In my own opinion, it matters a bit. I don't think the fatigue penalty on player performance is near what it was in previous versions of the game; a player at 85% in v3.1 is playing better than he would in v2. That said, I've heard other coaches say they don't think it matters at all in v3.1, or virtually not at all. And I've seen some coaches--winning coaches--run their key players down as low as 70% and well into the red. But I've also watched those same teams appear to fade a bit in the second half against teams of equal or near equal talent.
Regardless, your player's production isn't simply an internal factor; it's relational to the performance of the players for the other team. That's why I set my substitution values according to when I think the next guy on my depth chart is likely to match up best (or least bad) against the other team's player(s). My PhD in statistics and actuarial sciences has allowed me to devise a calculation of this that is precise to the sixth sigma -- okay, that's a lie. I basically drink a couple six beers and make a decision based on fake news and the general orientation of the seat of my pants. But I do try to at least guess in the right direction.
While I know all this didn't provide a direct answer to your question or provide concrete actionable advice in any way shape or form, I do hope it helped.
9/16/2017 1:21 PM (edited)