The Boston Globe has an interesting article on the role of bullet voting in Boston local elections.

Four of 13 city council seats are elected at-large. Seeking bullet votes to at-large seats is a semi-effective way for political minorities to marshal citywide support and win seats. As a strategy, bullet voting involves using less than one’s full voting strength. In other words, the voter ‘withholds’ votes from all but the most preferred candidate. Bullet votes seem to be the norm in Boston:

In the last election for Boston City Council, each ballot contained an average of 2.9 votes for at-large seats.

Parts of the article suggest that some view the practice in a negative light. At any rate, there’s a hint of how STV could preserve the benefits of and eliminate many of the problems with bullet voting.

The danger is when voters don’t use one of their votes for a popular candidate they support, gambling that that candidate will win anyway. If enough voters make that same calculation, the candidate could lose.

(In Cambridge’s more complex electoral system, however, mathematicians say there is no value to bullet voting. Voters rank the order of their votes, so their first-choice candidate automatically receives a boost over others they may choose but give lower ranks.)

From 1938 to 1967, Massachusetts cities could adopt proportional representation for local elections. Plan E was one of six state-sanctioned charter formats, this one mandating council-manager government with PR elections and no party primaries.

For the Globe, effective bullet voting has high search and info costs:

But deciding whether you should cast a single ballot for your first-choice candidate, mathematicians say, you have to make strategic judgments about the race and how others will vote. And that requires a sophisticated view of the election.

With PR-STV and Plan E, everybody basically bullet votes. Surplus transfers and eliminations increase one’s likelihood of casting an effective vote absent tortuous calculation about whom to rank first.