Funny little fact: I use a home-brewed plagiarism detection software to scan assignments from my students. Most of it is straight forward, but it has some smarts to spot and highlight patterns that may be missed by a simple linguistic analysis. I'm pretty proud of it.
To make a long story short, I randomly chose 20 "Politics" forum posts by: Dos Equis ("Student D.E."), iwantmass ("Student W.M."), The Coach ("Student T.C."), Agnostic007 ("Student J.B.") and myself, and I ran them through the fancy tool.
The software thinks there's an 11% probability of Dos Equis plagiarizing iwantmass and only a 9% chance of iwantmass plagiarizing Dos Equis - anything less than 20% is way too low in my opinion and likely a false positive. The sample size was small but that knife cuts both ways in this case.
Anyways... I'm watching all of you.
