digraph { layout=dot rankdir=TB overlap=false concentrate=false bgcolor=lightblue //splines="" node [style=filled shape=box color=blue4 fontcolor=white] label="Estimate or model probabilities. Use user generated\nBayesian Belief Models when it matters enough. " Predictability [shape=hexagon] [shape=hexagon] }
Estimate or model probabilities. Use Bayesian Belief Models when it matters enough.
Predictability of the future effects of actions.
Predictability requires sufficiently accurate models of the system that is being regulated and of the regulator itself, to be able to rank the actions and strategies that will give the best outcome. This is nothing other than the good regulator theorem.