challenge: find the closest match to titor at tti.com. and if your feeling extra frisky, find the closest match to titor on the whole net!
The closest match was "Buzzmaker" on the Chi Square test. In the book, the same person wrote as the lawyer and as Mom Titor - but that writer wasn't Titor. The problem with these results is that the database of words is rather small - less than 2,500 words for each subject. The person who had the least chance of being Titor was TTA. As I recall it was less than one chance in 25,000 or thereabouts. TTA and Titor have substantial databases (15,000+ words each) so I'm quite satisfied with their comparison.
There are two huge problems with attempting to do what your challenge asks.
First, you have to have some specific targets. You then have to physically go through the entire forum to find their posts, copy & paste every post into a document. Not only do you have to copy and paste you have to cull out quotes from other posts because those are not the writer's words. That's several hours work for each target. But I have done about a dozen targets here and half that many on Anomalies over the past few years.
Second, you have to have a database of at least 5,000 words in the posts. It's best to have 20,000 words or more. Not many posters are that prolific. Having a too small database means that someone might "look" like Titor based on a dozen or so uses of key words in the test. That's a Type 1 error - False Positive. The same can happen the other way. Failing to find those key words in a small database makes it look like "Titor" (if he was the writer) isn't. THat's a Type 2 error - False Negative. With 20,000 words or more those anomalous and misleading results tend to a attenuated and you get a more accurate picture. WHen researchers apply the same test that I use to Shakespear vs. Bacon to see who wrote a certain play they have a database of 1 million+ words for each writer. The larger the database the more reliable the results.
Third, doing the entire Internet wouldn't work. Without specific targets you couldn't accomplish the task with an army of super-computers programmed to suck out, collate and file every word written by every individual ever to post (though I don't think that you were serious with this particular part of the challenge). But even if you could do this you'd still have to go back, by hand, and cull out the quotes, etc. That goes a fur-bit beyond a hobby...you'd spend the rest of your life preparing the data and probably never get it done.
