thanks for the pdf of Preskill's lecture. It is helpful.

regarding depolarization/state reduction, i would encourage you to think about the concept of impedance matching as it applies in quantum optics.  Professor Stroud of Rochester works with impedances in optics, though I don't think he has adopted a full quantum formalism for that approach. You might try emailing him with your question. 

for a more general perspective on the role of impedance quantization see my author page http://vixra.org/author/peter_cameron

and particularly the state reduction paper. Not exactly what you're looking for, but you might find it helpful

http://vixra.org/author/peter_cameron

Simply.....or intuitionally.....^^;;;;

Depolarizing channel means that the error occurs to all direction in a qubit. For example......three directions of spin (X, Y, Z).

So, quantum state (pure), which is on surface of Bloch shpere, will be the mixed (classical) state, which is in Bloch sphere.

Also, the depolarizing error is often considered to total error (it's not exactly physical terminology.....^^;;). 

I attached file for your question.

It may be helpful to think about depolarizing channels in another way:  with some probability the qubit is discarded and replaced with a qubit in a maximally mixed state; otherwise it is unchanged.  So the state of the qubit becomes a mixture of the original state and the maximally mixed state.

In the Bloch sphere picture, the original state is represented by some vector, and the maximally mixed state is represented by the zero vector.  A mixture of the two will be presented by a linear combination of the two vectors, which will be in the same direction as the vector for the original state, but shorter.