Scientific references for anti-cheating technology?

I've been digging and every web site I hit on this topic lacks any verifiable documentation of how it detects anti-cheats.  Basically the principle I've homed in on is that the software compares a player's moves to those of similar ranked players in similar situations and compares that to the output of an engine. What I want is to see some more of this technology.  I can understand it's easy to catch someone who plays well against an 1800 rated player just by talking with them. "Why did you play e3?" is followed by a poor understanding of the position for instance.  But what about that 1800 rated player who 'sometimes' cheats?  I've heard of GM's being 'busted' and I can't help but be suspicious of the methods when noone - even the professors working on it - have any real documentation I can look at and then run some tournament games through it and see if it falsely accuses.  

My interest is purely scientific, just look at my awful play and you can see I am improving but still suck at 1350ish. I just want to see some of the 'proof' that's out there.  Chess.com can keep their methods secret, they are in a position where I think that's best.  But chess.com isn't the only source of anti-cheating activity.  I suppose their access to 1M games / day conveniently in electronic format helps with the database comparisons of rating vs. move choice.  However I'm still just stumped in situations where a game has like 4 moves per position that are within 50 centipawns of each other with the 5th best move being 75 centipawns off. How can a distinction be made in that area? Is it some statistics anomaly that I'm not educated on?

Could a chessbase database be useful for detecting cheating at the < 1800 level?  Does anyone really cheat who is rated under 1800? 

I almost had a minor in psychology when I went to college so I have a fascination with predicting human behavior and chess is a very interesting area to study it.  Just look at how accurate ratings are at predicting odds of a win.  Vegas can set odds and guarantee profit over hundreds of games using ratings.

Any references you can find that show more about HOW the checks are made, preferably with public information on the databases used, would be appreciated.  

K.  I'll look there.

 If the topic is forbidden here, that explains why I kept finding dead threads on it. 8)

He was informed this was the wrong place to discuss such things so did the right thing and deleted the OP since you can't delete threads. He did the honourable thing in my opinion. He should be commended.