I agree it is better for the pitcher, but I do think, in the single game that someone is managing, having a fresh pitcher every 2-3 batters yields fewer runs. Bringing in guys for 1 inning or less and being able to tell them "dont save anything on every pitch" actually yields better (provided you have a supply of pitchers that can do that). So, I get why managers use this model when they can, short term gain rather than long term gain.
It is odd, because Baseball is the ultimate "care about the long term" sport since it is a very long season, counting in playoffs it is around 180 games to win the WS (and more if you count spring training).
It would be an interesting study, I wonder which is correct? It is very hard thing to create a valid experiment for. So many of the fans on this board are going to say "fewer pitching changes", but that is partially because we dont have a very young board and to some degree we are looking back to the game we watched as kids, younger adults. If I ask my 18 year old niece, she may be more likely to say "wow, 102 MPH, you cant hit that no matter what, why would you not try to keep that level of pitching up if you could".