Margarita,

Thanks for letting me know that my name had come up in a discussion of the Council. I have a couple of comments, that you may forward to the Council if you think it is appropriate.

(As Patricio and maybe others will recall I was instrumental in the decision of the original ccTLD constituency to use STV in multi-member elections, which was dropped when ICANN V2.0 was rebooted)

Nigel is right. The ccTLD Constituency did adopt the STV method for elections with many candidates, after he convinced us it was a good idea.

About this particular election, it seems that Pablo would have won in any case, but still, it may be good to make the rules more explicit for the future. Joke says:

As agreed the candidate with highest level of preference is considered to be selected.

This is OK, as long as everybody agrees about how the "level of preference" is to be computed. However, I see in the conversations that TWO methods are mentioned:

Method 1: Assign scores (1=best preference, 2=second best, ..., 5=not preferred at all), add up all scores and the winner is the candidate with the lowest score.

Method 2: Count the number of first preference votes.

In this case, Pablo wins under both methods, but it worries me that the Secretariat seems to believe that Method 1 and 2 are always equivalent:

Pablo Rodriguez still has the lowest total score, in other words: the largest number of first preference votes.

But the two methods do not necessarily give the same result, as the following small example (with three candidates A, B, C) shows:

        A  B  C

voter1  1  2  3

voter2  1  2  3

voter3  3  2  1

voter4  3  2  1

voter5  2  1  3

Score: 10  9 11

So, with Method 1 (lowest score) the winner would be B, but B would lose with Method 2, because only one voter ranked him first.

Therefore, I think it would be important for the ccNSO to adopt more explicit rules for future elections.

Patricio