Rating investment products

Although not the sole factor (by far) that led to the financial crisis, a failure of ratings agencies to provide accurate estimates of the risk associated with various investment products definitely played a big role. (i.e. investments were rated as being safer than they actually turned out to be -- at least this is what is commonly believed)

I'm wondering whether ratings agencies should take notes from places that continually calculate risk, like authorized bookies in Vegas who produce odds for various bets. If the odds of an outcome are set properly, one would expect roughly even numbers of bettors for both outcomes. If an overly large number of bettors choose to bet on one outcome instead, then that may be taken as a signal that the odds are not set correctly and a bookie (or whomever is handling the bets) can adjust the odds based on the behavior of the crowd.

As has been reported recently, many firms and organizations bet that certain CDOs would fail. A ratings system based on crowd-sourced information would consider these bets as an indication that these produces actually have more risk than previously assessed. In other words, treat all players in the system as more or less rational actors and use their betting behavior as economic indicators to set risk rather than relying solely on independent estimates that may be biased/incorrect. This type of rating system would, of course, depend heavily upon a databased of what bets people are taking, but that's what the SEC is for, right? (i.e. don't allow firms to make bets hidden to the SEC)