Dale Sheldon writes:
> This is why standard ranked-pairs is to have all unranked candidates lose
> to any ranked candidates on a ballot; honestly, that's probably how voting
> aught to be done for this, because counting partial ballots the way we do
> /does/ break the algorithm, just as you described. Does doing so change
> the outcome in this case?
Yes. The results would look like this instead:
Geom Mart Moon Peng Pylo Subd Trip Zamb
Geom X 18 3 -1 4 9 20 -6
Mart -18 X -20 -27 -18 -13 5 -27
Moon -3 20 X -9 3 16 26 -13
Peng 1 27 9 X 7 11 23 -2
Pylo -4 18 -3 -7 X 12 14 -3
Subd -9 13 -16 -11 -12 X 7 -8
Trip -20 -5 -26 -23 -14 -7 X -25
Zamb 6 27 13 2 3 8 25 X
1. Zamboni Wars
2. Penguin Soccer
3. Geomancy
4. Moon Shot
5. Pylon
6. Subdivision
7. Martian Coaster Chaturanga
8. Trip Away
However, I am sure that the ballots would not look the same if the
voters knew that unranked candidates would lose. For one thing, every
designer's ballot would have his own game added to the top. For
another, some people would either not vote at all, or impose some
hasty ranking on the unplayed games based on browsing the rules. It
would have been a completely different contest.
I am totally in favor of allowing incomplete ballots. I just think
there is some research to be done as to how best to adapt methods that
were designed for complete ballots.
--dougorleans@xxxxxxxxx