So what are the commonalities between the following
- Delusional belief systems
- Sock puppets
- Cults
- CheckUser
What these have in common is that: the intellectual basis of the processes is never displayed in public; the relationship with reality is one of control rather than reflection; they exist solely to perpetuate existing structures.
Let me focus on Checkuser. This is a secret process, apparently capable of providing incontrovertible and absolute truth, but which is so fragile that it cannot be exposed to public scrutiny for fear that it will be irreparably damaged.
Let us clear away the minor issue that Checkuser does not, and cannot, provide absolute certainty. Full stop. But then, this is not its true purpose.
So what is it all about? It is a process which declares its victim in advance to be irrefutably guilty. In other words, it is a blatant exercise of the power of the system. The arbitrariness of the process is part of its intimidatory value. Users who believe in its veracity will be impressed by the power, authority and superior knowledge of the admins. Users who know full well that it cannot do what it claims will understand that it represents the exercise of arbitrary power. Some may believe that the system is capable of improvement or reform: this is a category error. It is already superbly adapted to its true function.
A clear indication of the intrinsically corrupt nature of the process can be seen in the number of sock-puppet blocks which follow a mere assertion of checkuser evidence. It is clear, and intended to be clear, that these are punishment blocks against users who happen to fall foul of influential and well-connected admins.
There's not a lot of point in trying to establish the intrinsically unreliable nature of the underlying technical process. To do that we would need to assemble a statistically adequate sample of the four categories: sock-puppets who were found out by CU; sock-puppets who were not found out by CU; non-socks who were falsely accused by CU and non-socks who were acquitted by CU. This is after all impossible -- and has certainly never been attempted by those who trumpet, and possibly even believe, the infallibility of CU.
So let's attack it from the other end. There are three levels to CheckUser abuse: overstating CU results; premature or improper CU cases; and outright mis-statement of CU results (or even non-results). The first is technically ; the second is easy to see from the record; the third has certainly happened.
Are there any particularly well-documented and egregious cases that anyone would like to share?
Checkuser, someone must have been telling lies about Joseph K.
Grep Checkuser 
