It's a pretty dumb methodology, but it's kind of OK as a basic meter of how 30 big companies are doing and as a proxy for big companies in general. As long as you're mostly comparing between time periods and adjusting for splits and what not.