A possible vulnerability with BAV might arise when the voters are polarized. Being unwilling to consider compromise, such voters may (in the extreme) vote against every candidate other than the one candidate they most favor and the result would be that no candidate would receive more than a very small positive net vote. Quite possibly, every candidate might even have a negative net vote. If there were a candidate with negligible name recognition, that candidate might actually have an advantage over other candidates because some voters would feel it unreasonable to vote against a candidate they know nothing about.
An organization (such as a large union or church, the NRA or the KKK) might recognize this as an opportunity. They could arrange for a stealth candidate to be on the ballot who subsequently would avoid responding to media-interview requests. Internally the organization would urge its loyal members to support this stealth candidate (while also urging those members to avoid discussing the matter publicly). The hope of the organization would be for the stealth candidate receive a net vote that, while small, is nevertheless larger than that of any other candidate. The danger of this would seem particularly high when BAV is first adopted and voters are both polarized and in the habit of thinking that no little-known candidate has even a remote chance of winning an election.
Even before an election is held, someone aware of the possibility of this kind of scam might take note of the odd behavior of such a stealth candidate. But what could be done?
A government agency or even a volunteer organization could take on the task of bringing this kind of stealth candidate activity to the attention of voters (this might seem to be a responsibility of the fourth estate but that does not mean the media would see it that way). Such an organization could actively seek to interview candidates to document their stands on issues as a guide to voters. A web site could make these interviews and other relevant information easily available for citizens seeking to learn about the candidates. But if some candidate refuses to be interviewed or provide such information then that refusal would appear prominently on that web site, perhaps with an alert to voters to the danger of such a candidate winning. Some judicial mechanism for challenging candidates on the basis of being a stealth candidate might even be considered.
After a BAV election is held, one would expect statistical oddities to be apparent in voting patterns that would suggest a reason to disqualify a candidate. For example, a successful stealth candidate would likely have only a small number of support votes, but also an exceptionally low number of opposition votes compared to support votes; there would be an unusually large proportion of abstentions. When viewed along with a past history of avoiding public campaigning, this could serve as supporting evidence for disqualifying an otherwise winning candidate.
A weakness in interpreting such statistical evidence is that when a voter fails to express either opposition or support that may be in spite of the voter knowing a great deal about the candidate. The abstention may simply be due to the voter being indifferent or undecided about a candidate even though being knowledgeable about the candidate.
But ballots could be altered to allow voters to report on that distinction, perhaps by checking a third box or perhaps by asking voters to draw a line through unrecognized candidates' name as an indication of having no knowledge about the candidate or perhaps even no awareness of the candidate. In tallying the net vote, that would be a distinction making no difference, but the extra information would serve as evidence for distinguishing a stealth candidate (who should be disqualified) from a candidate whom voters may widely deem undistinguished, but who is nonetheless the candidate whom the voters, on balance, actually do prefer as the best available consensus candidate.
