APBRmetrics

When a starting player is taken out of the game he is often taken out along with other starters. A team doing worse when a player is taken out could be partly because of the good players taken out with him. Does this make individual player plus minus ratings inaccurate? Harrison Barnes was a starter last year and his plus minus per 36 minutes was 0.8. This year Andre Iguodala came to the warriors and took the starting spot. Barnes comes off the bench now and plays with the starters less. He has been playing a little bit better than last year but his plus minus per 36 minutes is now -1.4.
Would plus minus ratings for each combination of five players when playing together be a better stat?
To find an individual players impact and compare him with other players you would have to compare the scoring differentials of the team with four players and player x and the same four players with a different player. For example to see the if Mareese Speights is better or worse than Andrew Bogut you would see the teams point differntial while playing Curry, Thompson, Igoudala, Lee, and Speights and compare it to the point differential for the team with the same lineup but with Bogut instead of Speights. It would be interesting to find the best five player combination point differential for teams and see if those are the five starters. do people agree?

Oh I didn't realize that adjusted plus minus was a thing but is there a type of plus minus for player combinations that accounts for the opposing team?

Yes adjusted plus minus is the general strategy for dealing with the differences in lineups played in and faced that affect raw +/- for individual players.

Yes there does exist the technique to calculate adjusted plus minus for 5 man lineups. JE here had shared such values in the past. I am not aware of any public source of this data for this season. It would be nice to see it again for 5 man lineups and pairs.