It's funny, because in the past, on/off was Tatum's calling card. The team was almost always awful with him off the court or when he missed the game. And it was quite the opposite with Jaylen where the team didn't seem to miss him whether or not he was there.
I agree that it is a factor and probably should prevent him from being the actual MVP, but there are other factors, too - namely who he plays with when he is with the bench and who the other starters play with when they are with the bench.
Glancing at the most used line-ups on NBA.com, the top two 5-man units without Tatum are the other 4 starters with either Horford or Hauser. The top two 5-man units with Tatum and without Brown all include Jrue, Pritchard, and Hauser - and then one of Horford or Kornet.
So I would conclude that the team is often quite a bit less talented when Tatum is with the 'bench players' than when Brown is out there without Tatum. The minutes samples are fairly low overall, though, so I'm sure there is a lot more nuance than what I'm describing. However, it does seem to follow the eye test when watching games - you see a lot of Brown, White, and KP out there without Tatum (our 2nd, 3rd, and 4th best players).
Yeah, the "eye test" is the thing that is hard to quantify. Everyone sees something different. I am with you regarding the rotations. And also on the past few seasons (which I know shouldn't matter for this season's MVP voting), but Tatum was always the one with the off the chart differentials while Shai was the best player on a lottery team.
So now the rotation is that when Tatum is resting, it is Brown as the primary scorer on the floor more often with 3 other starters and Tatum is deployed more often to kind of prop up the units with more bench players when Brown is resting. Brown gets White and Holiday and Porzingis when Tatum is resting, Tatum gets Pritchard and Hauser and Kornet when Brown is resting. Just to further back this up (2-man units minutes), Tatum is playing more with the bench and they play about the same with the other starters (Tatum more with Holiday, Brown more with White and Porzingis). Tatum has also played more than Brown with Brissett, Queta, and Tillman. So there is some tangible explanation to why Tatum's on-off differentials are less than past years. The coaching staff has made what appears to be adjustments to mitigate what had been a large differential in the past. In the past, it was Brown out there with the bench more and people were all over his on-off differential. This appears to be a tactical measure to sacrifice Tatum's individual performance to help the overall team.
Tatum + Pritchard/Hauser/Kornet 755 min / 743 min / 416 min
Brown + Pritchard/Hauser/Kornet 468 min / 575 min / 266 min
Tatum + White/Holiday/Porzingis 1654 min / 1489 min / 1065 min
Brown + White/Holiday/Porzingis 1345 min / 1588 min / 1121 min
The most used non-SGA line up for OKC is J. Giddey - A. Wiggins - C. Holmgren - J. Williams - C. Wallace. For some reason, they don't play Dort when SGA is off the court but he is part of the top 3 most used line ups with SGA. So it isn't just Wiggins in for SGA, it is also Wallace in for Dort. Rotations and other factors matter when looking at these on-off differentials. For example, SGA's DRtg is helped because he is on the court a lot with Dort. When SGA is off the court, for some reason, Dort is off the court also.
Maybe without SGA and Dort, OKC is a 35 win team, but I am not sure how you can state that as fact. Too many variables. It looks like PHI is a 45 win team without Embiid, that you can state as fact. I am not so sure OKC is as bad as a 35 win team without SGA. But they have been a lottery team with him. I think the main point is that over-relying on stats may not capture the whole picture.