Here in Ecuador, there are 14 candidates for President at the moment, and most voters polled (70% last I heard) were undecided. I don't know what the election system is --- I'm guessing FPTP with runoffs, in which case Cynthia Viteri will probably win, just because her supporters have enough money to put her on billboards and make her familiar to the electorate.
The electorate would probably prefer a different candidate in the Condorcet sense, so a different choice, a better choice, would arise if the election used a better method such as approval voting. >From a certain point of view, the reason FPTP produces lousy results is that the voters don't effectively coordinate their votes. Political parties are one mechanism for doing this, but they have their limits, and they're not the only possible mechanism. For a strong example, imagine that 90% of the eligible voters enter a pact before the election. The day before the election, they will take part in an unofficial poll, using e.g. STV or approval voting. They pledge to vote, in the real election, for the winner of the unofficial poll. Clearly the unofficial poll would determine the official election results, and the system is fairly stable: voters will want to participate in order to affect the election results, presumably most people will keep their promise (in part because they'll like the poll winner better than the other likely winner --- indeed the promise is probably unnecessary), and it's fairly immune from interference. I speculate that just running the poll and publishing the results would ultimately have this effect, and we've only failed to see it so far because most existing polls are FPTP with a short candidate list. The more trustworthy the poll, of course, the more effective it will be. If it develops local chapters, volunteer poll workers, voter-verified paper ballots, etc., more of the electorate will trust its results and vote alongside its participants in the real election. It need not have a huge number of *participants* to have an effect; it needs to have a lot of *readers*, and more importantly, each reader needs to know about the number of othe readers, and trust that the poll results accurately reflect their preferences. Accordingly, media like websites with visitor counters, state and national news broadcasts and papers, and posters in public places count more than email, word of mouth, doorknob fliers, and perhaps even local papers.
