[EM] The Equal Vote Coalition and robla

[Kristofer Munsterhjelm]

On 2025-04-22 02:33, Richard via Election-Methods wrote:

> * When software developers write IRV code to correctly count overvotes, 
> usually they use decimal numbers, such as giving one ballot half 
> strength for one equal-ranked candidate and half strength for the 
> voter's other equal-ranked candidate.  This shortcut yields the correct 
> result, in the sense that the correct candidate gets elected.  However 
> this decimal-number shortcut is not acceptable for governmental elections.

What makes this a shortcut and why is it not acceptable? I imagine that 
what that is must be specific to the US, since some local governmental 
elections in NZ use Meek STV which can produce keep values with no 
closed form. However, I don't know what it is.

If it's a matter of numerical precision, then for overvotes (but not 
Meek) one can use exact fractional arithmetic.

> * Here's how to correctly count overvotes in governmental elections, 
> without departing from the principle of one voter supporting only one 
> candidate in each round of voting.  First, imagine IRV is being used in 
> a huge convention hall filled with all the voters and all the 
> candidates.  Each voter lines up behind their first-choice candidate. 
> The candidate with the shortest line is eliminated.  The voters in that 
> shortest line then move to stand in line behind their second choice. 
> (All other voters stay where they are.)  Now imagine there were two 
> voters in that shortest line who have an equal preference for candidates 
> B and D as their second choice.  These two voters agree that one of them 
> will stand in line behind candidate B, and the other voter will stand in 
> line behind candidate D.  This "pairing" approach easily extends to five 
> candidates being marked in the same "rank" column.  (More than five 
> "equal ranks" can be calculated, but not as easily, and not as fast.)

What would happen if there were three such voters, all voting B=D? To be 
fair, it would seem you would have to place one and a half voter in B's 
line and the other in D's. But your constraints say you can't divide the 
voters into fractions.

-km