Are prediction markets too stable, because people trust them too much and assume new information already baked in? I explore this in a post on Slate with Andrew Gelman.