Do You Have Any Concerns About Doing Too Many Comparisons?

With all these efforts, it seems like you've really lowered the bar for doing a lot of these comparisons and looking at a broad spectrum of data. With that in mind, do you have any concerns about opening up the opportunity for doing too many comparisons and finding some spurious correlation? Now we're kind of making it up front that you're explicitly comparing these thousands of different things until you can be statistically open about that comparison. Under that it tends to work really well the signals that people want to measure. Normally, they do have the power to compare across thousands of different thing.