[EM] Teams

[Daniel Carrera]

On Fri, Jul 9, 2021 at 4:27 PM Susan Simmons <suzerainsimmons at outlook.com>
wrote:

> Did anybody recognize Teams as a reformulation of River?
>
> It shows that it is not necessary to explicify the implicit tree structure
> induced by the successive engulfing of one team by another.
>
> Does this simplified formulation increase River's chances of public
> adoption?
>

It took me some effort to understand the Teams method, but maybe I'm not a
good test. I'm only a little smarter than your average koala ;-)  But let
me offer a different way to write the Teams / River method that might be a
little easier to follow. Among other things, I removed the sports analogy,
moved some of the core concepts higher up, and inserted some redundancy.
Let me know what you think:

-----
Overview... the count is based on the pairwise support matrix.

The winner is elected via a series of rounds in which candidates are merged
into groups, where each group has a unique head. The first round begins
with every candidate being the sole member of a one-member group. Each
round consists of two steps:

1. Identify two groups to merge. Find the group “A” for which the head of
the group has the greatest pairwise loss against any member of another
group.

2. Group “A” is merged into the group of the pairwise winner from step (1)
and the previous head of group “A” is no longer the head of any group.

The rounds end when all candidates have been merged into one group. The
head of this one group is elected.
-----

Cheers,
Daniel
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.electorama.com/pipermail/election-methods-electorama.com/attachments/20210709/5b3d7cd2/attachment.html>