Monday, November 07, 2005

Number 1-2-3-4 with a Bullet

Tuesday’s at-large councilor vote in Boston has some of us weighing bullet voting again. Do you vote for fewer than the maximum allowable candidates, to give your favorite(s) a better shot in a smaller number of total votes cast?

Mathematicians don’t favor it. Then again, they and some policy wonks often favor preference voting to cumulative voting. However, we do not have that option.

A search for bullet voting offers too many choices. Fortunately, a Dorchester blog learning strategies not only discusses the key issues, but it also plugs in the names of the local candidates.

Come Tuesday at 7 a.m., Boston voters can choose up to four of eight at-large candidates. Of course, most vote for four, even if they only feel strongly in favor to one or two.

By the bye, in case you have forgotten, the two that you need to feel strongly about are Felix Arroyo and Sam Yoon.

As Larry Davidson at learning strategies notes, whether you vote for one, two, three or four candidates depends mathematically on how you think the rest of the electorate views the race.
  1. If you think a favorite is on the edge of winning or losing by a few votes, a single bullet vote for that candidate will least dilute the outcome and may make the difference.

  2. On the other hand, if a favorite is very likely to lose, a bullet vote might rob your second or third choice of enough votes to beat someone you don’t like.
Note: Start with the blog post and check the Bayesian probabilities link if you want to see the stats to back up the concepts.

In this case, I am betting that my two favorites have high enough profiles and enough endorsements to win. I still have to decide whether I feel strongly enough about other candidates to try to make sure no one climbs a ladder run because of my behavior. Think. Think. Think.

No comments: