one or two pt close game? i'd rather have him in there than bledsoe and pj tucker or even dragic.

[Cap]

I'm working on a mathematical model to try to estimate the impact of individual players on +/-. It's based on the 5-man-unit data at 82games.com, so it's based on a player's +/- and who they play with (but not who they play against).

So far, it says that the most valuable Sun by a wide margin is Kieff at a crazy +.52 points per minute, with IT also impressing at +.21. Thomas, of course, benefits from playing against the other team's bench.

Our worst is Plumlee at -.17, just behind Bledsoe at -.14. (Goodwin is far worse than either at -.58, but I'm talking about rotation players.)

The Bledsoe number surprised me, but it does seem to fit the data. In 245 minutes without Kieff on the floor, Bledsoe has put up a -62. In 115 minutes with neither Kieff nor IT on the floor, he has put up a -52.

[SDC]

I'm working on a mathematical model to try to estimate the impact of individual players on +/-. It's based on the 5-man-unit data at 82games.com, so it's based on a player's +/- and who they play with (but not who they play against).

So far, it says that the most valuable Sun by a wide margin is Kieff at a crazy +.52 points per minute, with IT also impressing at +.21. Thomas, of course, benefits from playing against the other team's bench.

Our worst is Plumlee at -.17, just behind Bledsoe at -.14. (Goodwin is far worse than either at -.58, but I'm talking about rotation players.)

The Bledsoe number surprised me, but it does seem to fit the data. In 245 minutes without Kieff on the floor, Bledsoe has put up a -62. In 115 minutes with neither Kieff nor IT on the floor, he has put up a -52.

show me your math equations.

[SDC]

read somewhere that last year the offense efficiency was higher with dragic on the floor than bledsoe.

[Cap]

show me your math equations.

Maybe later. It's a work in progress.

FWIW, a parameter tweak to combat overfitting has resulted in more reasonable-seeming results:

Code: Select all

Markieff.Morris = +0.063
         Thomas = +0.032
        Bledsoe = +0.017
       G.Dragic = +0.013
  Marcus.Morris = +0.004
         Tucker = +0.004
         Wright = +0.003
            Len = +0.001
          Green = +0.001
          Ennis = +0.001
       Z.Dragic = -0.000
        Bullock = -0.002
       Tolliver = -0.005
       Randolph = -0.007
        Plumlee = -0.009
         Warren = -0.011
        Goodwin = -0.017

[Cap]

A note on how critical Kieff and IT are to our success:

• With Kieff and IT on the court, we are +139 in 302 minutes (+22.1/48)
• With Kieff on the court and no IT, we are +133 in 962 minutes (+6.6/48)
• With IT on the court and no Kieff, we are -12 in 459 minutes (-1.3/48)
• With neither player on the court, we are -75 in 201 minutes (-17.9/48)

A note on a couple of garbage-time players:

• Goodwin but no Ennis: -34 in 39 minutes (-41.8/48)
• Goodwin with Ennis: -6 in 40 minutes (-7.2/48)
• Ennis but no Goodwin: +7 in 11 minutes (+29.2/48)

Our favorite whipping boy, Anthony Tolliver, does surprisingly well in the model, despite having quite a negative +/-.

Tolliver played in only one lineup with Kieff. Thomas-Tucker-Green-Tolliver-Markieff put up +19 in 8 minutes (+114.0/48).

When Thomas is on the floor without Markieff:

• w/ Tolliver: +0 in 101 minutes (+0.0/48)
• w/o Tolliver: -12 in 358 minutes (-1.6/48)

When Thomas and Markieff are both off the floor:

• w/ Tolliver: -23 in 103 (-10.7/48)
• w/o Tolliver: -52 in 98 (-25.5/48)

So his +/- is crappy, but less crappy than comparable lineups without him.



Note that this is based on the 5-man unit data from 82games.com, which isn't totally complete. (We have, in fact, played more than 1924 minutes this season.)

I think that not having data on opposing lineups is a fatal flaw to this approach. I don't believe IT is our second-most valuable player; his apparent impact on +/- comes from being a starting-caliber player going up against other teams' benches.

[SDC]

utah does suns a favor again by beating NOHlins.

[INFORMER]

That's a bit of a surprise. The Pelicans were up by 15 at one point.