Adaptive Markov

Neither the House, Senate or state capitals have sufficient banwdith to maintain partition. We know we are short N.  So we engage in continual adaption and the color rates for our three color is determined by the adapting encoding of the state, House and Senate events with the voter/taxpayer.

Sand box trading pits do this, keep a running encoding of S to L while maintaining bound variance at the pit boss. Sandbox runs a low enough Markov to be separable and no more, the second level up. But we can show quasi stability via the three principles, free entry/exit, round robin access to the bidding queues and bound market making risk. Quasi stable because it will step down incrementally without queue collapse, adiabatic, one node change at a time with round robin robin fair peeking.  Sandboxers know this quite well. Most of the major on line exchanges get this problem and can be quite stable.

So, given constant rates in the coloring problem, we can algorithmicly trim the factorial tree ex ante, this implies constant N. We should be able to write an algorithm to do that for any set of prime counters. Counters no sequentially prime will color some other object for the topologists to discover. I was doing a lot of finger counting, couldn't do seven, required two hands/.