Businessweek has an interesting section online - the CEO's guide to prediction markets. Yes, they can be cool and work well, but market participants need some incentive to take them seriously. At this point, there is probably a better chance of setting them up wrong than getting them right.
