Thanks for that input. Fully aware of that. Just saying that using a statistical model would be much better than huge lists of URLs.