The problem is that the statistical algorithm used by PB is BAD!!!!! There is no good explanation for the results.
I have been involved in statistics for many years. Even wrote a baseball simulation program back in 1979. You shouldn't need 20 or 30 leagues to have it balance out, not if you're using a model that is actually based upon the real-life statistics.
I had a problem last year with my pitchers getting hammered so I changed ballparks (to SF) to get one that is the most pitcher friendly available. In my first series pitching WMiller, Milton, Dempster, JAnderson, and Tapani I got touched up for 19 homeruns and a team era of 9.00!!! After 10 games my team ERA is 8.08 and my above average pitching staff has given up 28 homeruns, an average of almost 3 per game!!!
I've been with this organization for over 10 years and have many hours invested. I am so close to just throwing in the towel. Last year I had Eric Milton throw about a 100 homerun balls and not have but 2 quality starts all season!! I got tired of the excuses, "well lefthanded pitchers don't fair as well in baseball simulations such as PB".
This is such a wonderful concept, I just wish they could have gotten it right!!!!!!
If they want it to be based on actual stats then do it!!!
If they want it to be completely open-ended then do that.
BUT YOU CAN"T HAVE BOTH!!!