I could see this. I felt that last year's end-of-year roster was one that gelled and worked better than the one from the beginning of the year yet overachieved and won many games tha tcould have easily gone the other way. (The C's had something like 6 game-winning, buzzer-beater shots!)
I believe mean regression will not make the torrid pace from the end of last year sustainable, but a playoff roster with returning players and slight improvements will have a higher "median" win total. So the team might finish with 40, but less of them will be "flukey" ones.