Wel, yeah, the 100 points rule sucks and I won't use it anymore, even though it could have been more interesting with less 80/90% votes.
As for the groups, of course I'm going t disclose it now !
The groups formation is based on the results of the results of the Special round (the one when you had to give your top 5). But to rank each election, I used, instead of approval voting (each election in the top 5 gets one point), the Borda method (ie giving 5 points to an election when it's in the first place, 4 points when it's in the second, etc...). The results were :
1968 : 21
1912 : 20
1960 : 18
1932 : 16
1980 : 12
1948 : 12
1824 : 8
1896 : 7
2008 : 7
1992 : 6
1796 : 5
1976 : 3
Based on these results, I created four pots of three elections each.
Pot 1 (favorite elections)
1968
1912
1932
Pot 2 (apreciated elections)
1948
1980
1824
Pot 3 (poorly apreciated elections)
1896
2008
1992
Pot 4 (least favorite elections)
1796
1976
1884
Each group had to consist in one election from each pot, in oder to make it fairly balanced. So I made those groups :
Group 1
1968 (from pot 1)
1980 (from pot 2)
1992 (from pot 3)
1976 (from pot 4)
Group 2
1912 (from pot 1)
1948 (from pot 2)
2008 (from pot 3)
1796 (from pot 4)
Group 3
1932 (from pot 1)
1824 (from pot 2)
1896 (from pot 3)
1884 (from pot 4)
My main criterion to put determinate elections from different pots in the same group was mainly chronological coherence. Thus, I grouped most of recent-times elections in the group 1 (having to choose, from pot 3, between 1992 and 2008, I chose 1992 because it was closer to 1980 and the others), the oldest in group 3 (I could have put 1912 instead of 1932, but, to be honest, my goal was to get both 1932 and 1948 qualified for the top 5 - as a result they were both eleiminated
), and the remaining in the group 2.