So, I had been planning on keeping Akismet running until Thursday, but well, it wasn't really working. With all my previous efforts in place, I had about a 99.2% block rate. Sure, it wasn't perfect, and there were a few false positives, but with Akismet along, we were sitting at 89.5%. Now, while I was running Akismet, we had 550 attempted posts, 492 were blocked by Akismet. Had I been running my blocks, 544 would have been blocked. We have had 6 actual/non-spam posts. 550-544=6. Looks like I win.
The experiment did force me to think about a few things. First and foremost, I need to start handeling repeat offenders. I have about a dozen ip addresses and ranges of addresses that need to be dealt with before they get an opportunity to even load a page. If nothing else, this should minimize overhead. I have quite a few more tests I want to do with Akismet, that I would not have thought of beforehand, but they will require a much more robust comment tracking system than I have...so I guess I will have to write one.
