Yeah, I wish there was a fast way to weigh mpg that was traded/kept. I'm not sure I can used Wins Produced accurately enough, or weigh it properly, and I take WP with a grain of salt because I don't think it can account for player synergies. I guess if I have more time I might dig deeper into the numbers (say, pre-ASB versus post ASB point differentials to take into account trades) but that would take more time than I'd have on hand atm.
I thought it would be interesting to look at roster turnover over multiple seasons. I used the data at dougstats.com for this. It may not be perfect due to the player name changes from one season to the next, but I think it should be good enough. I came up with two different stats than what's described in the OP. RSI, or roster stability index, is the percentage of players who've played at least 20% of the available minutes in the current season who also played on the team in the prior season. I chose that 20% cutoff because I didn't feel a player who's hardly in the rotation in the current season should count as a "returning player". MSI, or minutes stability index, is the percentage of the player minutes allocations in the current season that was accounted for in the prior season. Here are the results: As one would expect, the perennially good teams have had the most stable rotations, while the perennially bad teams have had very unstable rotations. I highlighted in green the most stable team-seasons, and in red the least stable team-seasons by both metrics.
