You're basically saying the so called "distributed trustless system" of Bitcoin relies entirely on trusting 1-3 guys.  This is obviously not a valid security model.  What is my personal definition of when a PoW network is secure?  If it relies on pools, I would say most likely never.  Best case scenario, secure temporarily over a brief period of time.  The sheer amount of attack vectors for pools renders the entire model nonfunctional.

PoW in current form, with pool mining, is an obvious dead end.  It's excellent for distribution (i.e. before an ASIC is created), but useless for anything else.  The only solution I see to move forward out of this already failed model, is to utilize proof of stake and introduce the variable of reputation to fix most of proof of stake's current issues, such as having no finite resource in the system, the root of most of stake's problems.

If someone is going to claim I'm wrong, then I hope you have some kind of method for enforcing p2pool for PoW at the protocol level?  And can defeat share withholding attack at the same time?  I'm going to try to move forward instead with the other system I described.  The current system is a complete dead end.

I have created a TLDR version of my post: