What I've tried to do is draft 5 starters who are relatively equal in value so that there's little drop off from Starter 1 to starter 2, 3, and so on. What seems to happen is that at least one of my starters will crash and burn while the other guys do basically what I expected.
I usually shoot for 900-1100 IP from my starters. If I want a 5 man staff, I look for 190-220 IPs per guy.
In other words, if I want my starters to say average 1.27 WHIP#, 5.75 K/9#, 3.3 BB/9#, and 0.7 HR/9, I'll draft 5 starters to hit those marks. But what seems to happen is that at least one of these guys will completely tank. The tanker could be a y or a righty.
Inevitably, I'll put the tanker into the pen as my Long A guy. So instead of 1000 or so innings from my rotation, now I'm down to 800. I'll keep the pitch count the same for my 4 starters as I did for my original 5. Whenever I've had to do that, my team immediately seems to play much better.
My most recent team where I did this, my guys started 4-12. After the switch from a 5 man rotation to a 4 man rotation, they have gone 14-5. I know that that isn't a big sample size, but I swear, that phenomenon has taken place for me before.
I'm sorry that I don't have more specific data for you guys to chew on. I guess what I'm asking is that is there some kind of a "glitch" in the SIM where it favors shorter rotations.