Okay, here's a broadened picture of the +/- statistics I was presenting earlier. I took the top 10 most effective 5 man rotations for the Jazz, and here's how many times each players appeared in those 10 line-ups. The sample size is still relatively small, but between the various line-ups, each player with at least 3 appearances logs more than 50 minutes total, others substantially more.
Gordon Hayward: 7
Derrick Favors: 6
Jamaal Tinsley: 6
Enes Kanter: 5
Marvin Williams: 5
DeMarre Carroll: 5
Paul Millsap: 4
Al Jefferson: 3
Mo Williams: 3
Randy Foye: 3
Earl Watson: 1
Jeremy Evans: 1
Alec Burks: 0
What I've found getting into this stuff is that Foye and Haywad can't play together. Haywad is a great 2 and so is Foye (in the right situation), but playing two 2's together is too easy to exploit from a size standpoint. Haywad might get good enough to guard bigger threes, I dunno. Foye isn't big enough to guard 2's, even though he works his *** off trying. Add an undersized 2 and an undersized 3 and you're asking for trouble. Any other combination of Foye or Haywad with Marvin or Carroll has been pretty damn effective. This is why playing Haywad off the bench has worked so well.
The only other combo that has been noticeably & statistically poor is the Millsap as a three experiment.
.
The easiest way to interpret this is Jazz interior defense is only bad when the exterior defense is before.