optimal value, belief recursion: 0.716320;  history enumeration: 0.716320;  difference 0.0e+00
belief after (n1, n0) observations:  (1,0): 0.8000  (2,0): 0.9412  (2,1): 0.8000  (3,0): 0.9846  (3,1): 0.9412  (4,0): 0.9961
value of answering immediately: 0.000;  value of the best fixed number of probes n then majority answer:
   n = 0: 0.0000
   n = 1: 0.5500
   n = 2: 0.5000
   n = 3: 0.6420
   n = 4: 0.5920
   n = 5: 0.6342
   n = 6: 0.5842

best value with bounded memory (exhaustive search over deterministic memory policies)
memory                        states     value
none (answer at once)              1    0.0000
window of last 1 observation(s)       3    0.5500
window of last 2 observation(s)       7    0.6949
counter clipped to +-1             3    0.5500   policy: -1:ansD, +0:probe, +1:ansH
counter clipped to +-2             5    0.7112   policy: -2:ansD, -1:probe, +0:probe, +1:probe, +2:ansH
counter clipped to +-3             7    0.7112   policy: -3:probe, -2:ansD, -1:probe, +0:probe, +1:probe, +2:ansH, +3:probe
full belief (unbounded)          inf    0.7163
