number of prefixes (states)        : 341   formula (V^(T+1)-1)/(V-1) = 341
J by enumerating 256 responses   : 0.6549259034
J = V^pi(empty prefix) by recursion: 0.6549259034
max |E[R | prefix] - V(prefix)|      : 8.88e-16
max |E_a A(s, a)| over all states    : 2.71e-16

example response (1, 1, 2, 1): prob=0.00269, reward=+1.50
  V(prefix ()) = +0.6549
  V(prefix (1,)) = +0.6938
  V(prefix (1, 1)) = +1.0524
  V(prefix (1, 1, 2)) = +1.4354
  V(prefix (1, 1, 2, 1)) = +1.5000
