By Bill James What Nate Silver does on fivethirtyeight.com — and this is my understanding of what he does, not his description of it — is search data for the underlying order of the universe that is depicted by those facts and statistics. Every topic in the political cosmos is blanketed by data. What Silver, 31, is very, very good at is finding and assembling that information on what you and I might envision as some terrifying backroom machine but that in reality is probably just a pretty good laptop. Silver is equally good at avoiding searches that produce only the elements that prove his point of view. Let us take, for example, gay marriage. It's not that Silver doesn't have an opinion on the issue; he probably does. It's that he doesn't start his analysis with his opinion; he starts with the data. He finds every vote on the issue of gay marriage that he can — every poll, every survey, all the related factors. Then he looks for aspects that might reveal changing attitudes. Where was the public on this issue in 2004? Where was it in 2006? Where is it now? What are the numbers from Nebraska? What are the numbers from California? How are the numbers from Nebraska different from those from California? After studying all of it, he reaches a conclusion: Gay marriage is gathering voter support by about 2 points a year, and within three or four years, gay-marriage advocates are going to start winning elections, and here's when that's going to happen in California and in Nebraska and in Arkansas. Is he right? I don't know. The point is not how precisely he calls the results but that after reading his analysis, you actually know something you didn't know when you started. In a world choking on retreaded arguments long worn bald of the facts, this type of analysis has proved to be stunningly — and reassuringly — popular.