Cutting to the chase, the fundamental error has been to assume that
write-in or so-called inconsequential candidates can be batch-
eliminated before having results from the whole election. No precinct
knows what can be eliminated until it has the results from other
precincts for the first round. Further a method must accomodate not
just a most-common scenario but also all possible scenarios. Runoff
voting in general encourages candidate counts to increase. Cf. San
Francisco.
We are talking about the matrix size necessary to fully canvass an IRV
election centrally from initial data provided by each precinct. That
initial data might categorize all write-candidates into a single pile,
but the risk is that if reports from other precincts indicate possible
significance, it would be necessary to ask the precincts to tabulate
the write-in pile. If you did this with so-called minor candidates,
you'd see a lawsuit, which is less likely with write-ins.
----
Election-Methods mailing list - see http://electorama.com/em for list info
