The following shows the percentage of starts that has gone to the top 4 starters of each team, by year.
The peak post-1920 occurs in 1974, with a rate of 79%. If you had 5 full-time and healthy starters, you'd expect 4 of them to get 80% of the starts. The median from 1901 to 2002 is 73%. In the last 5 years, the average was 72%.
Ok, think about it for a minute. In the 10-year span from 1993 to 2002, the rate is for 73% of the starts to go to your top 4 starters. From 1968 to 1977, the modern peak, 77% of the starts went to the top 4 starters. That difference, 4% of 162 games, comes out to 6 starts, somehow spread out among the 4. You can give all those to your top starter, so your current day 34-starter, becomes a latter day 40-starter. Or you start spreading it among the 4 starters (on a maybe 3,2,1,0 basis). Assuming that the win% of your top 5 starters is .60, .54, .50, .46, .42, we are talking about finding at most 1 win (shifting the 6 games from the .42 winner to the .60 winner). But if every team does this, the advantage is now lost. 1 win is large enough to consider. And who knows if the extra pitching will hurt/help the pitcher. These are all considerations for management and the fans to ponder.
For seasons with 16 teams, I selected the 64 pitchers with the most starts. For seasons with 30 teams, I selected the 120 pitchers with the most starts, etc. The rates shown below is the total of these starters, compared to all starts made for that year.