Princeton University neuroscience professor Sam Wang (whose approach to debunking political smears is must-read material) has revived his meta-analysis of state polls.
There are two key things that makes Sam's analysis unique. First, he's not trying to predict anything; rather, his analysis aims to determine where we are at now -- not where we will be on November 4. Second, instead of running simulations or simply averaging available data, he uses a mathematical formula to determine the probability of electoral vote distributions for each state based on current polling data.
A simpler way of thinking about it is that his approach should quickly reflect changes based on events in the campaign. Basically, he's trading off the potential of predictive value for the precision of a real-time snapshot.
With that preamble out of the way -- and I hope I haven't butchered anything -- Sam's model is already showing a bounce from McCain's "Seven Kitchen Tables" Gaffe, boosting Obama's lead over McCain in the current snapshot by about 20 electoral votes.
© Jed Lewison