I am with houston. He does not have a problem understanding you. But, the divisional reshuffling would be more equal if you mixed up the rankings instead of always putting the 3rd best team in Division C. Which is unarguably (sorry this is just math, no feelings involved) a weaker collection of teams when compared to Division A.
This would be a more equivalent way to mix the divisions: A (1,5,8,12) B (2,4,9,11) C (3,6,7,10)
To explain: Once you have your list of teams ranked 1 through 12 (based on a 3 yr cumulative regular season record), lets break those 12 teams into 4 groups. Group W : [ 1,2,3] Group X : [ 4,5,6] Group Y: [7,8,9] and Group Z: [10,11,12] Now, lets assume that each of these three groups represents a group that is statistically proven to be similar in performance, that the team with the lowest number in each group is also the team from each group that is the best performer, and furthermore that the entire purpose of this exercise is to balance the power of the league across each division as much as possible. If we take 1,4,7, and 10 and put them into the same division ( as your method does) then you are taking the most powerful member of each group and forming a division of the most powerful teams from their respective group. If you wanted to optimally balance the power you would distribute the members from each group into the divisions in an alternating pattern to ensure that no division represents more top performers or lower performers from across the different groups. How do we do this?
Make sure that each division has an average rank of the 2nd ranked team from each group:
A( W1 X2 Y2 Z3 )
B( W2 X1 Y3 Z2 )
C( W3 X3 Y1 Z1 )