Fun fact

[Krodis]

Coming into the Kings game (I don't suspect it changed much) our best lineup of at least 100 minutes together by differential (per 100 possessions) has actually been:

Maynor
Cook
Harden
Collison
Mohammed

Have I mentioned I really like the way our bench plays?

[Krodis]

Although really, pretty much every lineup we throw out there with the players we currently have. (The only lineup that's played any significant minutes with a negative differential is Westbrook/Thabo/Durant/Ibaka/Collison, although it does have a positive adjusted +/-.

[slick_watts]

Adjusted +/- is too noisy with these small samples, but you are right that our second unit tends to blow away other second units. We have two starting quality players in there (Nazr, Harden), and good role players. They also play very well together.

[Krodis]

I do agree that +/- is too noisy with small samples. I do tend to trust it more with entire lineups as opposed to individual players, although my trust might very well be misplaced (and admittedly, the sample is smaller with entire lineups vs individual players). I actually used the raw differential in the original post, that lineup drops to fourth when adjusted, still pretty good.

Our bench has some good players (Harden, Collison, Nazr), but also plays well together. Maynor distributes, Harden carries the load, Cook finds open threes. Collison and Nazr do their thing. Harden and Collison play really well together. So do Harden and Cook. And Collison and Nazr.

Our bench actually features what have been our two best players for 1 year adjusted +/- (way too noisy to trust imo, though) in Collison and Maynor. Even with two year adjusted +/-(less noisy, but still pretty noisy) they're still second and third (Durant jumps to first). I don't really love the reliability of +/- though, so I won't read too far into it.